CN114613005A - Pedestrian trajectory determination method, system, device and medium - Google Patents
Pedestrian trajectory determination method, system, device and mediumDownload PDFInfo
- Publication number
- CN114613005A CN114613005ACN202210210798.7ACN202210210798ACN114613005ACN 114613005 ACN114613005 ACN 114613005ACN 202210210798 ACN202210210798 ACN 202210210798ACN 114613005 ACN114613005 ACN 114613005A
- Authority
- CN
- China
- Prior art keywords
- human body
- pedestrian
- image acquisition
- feature
- pair
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
- Traffic Control Systems (AREA)
Abstract
Translated fromChinese
Description
Translated fromChinese技术领域technical field
本发明涉及轨迹追踪技术领域，具体提供一种行人运动轨迹确定方法、系统、装置和介质。The invention relates to the technical field of trajectory tracking, and in particular provides a method, system, device and medium for determining a pedestrian movement trajectory.
背景技术Background technique
多目标轨迹聚合是指通过多个相机获取多个目标的运动轨迹，以确定每个目标各自的整体运动轨迹。多目标轨迹聚合在商场、园区、机场、安防等场景都有着广泛的应用。现有技术中的多目标轨迹聚合方法往往使用单镜跟踪算法获取目标在单个相机内的运动轨迹，再通过目标的人体外貌特征进行多个相机之间的运动轨迹之间的关联，从而获取目标在多个相机下的完整行动轨迹。然而，人体外貌特征并不会始终有效，如会存在人体外貌特征被遮挡、模糊、着装类似等情况，从而导致现有的多目标轨迹聚合方法获取的目标的整体运动轨迹的效果较差。Multi-target trajectory aggregation refers to acquiring the motion trajectories of multiple targets through multiple cameras to determine the respective overall motion trajectories of each target. Multi-target trajectory aggregation has a wide range of applications in shopping malls, parks, airports, security and other scenarios. The multi-target trajectory aggregation methods in the prior art often use a single-lens tracking algorithm to obtain the target's motion trajectory within a single camera, and then use the target's human appearance features to associate the motion trajectories between multiple cameras to obtain the target. Full action trail with multiple cameras. However, human appearance features are not always effective. For example, human appearance features may be occluded, blurred, or similar in clothing, etc., resulting in poor overall motion trajectories of targets obtained by existing multi-target trajectory aggregation methods.
本领域需要一种新的行人运动轨迹确定方案来解决上述问题。There is a need in the art for a new pedestrian motion trajectory determination solution to solve the above problems.
发明内容SUMMARY OF THE INVENTION
本发明旨在解决上述技术问题，即，解决或者部分解决如何在不保证行人的人体外貌特征始终有效的条件下，仍然能够实现准确、有效地将位于图像采集范围内的行人运动轨迹进行关联，获得行人的完整的运动轨迹的问题，本发明提供了一种行人运动轨迹确定方法、系统、装置和介质，包括：The present invention aims to solve the above-mentioned technical problems, that is, to solve or partially solve how to accurately and effectively associate the motion trajectories of pedestrians within the image acquisition range without guaranteeing that the pedestrian's human appearance features are always valid, For the problem of obtaining the complete motion trajectory of pedestrians, the present invention provides a pedestrian motion trajectory determination method, system, device and medium, including:
在第一方面，本发明提供一种行人运动轨迹确定方法，所述方法包括：In a first aspect, the present invention provides a method for determining a pedestrian motion trajectory, the method comprising:
获取位于不同图像采集设备的图像采集范围内的行人运动轨迹；Acquire pedestrian motion trajectories within the image acquisition range of different image acquisition devices;
获取不同图像采集设备之间的关系特征；Obtain the relationship characteristics between different image acquisition devices;
获取每个行人运动轨迹各自对应的人体特征信息；Obtain the human body feature information corresponding to each pedestrian movement trajectory;
根据所述人体特征信息和所述关系特征，对属于同一行人的行人运动轨迹进行关联，生成对应各行人最终的行人运动轨迹。According to the human body feature information and the relationship feature, the motion trajectories of pedestrians belonging to the same pedestrian are associated to generate the final pedestrian motion trajectories corresponding to each pedestrian.
在上述行人运动轨迹确定方法的一个技术方案中，“根据所述人体特征信息和所述关系特征，对属于同一行人的行人运动轨迹进行关联，生成对应各行人最终的行人运动轨迹”的步骤具体包括：In a technical solution of the above-mentioned pedestrian motion trajectory determination method, the steps of "associating pedestrian motion trajectories belonging to the same pedestrian according to the human body feature information and the relationship feature, and generating the final pedestrian motion trajectory corresponding to each pedestrian" are specific. include:
根据所述人体特征信息确定不同行人运动轨迹之间的特征相似度；Determine the feature similarity between different pedestrian motion trajectories according to the human body feature information;
根据所述特征相似度和所述关系特征，计算不同行人运动轨迹之间的关联分；According to the feature similarity and the relationship feature, calculate the correlation score between different pedestrian movement trajectories;
若两个行人运动轨迹之间的关联分大于预设的分数阈值，则判定所述两个行人运动轨迹属于同一行人的行人运动轨迹；If the correlation score between the two pedestrian movement trajectories is greater than the preset score threshold, it is determined that the two pedestrian movement trajectories belong to the pedestrian movement trajectory of the same pedestrian;
对属于同一行人的行人运动轨迹进行关联生成对应各行人最终的行人运动轨迹；和/或，Associating pedestrian motion trajectories belonging to the same pedestrian to generate final pedestrian motion trajectories corresponding to each pedestrian; and/or,
所述关系特征包括不同图像采集设备之间的空间关系特征和不同图像采集设备之间的时间关系特征，其中，两个图像采集设备之间的空间关系特征用于表示在所有图像采集设备相互关联形成的拓扑结构中所述两个图像采集设备之间的关联权重，两个图像采集设备之间的时间关系特征用于表示行人花费不同时长由一个图像采集设备的图像采集范围运动至另一个图像采集设备的图像采集范围的概率。The relationship features include spatial relationship features between different image capturing devices and time relationship features between different image capturing devices, wherein the spatial relationship features between two image capturing devices are used to indicate that all image capturing devices are related to each other. The association weight between the two image capture devices in the formed topology structure, and the time relationship feature between the two image capture devices is used to indicate that pedestrians spend different time periods moving from the image capture range of one image capture device to another image The probability of the image acquisition range of the acquisition device.
在上述行人运动轨迹确定方法的一个技术方案中，“根据所述特征相似度和所述关系特征，计算不同行人运动轨迹之间的关联分”的步骤具体包括通过下式计算所述关联分：In a technical solution of the above pedestrian motion trajectory determination method, the step of "calculating the correlation score between different pedestrian motion trajectories according to the feature similarity and the relationship feature" specifically includes calculating the correlation score by the following formula:
zij＝eij-decay+decay×sijzij = eij -decay+decay×sij
其中，所述zij表示两个行人运动轨迹之间的关联分，所述eij表示两个行人运动轨迹之间的特征相似度，所述sij表示两个行人运动轨迹各自对应的图像采集设备之间的关系特征的关系特征，所述decay表示预设的衰减因子；Wherein, the zij represents the correlation score between the two pedestrian motion trajectories, the eij represents the feature similarity between the two pedestrian motion trajectories, and the sij represents the image acquisition corresponding to the two pedestrian motion trajectories The relationship feature of the relationship feature between the devices, the decay represents a preset decay factor;
当所述关系特征包括所述关系特征包括不同图像采集设备之间的空间关系特征和不同图像采集设备之间的时间关系特征时，所述sij表示两个行人运动轨迹各自对应的图像采集设备之间的空间关系特征与时间关系特征的乘积。When the relationship feature includes that the relationship feature includes a spatial relationship feature between different image capturing devices and a temporal relationship feature between different image capturing devices, the sij represents the image capturing devices corresponding to each of the two pedestrian movement trajectories The product of the spatial relationship feature and the temporal relationship feature between.
在上述行人运动轨迹确定方法的一个技术方案中，所述方法还包括通过下式计算两个行人运动轨迹之间的特征相似度eij：In a technical solution of the above pedestrian motion trajectory determination method, the method further includes calculating the feature similarity eij between two pedestrian motion trajectories by the following formula:
eij＝eiT·ejeij =eiT ·ej
其中，所述ei表示一个行人运动轨迹的中心向量，所述ej表示另一个行人运动轨迹的中心向量，所述T表示中心向量ei的向量转置；Wherein, the ei represents the center vector of a pedestrian motion track, the ej represents the center vector of another pedestrian motion track, and the T represents the vector transpose of the center vector ei ;
在上述行人运动轨迹确定方法的一个技术方案中，所述关系特征包括不同图像采集设备之间的空间关系特征和不同图像采集设备之间的时间关系特征，其中，两个图像采集设备之间的空间关系特征用于表示在所有图像采集设备相互关联形成的拓扑结构中所述两个图像采集设备之间的关联权重，两个图像采集设备之间的时间关系特征用于表示行人花费不同时长由一个图像采集设备的图像采集范围运动至另一个图像采集设备的图像采集范围的概率，“获取不同图像采集设备之间的空间关系特征和时间关系特征”的步骤具体包括：In a technical solution of the above-mentioned pedestrian motion trajectory determination method, the relationship feature includes a spatial relationship feature between different image capturing devices and a temporal relationship feature between different image capturing devices, wherein the relationship between the two image capturing devices The spatial relationship feature is used to represent the association weight between the two image acquisition devices in the topology structure formed by the correlation of all image acquisition devices, and the temporal relationship feature between the two image acquisition devices is used to represent that pedestrians spend different time The probability that the image capture range of one image capture device moves to the image capture range of another image capture device. The steps of "obtaining the spatial relationship features and temporal relationship features between different image capture devices" specifically include:
获取通过每个图像采集设备采集到的行人图像得到的人体特征样本，其中，所述人体特征样本包括根据所述行人图像提取到的人体特征，采集所述行人图像的图像采集设备的设备编号；acquiring a human body feature sample obtained from a pedestrian image collected by each image capture device, wherein the human body feature sample includes the device number of the image capture device that collected the pedestrian image based on the human body feature extracted from the pedestrian image;
针对每个人体特征样本，从所有人体特征样本中获取与当前人体特征样本的人体特征相似的相似人体特征样本，并将当前人体特征样本分别与每个相似人体特征样本组成一个匹配样本对；For each human body feature sample, obtain similar human body feature samples from all the human body feature samples that are similar to the human body features of the current human body feature sample, and form a matching sample pair with the current human body feature sample and each similar human body feature sample respectively;
针对每个匹配样本对，获取当前匹配样本对中的两个人体特征样本xi和xj各自对应的图像采集设备的设备编号ci和cj，根据所述设备编号ci和cj建立图像采集设备对(ci，cj)；获取当前匹配样本对(ci，cj)中的两个人体特征样本xi和xj各自对应的行人图像的采集时刻tsi与tsj，确定两个采集时刻tsi与tsj之间的时间间隔|tsi-tsj|，将所述时间间隔|tsi-tsj|添加到当前图像采集设备对(ci，cj)的时间间隔列表中；For each matching sample pair, obtain the device numbers ci and cj of the image acquisition devices corresponding to the two human body feature samples xi and xj in the current matching sample pair, and establish the image acquisition device according to the device numbers ci and cj Image acquisition device pair (ci , c j) ; acquisition times tsi and tsj of pedestrian images corresponding to two human feature samples xi and xj in the current matching sample pair (ci , c j), respectively, Determine the time interval |tsi -tsj | between the two acquisition moments tsi and tsj , and add the time interval |tsi -tsj | to the current image acquisition device pair (ci , c j) time interval list;
针对每个图像采集设备对，根据预设的分箱数对当前图像采集设备对(ci，cj)的时间间隔列表进行分箱处理，根据分箱处理后落入每个分箱中的时间间隔的数量建立当前图像采集设备对(ci，cj)的直方图，其中，所述直方图的横坐标表示行人由设备编号为ci的图像采集设备的图像采集范围运动至设备编号为cj的图像采集设备的图像采集范围花费的时间，纵坐标表示花费所述时间的概率；For each image acquisition device pair, binning is performed on the time interval list of the current image acquisition device pair (ci , cj ) according to the preset number of bins, and the The number of time intervals establishes a histogram of the current image acquisition device pair (ci , c j) , wherein the abscissa of the histogram represents the movement of pedestrians from the image acquisition range of the image acquisition device with device number ci to the device number is the time spent in the image acquisition range of the image acquisition device of cj , and the ordinate represents the probability of spending the time;
根据每个图像采集设备对的直方图，确定每个图像采集设备对中两个图像采集设备之间的空间关系特征和时间关系特征。According to the histogram of each image acquisition device pair, the spatial relationship feature and the temporal relationship feature between the two image acquisition devices in each image acquisition device pair are determined.
在上述行人运动轨迹确定方法的一个技术方案中，“根据每个图像采集设备对的直方图，确定每个图像采集设备对中两个图像采集设备之间的空间关系特征和时间关系特征”的步骤具体包括：In a technical solution of the above-mentioned pedestrian motion trajectory determination method, "according to the histogram of each image acquisition device pair, determine the spatial relationship feature and the time relationship feature between the two image acquisition devices in each image acquisition device pair". The steps include:
根据所述图像采集设备对的直方图的纵坐标确定所述图像采集设备对中两个图像采集设备之间的时间关系特征；determining a time relationship feature between two image capturing devices in the pair of image capturing devices according to the ordinate of the histogram of the pair of image capturing devices;
根据下述公式确定所述两个图像采集设备之间的空间关系特征：The spatial relationship characteristics between the two image acquisition devices are determined according to the following formula:
其中，kij指匹配样本对中的两个人体特征样本xi和xj各自对应的图像采集设备ci和cj之间的空间关系特征，nij指图像采集设备对(ci，cj)的时间间隔列表中时间间隔的总数，
在上述行人运动轨迹确定方法的一个技术方案中，在“根据分箱处理后落入每个分箱中的时间间隔的数量建立当前图像采集设备对(ci，cj)的直方图”的步骤具体包括：In a technical solution of the above pedestrian motion trajectory determination method, in the process of “establishing a histogram of the current image acquisition device pair (ci, c j) according to the number of time intervals that fall into each bin after binning” The steps include:
对分箱处理后落入每个分箱中的时间间隔的数量进行高斯平滑，获得高斯平滑后的每个分箱中的时间间隔的数量；Perform Gaussian smoothing on the number of time intervals that fall into each bin after binning to obtain the number of time intervals in each bin after Gaussian smoothing;
对高斯平滑后的每个分箱中的时间间隔的数量进行归一化，获得每个分箱对应的归一化结果；Normalize the number of time intervals in each bin after Gaussian smoothing to obtain the normalized result corresponding to each bin;
将每个分箱对应的数值作为横坐标，将每个分箱对应的归一化结果作为纵坐标，建立当前图像采集设备对(ci,cj)的直方图。Taking the value corresponding to each bin as the abscissa and the normalization result corresponding to each bin as the ordinate, the histogram of the current image acquisition device pair (ci , cj ) is established.
在上述行人运动轨迹确定方法的一个技术方案中，所述人体特征信息包括由图像采集设备采集到的行人图像提取到的人体特征，所述行人图像的采集时刻，所述图像采集设备的设备编号和所述人体特征所属的行人运动轨迹的轨迹编号；“根据所述人体特征信息和所述关系特征，对属于同一行人的行人运动轨迹进行关联，生成对应各行人最终的行人运动轨迹”的步骤具体包括：In a technical solution of the above method for determining pedestrian movement trajectory, the human body feature information includes human body features extracted from a pedestrian image collected by an image collection device, the collection time of the pedestrian image, and the device number of the image collection device. and the track number of the pedestrian movement track to which the human body feature belongs; the steps of "associating the pedestrian movement track belonging to the same pedestrian according to the human body feature information and the relationship feature, and generating the final pedestrian movement track corresponding to each pedestrian" Specifically include:
步骤S1：针对每个人体特征信息，从所有人体特征信息中获取与当前人体特征信息的人体特征相似的相似人体特征信息，并将当前人体特征信息分别与每个相似人体特征信息组成一个匹配特征对；Step S1: For each human body feature information, obtain similar human body feature information similar to the human body feature of the current human body feature information from all the human body feature information, and form a matching feature with the current human body feature information and each similar human body feature information respectively. right;
步骤S2：针对每个匹配特征对，判断当前匹配特征对中两个人体特征信息对应的轨迹编号是否相同；Step S2: for each matching feature pair, determine whether the track numbers corresponding to the two human body feature information in the current matching feature pair are the same;
步骤S21：若相同，则判定当前匹配特征对中两个人体特征信息属于的行人运动轨迹是同一个行人的行人运动轨迹；Step S21: if they are the same, then determine that the pedestrian movement trajectories to which the two human body feature information in the current matching feature pair belong are the pedestrian movement trajectories of the same pedestrian;
步骤S22：若不同，则获取当前匹配特征对中两个人体特征信息对应的设备编号对应的图像采集设备对的直方图，根据所述直方图获取当前匹配特征对对应的两个图像采集设备之间的空间关系特征；Step S22: If different, obtain the histogram of the image acquisition device pair corresponding to the device numbers corresponding to the two human body feature information in the current matching feature pair, and obtain the one of the two image acquisition devices corresponding to the current matching feature pair according to the histogram. The spatial relationship between the characteristics;
获取由当前匹配特征对中两个人体特征信息对应的采集时刻确定的时间间隔，以所述时间间隔作为横坐标查询所述直方图得到相应的纵坐标，将所述纵坐标作为时间关系特征；Obtain the time interval determined by the collection moments corresponding to the two human body feature information in the current matching feature pair, use the time interval as the abscissa to query the histogram to obtain the corresponding ordinate, and use the ordinate as the time relationship feature;
根据所述人体特征信息、所述空间关系特征与所述时间关系特征，判断当前匹配特征对中两个人体特征信息所属的行人运动轨迹是否属于同一个行人的行人运动轨迹；若是，则将两个人体特征信息对应的轨迹编号修改成同一轨迹编号；According to the human body feature information, the spatial relationship feature and the time relationship feature, it is determined whether the pedestrian motion trajectories to which the two human body feature information in the current matching feature pair belong belong to the pedestrian motion trajectory of the same pedestrian; The track number corresponding to the personal feature information is modified to the same track number;
步骤S3：按照人体特征信息对应的采集时间由先至后的顺序，对具有相同轨迹编号的人体特征信息进行排列，根据排列后的人体特征信息生成对应各行人最终的行人运动轨迹。Step S3: Arrange the human body feature information with the same track number according to the first-to-last order of the collection time corresponding to the human body feature information, and generate the final pedestrian motion track corresponding to each pedestrian according to the arranged human body feature information.
在上述行人运动轨迹确定方法的一个技术方案中，所述方法还包括通过下列步骤从所有目标人体特征中获取与当前目标人体特征的相似人体特征，其中，所述目标人体特征是所述人体特征样本或所述人体特征信息：In a technical solution of the above-mentioned pedestrian motion trajectory determination method, the method further includes obtaining a human body feature similar to the current target human body feature from all target human body features through the following steps, wherein the target human body feature is the human body feature Sample or said human characteristic information:
分别获取当前目标人体特征与其他目标人体特征之间的特征相似度；Obtain the feature similarity between the current target body feature and other target body features respectively;
获取当前目标人体特征所属行人运动轨迹的开始时间begi和结束时间endi，获取每个所述其他目标人体特征所属行人运动轨迹的开始时间begij和结束时间endij；Obtain the start time begi and the end time endi of the pedestrian motion track to which the current target human body feature belongs, and obtain the start time begij and end time endij of the pedestrian motion track to which each of the other target body features belongs;
针对每个所述其他目标人体特征，若所述其他目标人体特征对应的特征相似度大于相似度阈值并且所述其他目标人体特征对应的开始时间begij和结束时间endij满足begij＜endi+tw且endij＜begi-tw的条件，则将所述其他目标人体特征作为当前目标人体特征的相似人体特征，其中，tw为预设的时间窗口。For each of the other target human body features, if the feature similarity corresponding to the other target human body features is greater than the similarity threshold and the start time begij and end time endij corresponding to the other target human body features satisfy begij <endi Under the condition of +tw and endij <begi-tw , the other target human body features are used as similar human body features of the current target human body feature, where tw is a preset time window.
在第二方面，本发明提供一种行人运动轨迹确定系统，所述系统包括：In a second aspect, the present invention provides a pedestrian motion trajectory determination system, the system comprising:
行人运动轨迹获取模块，其被配置为获取位于不同图像采集设备的图像采集范围内的行人运动轨迹；a pedestrian motion trajectory acquisition module, which is configured to acquire pedestrian motion trajectories located within image acquisition ranges of different image acquisition devices;
关系特征获取模块，其被配置为获取不同图像采集设备之间的关系特征；a relationship feature acquisition module configured to acquire relationship features between different image acquisition devices;
人体特征信息获取模块，其被配置为获取每个行人运动轨迹各自对应的人体特征信息；a human body feature information acquisition module, which is configured to obtain the human body feature information corresponding to each pedestrian movement trajectory;
最终行人运动轨迹生成模块，其被配置为根据所述人体特征信息和所述关系特征，对属于同一行人的行人运动轨迹进行关联，生成对应各行人最终的行人运动轨迹。The final pedestrian motion trajectory generation module is configured to associate pedestrian motion trajectories belonging to the same pedestrian according to the human body feature information and the relationship feature, and generate final pedestrian motion trajectories corresponding to each pedestrian.
在第三方面，提供一种控制装置，该控制装置包括处理器和存储装置，所述存储装置适于存储多条程序代码，所述程序代码适于由所述处理器加载并运行以执行上述行人运动轨迹确定方法的技术方案中任一项技术方案所述的行人运动轨迹确定方法。In a third aspect, a control device is provided, the control device comprising a processor and a storage device, the storage device being adapted to store a plurality of pieces of program code, the program code being adapted to be loaded and run by the processor to execute the above The pedestrian movement trajectory determination method described in any one of the technical solutions of the pedestrian movement trajectory determination method.
在第四方面，提供一种计算机可读存储介质，该计算机可读存储介质其中存储有多条程序代码，所述程序代码适于由处理器加载并运行以执行上述行人运动轨迹确定方法的技术方案中任一项技术方案所述的行人运动轨迹确定方法。In a fourth aspect, a computer-readable storage medium is provided, the computer-readable storage medium stores a plurality of program codes therein, and the program codes are adapted to be loaded and executed by a processor to execute the technique of the above-mentioned method for determining the motion trajectory of a pedestrian The pedestrian motion trajectory determination method described in any one of the technical solutions in the solution.
在采用上述技术方案的情况下，本发明能够获取位于不同图像采集设备的图像采集范围内的行人运动轨迹各自对应的人体特征信息，根据不同图像采集设备之间的关系特征以及行人运动轨迹对应的人体特征信息，将属于同一行人的行人运动轨迹进行关联，以获得对应各行人最终的行人运动轨迹。通过上述配置方式，本发明综合考虑了图像采集设备之间的关系特征以及行人运动轨迹对应的人体特征信息对行人运动轨迹关联的影响，防止人体特征信息相似的不同行人的行人运动轨迹的关联，增强了人体特征信息存在差异的相同行人的行人运动轨迹的关联，使得生成的对应各行人最终的行人运动轨迹更为准确。In the case of adopting the above technical solution, the present invention can obtain the human body feature information corresponding to the pedestrian motion trajectories located in the image acquisition ranges of different image acquisition devices. The human body feature information is used to associate the pedestrian motion trajectories belonging to the same pedestrian to obtain the final pedestrian motion trajectories corresponding to each pedestrian. Through the above configuration, the present invention comprehensively considers the relationship between the image acquisition devices and the influence of the human body feature information corresponding to the pedestrian movement track on the association of the pedestrian movement track, so as to prevent the association of the pedestrian movement tracks of different pedestrians with similar human body feature information. The association of pedestrian motion trajectories of the same pedestrians with differences in human feature information is enhanced, so that the generated final pedestrian motion trajectories corresponding to each pedestrian are more accurate.
附图说明Description of drawings
参照附图，本发明的公开内容将变得更易理解。本领域技术人员容易理解的是：这些附图仅仅用于说明的目的，而并非意在对本发明的保护范围组成限制。其中：The disclosure of the present invention will become more easily understood with reference to the accompanying drawings. It is easily understood by those skilled in the art that these drawings are only used for the purpose of illustration and are not intended to limit the protection scope of the present invention. in:
图1是根据本发明的一个实施例的行人运动轨迹确定方法的主要步骤流程示意图；1 is a schematic flowchart of the main steps of a method for determining a pedestrian motion trajectory according to an embodiment of the present invention;
图2是根据本发明实施例的一个实施方式的行人运动轨迹确定方法的主要步骤流程示意图；FIG. 2 is a schematic flowchart of main steps of a method for determining a pedestrian motion trajectory according to an embodiment of the present invention;
图3是根据本发明的一个实施例的行人运动轨迹确定系统的主要结构框图；3 is a main structural block diagram of a pedestrian motion trajectory determination system according to an embodiment of the present invention;
图4是根据本发明实施例的一个实施方式的行人运动轨迹确定系统的主要结构框图。FIG. 4 is a main structural block diagram of a pedestrian motion trajectory determination system according to an embodiment of the present invention.
具体实施方式Detailed ways
下面参照附图来描述本发明的一些实施方式。本领域技术人员应当理解的是，这些实施方式仅仅用于解释本发明的技术原理，并非旨在限制本发明的保护范围。Some embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principle of the present invention, and are not intended to limit the protection scope of the present invention.
在本发明的描述中，“模块”、“处理器”可以包括硬件、软件或者两者的组合。一个模块可以包括硬件电路，各种合适的感应器，通信端口，存储器，也可以包括软件部分，比如程序代码，也可以是软件和硬件的组合。处理器可以是中央处理器、微处理器、图像处理器、数字信号处理器或者其他任何合适的处理器。处理器具有数据和/或信号处理功能。处理器可以以软件方式实现、硬件方式实现或者二者结合方式实现。非暂时性的计算机可读存储介质包括任何合适的可存储程序代码的介质，比如磁碟、硬盘、光碟、闪存、只读存储器、随机存取存储器等等。术语“A和/或B”表示所有可能的A与B的组合，比如只是A、只是B或者A和B。术语“至少一个A或B”或者“A和B中的至少一个”含义与“A和/或B”类似，可以包括只是A、只是B或者A和B。单数形式的术语“一个”、“这个”也可以包含复数形式。In the description of the present invention, "module" and "processor" may include hardware, software or a combination of both. A module may include hardware circuits, various suitable sensors, communication ports, memory, and may also include software parts, such as program codes, or a combination of software and hardware. The processor may be a central processing unit, a microprocessor, an image processor, a digital signal processor, or any other suitable processor. The processor has data and/or signal processing functions. The processor may be implemented in software, hardware, or a combination of the two. Non-transitory computer-readable storage media include any suitable media that can store program code, such as magnetic disks, hard disks, optical disks, flash memory, read-only memory, random-access memory, and the like. The term "A and/or B" means all possible combinations of A and B, such as just A, just B, or A and B. The terms "at least one A or B" or "at least one of A and B" have a similar meaning to "A and/or B" and can include just A, only B, or A and B. The terms "a" and "the" in the singular may also include the plural.
参阅附图1，图1是根据本发明的一个实施例的行人运动轨迹确定方法的主要步骤流程示意图。如图1所示，本发明实施例中的行人运动轨迹确定方法主要包括下列步骤S101-步骤S104。Referring to FIG. 1 , FIG. 1 is a schematic flowchart of the main steps of a method for determining a pedestrian movement trajectory according to an embodiment of the present invention. As shown in FIG. 1 , the method for determining a pedestrian movement trajectory in the embodiment of the present invention mainly includes the following steps S101-S104.
步骤S101：获取位于不同图像采集设备的图像采集范围内的行人运动轨迹。Step S101: Acquire pedestrian movement trajectories located within the image acquisition range of different image acquisition devices.
在本实施例中，可以从不同图像采集设备的图像采集范围内获取行人运动轨迹。In this embodiment, the pedestrian movement trajectories can be acquired from image acquisition ranges of different image acquisition devices.
一个实施方式中，可以使用多目标跟踪算法获取图像采集设备的图像采集范围内的多个行人的行人运动轨迹。多目标跟踪算法包括但不限于DeepSort(Simple Online andRealtime Tracking with a Deep Association Metric，基于深度关联度量的简单在线实时跟踪)等。In one embodiment, a multi-target tracking algorithm can be used to acquire the pedestrian movement trajectories of multiple pedestrians within the image acquisition range of the image acquisition device. The multi-target tracking algorithm includes, but is not limited to, DeepSort (Simple Online and Realtime Tracking with a Deep Association Metric, simple online real-time tracking based on a deep association metric) and the like.
步骤S102：获取不同图像采集设备之间的关系特征。Step S102: Acquire relationship characteristics between different image capturing devices.
在本实施例中，可以获取不同图像采集设备之间的关系特征。关系特征包括但不限于时间关系特征、空间关系特征、时间关系特征和空间关系特征之间组合获得的关系特征等。In this embodiment, the relationship characteristics between different image capturing devices can be acquired. The relationship features include, but are not limited to, temporal relationship features, spatial relationship features, relationship features obtained by combining temporal relationship features and spatial relationship features, and the like.
一个实施方式中，关系特征可以包括不同图像采集设备之间的空间关系特征和不同图像采集设备之间的时间关系特征，其中，两个图像采集设备之间的空间关系特征用于表示在所有图像采集设备相互关联形成的拓扑结构中两个图像采集设备之间的关联权重，两个图像采集设备之间的时间关系特征用于表示行人花费不同时长由一个图像采集设备的图像采集范围运动至另一个图像采集设备的图像采集范围的概率。两个图像采集设备之间的空间关系特征可以用来表征在所有图像采集设备相互关联形成的拓扑结构中两个图像采集设备之间的关联权重，关联权重能够代表两个图像采集设备之间的关联程度。如一个行人在两个图像采集设备之间频繁出现，则可以认为这两个图像采集设备之间的关联权重较高。两个图像采集设备之间的时间关系特征可以用来表征行人花费不同时长由一个图像采集设备的图像采集范围运动至另一个图像采集设备的图像采集范围的概率。根据图像采集设备之间的拓扑结构，可以确认两个图像采集设备之间的距离关系，那么当行人花费一个较短的时长从一个图像采集设备的图像采集范围内运动到另一个距离较远的图像采集设备的图像采集范围内，这种情况的概率是较低的。In one embodiment, the relationship feature may include a spatial relationship feature between different image capturing devices and a temporal relationship feature between different image capturing devices, wherein the spatial relationship feature between two image capturing devices is used to represent all images The weight of the association between two image acquisition devices in the topology structure formed by the correlation of acquisition devices, and the time relationship feature between the two image acquisition devices is used to indicate that pedestrians spend different time moving from the image acquisition range of one image acquisition device to another. The probability of the image acquisition range of an image acquisition device. The spatial relationship feature between two image acquisition devices can be used to represent the correlation weight between the two image acquisition devices in the topology structure formed by the correlation of all image acquisition devices, and the correlation weight can represent the relationship between the two image acquisition devices. degree of association. If a pedestrian appears frequently between two image capturing devices, it can be considered that the association weight between the two image capturing devices is high. The temporal relationship feature between two image capture devices can be used to characterize the probability that pedestrians spend different time periods moving from the image capture range of one image capture device to the image capture range of another image capture device. According to the topology between the image capture devices, the distance relationship between the two image capture devices can be confirmed. Then, when a pedestrian spends a short time moving from the image capture range of one image capture device to another farther away Within the image acquisition range of the image acquisition device, the probability of this situation is low.
一个实施方式中，可以读取大量未进行行人运动轨迹标注的位于不同图像采集设备的图像采集范围内的行人运动轨迹数据，应用这些行人运动轨迹数据对预设的关系特征模型的训练。使用训练好的关系特征模型来获取不同图像采集设备之间的关系特征。In one embodiment, a large number of pedestrian motion trajectory data within the image acquisition range of different image acquisition devices without pedestrian motion trajectory annotation can be read, and the pedestrian motion trajectory data can be used to train a preset relational feature model. Use the trained relational feature model to obtain relational features between different image acquisition devices.
步骤S103：获取每个行人运动轨迹各自对应的人体特征信息。Step S103: Acquire the human body feature information corresponding to each pedestrian movement track.
在本实施例中，可以根据每个行人运动轨迹对应的图像采集设备获取的抓拍图像，来获取行人运动轨迹对应的人体特征信息。In this embodiment, the human body feature information corresponding to the pedestrian movement trajectory may be acquired according to the snapshot image obtained by the image acquisition device corresponding to each pedestrian movement trajectory.
一个实施方式中，人体特征信息可以为人体特征向量，可以对图像采集设备获取的抓拍图像进行特征提取，获取人体特征向量。In one embodiment, the human body feature information may be a human body feature vector, and feature extraction may be performed on a captured image obtained by an image acquisition device to obtain a human body feature vector.
一个实施方式中，人体特征信息可以为人体特征序列，即，根据行人运动轨迹的对应的多张抓拍图像，提取每张抓拍图像的人体特征向量，将多张抓拍图像的人体特征向量的集合作为人体特征序列。人体特征信息也可以为人体特征序列中的一个人体特征向量，如在人体特征序列中任意获取其中一个人体特征向量作为人体特征信息，或者取人体特征序列中包含多个人体特征向量的平均向量作为人体特征信息。In one embodiment, the human body feature information may be a sequence of human body features, that is, according to the corresponding multiple snapshot images of the pedestrian movement trajectory, extract the human body feature vector of each snapshot image, and use the collection of the human body feature vectors of the multiple snapshot images as the set of human body feature vectors. Human feature sequence. The human body feature information can also be a human body feature vector in the human body feature sequence, for example, one of the human body feature vectors in the human body feature sequence is arbitrarily obtained as the human body feature information, or the average vector containing multiple human body feature vectors in the human body feature sequence is taken as the Human characteristics information.
步骤S104：根据人体特征信息和关系特征，对属于同一行人的行人运动轨迹进行关联，生成对应各行人最终的行人运动轨迹。Step S104: Correlate pedestrian movement trajectories belonging to the same pedestrian according to the human body feature information and relationship characteristics, and generate final pedestrian movement trajectories corresponding to each pedestrian.
在本实施例中，可以根据行人运动轨迹对应的人体特征信息，图像采集设备的关系特征来综合判断不同的行人运动轨迹是否属于同一行人，并将属于同一行人的行人运动轨迹进行关联，以生成对应各行人最终的行人运动轨迹。In this embodiment, it is possible to comprehensively judge whether different pedestrian motion trajectories belong to the same pedestrian according to the human body feature information corresponding to the pedestrian motion trajectories and the relationship characteristics of the image acquisition device, and to associate the pedestrian motion trajectories belonging to the same pedestrian to generate Corresponding to the final pedestrian motion trajectory of each pedestrian.
基于上述步骤S101-步骤S104，本发明实施例能够获取位于不同图像采集设备的图像采集范围内的行人运动轨迹各自对应的人体特征信息，根据不同图像采集设备之间的关系特征以及行人运动轨迹对应的人体特征信息，将属于同一行人的行人运动轨迹进行关联，以获得对应各行人最终的行人运动轨迹。通过上述配置方式，本发明实施例综合考虑了图像采集设备之间的关系特征以及行人运动轨迹对应的人体特征信息对行人运动轨迹关联的影响，防止人体特征信息相似的不同行人的行人运动轨迹的关联，增强了人体特征信息存在差异的相同行人的行人运动轨迹的关联，使得生成的对应各行人最终的行人运动轨迹更为准确。Based on the above steps S101-S104, the embodiment of the present invention can obtain the human body feature information corresponding to the pedestrian motion trajectories located in the image acquisition ranges of different image acquisition devices. The human body characteristic information of , associates the pedestrian motion trajectories belonging to the same pedestrian to obtain the final pedestrian motion trajectories corresponding to each pedestrian. Through the above configuration, the embodiment of the present invention comprehensively considers the relationship between the image acquisition devices and the influence of the human body feature information corresponding to the pedestrian movement track on the pedestrian movement track association, so as to prevent the pedestrian movement track of different pedestrians with similar human body feature information. The association enhances the association of the pedestrian motion trajectories of the same pedestrians with differences in human feature information, so that the generated final pedestrian motion trajectories corresponding to each pedestrian are more accurate.
下面对步骤S102和步骤S104作进一步地说明。Steps S102 and S104 are further described below.
在本发明实施例的一个实施方式中，步骤S102可以进一步包括以下步骤S1021至步骤S1025：In one implementation of the embodiment of the present invention, step S102 may further include the following steps S1021 to S1025:
步骤S1021：获取通过每个图像采集设备采集到的行人图像得到的人体特征样本，其中，人体特征样本包括根据行人图像提取到的人体特征，采集行人图像的图像采集设备的设备编号。Step S1021: Acquire a human body feature sample obtained from a pedestrian image collected by each image acquisition device, wherein the human body feature sample includes the device number of the image acquisition device that collects the pedestrian image according to the human body feature extracted from the pedestrian image.
在本实施方式中，可以对每个图像采集设备采集到的行人图像进行特征提取，获取多个人体特征样本。人体特征样本的集合可以表示为X＝{x1，x2，…，xn}，对应的图像采集设备的设备编号的集合可以表示为C＝{c1，c2，…，cn}，其中，n为人体特征样本的数量。In this embodiment, feature extraction may be performed on the pedestrian images collected by each image collection device to obtain a plurality of human body feature samples. The set of human feature samples can be expressed as X={x1 , x2 ,..., xn }, and the set of device numbers of the corresponding image acquisition devices can be expressed as C={c1 , c2 ,..., cn } , where n is the number of human feature samples.
步骤S1022：针对每个人体特征样本，从所有人体特征样本中获取与当前人体特征样本的人体特征相似的相似人体特征样本，并将当前人体特征样本分别与每个相似人体特征样本组成一个匹配样本对。Step S1022: For each human body feature sample, obtain similar human body feature samples from all the human body feature samples that are similar to the human body features of the current human body feature sample, and form a matching sample with the current human body feature sample and each similar human body feature sample respectively. right.
一个实施方式中，可以检索人体特征样本topk近邻人体特征样本。可以将每个人体特征样本xi对应的topk近邻表示为n(xi)＝{Xi1，xi2，…，Xik}。检索过程中使用算法包括但不限于暴力检索、IVFPQ算法(Inverted File System Product Quantization，倒排系统乘积量化)、HNSW算法(Hierarchcal Navigable Small World graphs，分层可导航小世界图)等。In one embodiment, the human feature samples topk nearest neighbor human feature samples can be retrieved. The topk nearest neighbors corresponding to each human feature samplexi can be expressed as n(xi )={Xi1 , xi2 , . . . , Xik }. Algorithms used in the retrieval process include but are not limited to brute force retrieval, IVFPQ algorithm (Inverted File System Product Quantization), HNSW algorithm (Hierarchcal Navigable Small World graphs, Hierarchical Navigable Small World graphs), etc.
一个实施方式中，根据以下步骤S10221至步骤S10223，获取与当前人体特征样本的人体特征相似的人体特征样本，其中，步骤S10221至步骤S10223中的目标人体特征为人体特征样本：In one embodiment, according to the following steps S10221 to S10223, a human body feature sample similar to the human body feature of the current human body feature sample is obtained, wherein the target human body feature in steps S10221 to S10223 is a human body feature sample:
步骤S10221：分别获取当前目标人体特征与其他目标人体特征之间的特征相似度；Step S10221: respectively acquiring the feature similarity between the current target human body feature and other target human body features;
步骤S10222：获取当前目标人体特征所属行人运动轨迹的开始时间begi和结束时间endi，获取每个其他目标人体特征所属行人运动轨迹的开始时间begij和结束时间endij；Step S10222: Obtain the start time begi and the end time endi of the pedestrian motion track to which the current target human body feature belongs, and obtain the start time begij and end time endij of the pedestrian motion track to which each other target body feature belongs;
步骤S10223：针对每个其他目标人体特征，若其他目标人体特征对应的特征相似度大于相似度阈值并且其他目标人体特征对应的开始时间begij和结束时间endij满足begij＜endi+tw且endij＜begi-tw的条件，则将其他目标人体特征作为当前目标人体特征的相似人体特征，其中，tw为预设的时间窗口。Step S10223: For each other target human body feature, if the feature similarity corresponding to the other target human body features is greater than the similarity threshold and the start time begij and end time endij corresponding to the other target human body features satisfy begij <endi +tw and Under the condition of endij <begi-tw , other target human body features are used as similar human body features of the current target human body feature, where tw is a preset time window.
在本实施方式中，目标人体特征对应的特征相似度可以为当前目标人体特征和其他目标人体特征对应的两个人体特征向量之间的向量内积。本领域技术人员可以根据实际应用的需要设置相似度阈值的取值。In this embodiment, the feature similarity corresponding to the target human body feature may be the vector inner product between the current target human body feature and two human body feature vectors corresponding to other target human body features. Those skilled in the art can set the value of the similarity threshold according to the needs of practical applications.
一个实施方式中，针对每个其他目标人体特征，当其他目标人体特征满足以下条件，则可以将其他目标人体特征作为当前目标人体特征的相似人体特征：In one embodiment, for each other target human body feature, when the other target human body feature satisfies the following conditions, the other target human body feature can be used as a similar human body feature of the current target human body feature:
(1)其他目标人体特征对应的特征相似度大于相似度阈值；(1) The feature similarity corresponding to other target human features is greater than the similarity threshold;
(2)当前目标人体特征对应的行人运动轨迹的获取时间和其他人体特征对应的行人运动轨迹的获取时间之间的最大时间间隔小于预设的第一轨迹时间间隔阈值。(2) The maximum time interval between the acquisition time of the pedestrian motion trajectory corresponding to the current target human body feature and the acquisition time of the pedestrian motion trajectory corresponding to other human body features is less than the preset first trajectory time interval threshold.
本领域技术人员可以根据实际应用的需要设置第一轨迹时间间隔阈值的取值。Those skilled in the art can set the value of the first trajectory time interval threshold according to the needs of practical applications.
一个实施方式中，针对每个其他目标人体特征，当其他目标人体特征满足以下条件，则将可以其他目标人体特征作为当前目标人体特征的相似人体特征：In one embodiment, for each other target human body feature, when the other target human body feature satisfies the following conditions, the other target human body feature can be used as a similar human body feature of the current target human body feature:
(1)其他目标人体特征对应的特征相似度大于相似度阈值；(1) The feature similarity corresponding to other target human features is greater than the similarity threshold;
(2)当前目标人体特征对应的行人运动轨迹的获取时间的中间值和其他人体特征对应的行人运动轨迹的获取时间的中间值之间的中间值差值小于预设的第二轨迹时间间隔阈值。(2) The difference between the median value of the acquisition time of the pedestrian motion trajectory corresponding to the current target human body feature and the median value of the acquisition time of the pedestrian motion trajectory corresponding to other human body features is less than the preset second trajectory time interval threshold .
本领域技术人员可以根据实际应用的需要设置第二轨迹时间间隔阈值的取值。Those skilled in the art can set the value of the second trajectory time interval threshold value according to the needs of practical applications.
步骤S1023：针对每个匹配样本对，获取当前匹配样本对中的两个人体特征样本xi和xj各自对应的图像采集设备的设备编号ci和cj，根据设备编号ci和cj建立图像采集设备对(ci，cj)；获取当前匹配样本对(xi，xj)中的两个人体特征样本xi和xj各自对应的行人图像的采集时刻tsi与tsj，确定两个采集时刻tsi与tsj之间的时间间隔|tsi-tsj|，将时间间隔|tsi-tsj|添加到当前图像采集设备对(ci，cj)的时间间隔列表中。Step S1023: For each matching sample pair, obtain the device numbers ci and cj of the image acquisition devices corresponding to the two human body feature samples xi and xj in the current matching sample pair, according to the device numbers ci and cj Establish an image acquisition device pair (ci , c j) ; obtain the acquisition times tsi and tsj of the pedestrian images corresponding to the two human feature samplesxi and xj in the current matching sample pair (xi , xj ) , determine the time interval |tsi -tsj | between the two acquisition moments tsi and tsj , and add the time interval |tsi -tsj | to the time of the current image acquisition device pair (ci , c j) in the interval list.
在本实施方式中，可以将匹配样本对(xi，xj)对应的图像采集设备的设备编号ci和cj，建立图像采集设备对(ci，cj)，并将xi和xj对应的行人图像的采集时间的时间间隔添加至图像采集设备对(ci，cj)的时间间隔列表中。In this embodiment, the device numbers ci and cj of the image acquisition devices corresponding to the matched sample pair (xi , xj ) can be used to establish the image acquisition device pair (ci , c j) , and the xi and the The time interval of the acquisition time of the pedestrian image corresponding to xj is added to the time interval list of the image acquisition device pair (ci , c j) .
步骤S1024：针对每个图像采集设备对，根据预设的分箱数对当前图像采集设备对(ci，cj)的时间间隔列表进行分箱处理，根据分箱处理后落入每个分箱中的时间间隔的数量建立当前图像采集设备对(ci，cj)的直方图，其中，直方图的横坐标表示行人由设备编号为ci的图像采集设备的图像采集范围运动至设备编号为cj的图像采集设备的图像采集范围花费的时间，纵坐标表示花费时间的概率。Step S1024: For each image acquisition device pair, perform binning processing on the time interval list of the current image acquisition device pair (ci , cj ) according to the preset binning number, and fall into each bin according to the binning process. The number of time intervals in the bins establishes a histogram of the current image acquisition device pair (ci ,cj ), where the abscissa of the histogram represents the movement of pedestrians from the image acquisition range of the image acquisition device with device numberci to the device The time spent in the image acquisition range of the image acquisition device numbered cj , the ordinate represents the probability of the time spent.
在本实施方式中，步骤S1024可以进一步包括以下步骤S10241至步骤S10243：In this embodiment, step S1024 may further include the following steps S10241 to S10243:
步骤S10241：对分箱处理后落入每个分箱中的时间间隔的数量进行高斯平滑，获得高斯平滑后的每个分箱中的时间间隔的数量；Step S10241: Perform Gaussian smoothing on the number of time intervals that fall into each bin after the binning process, and obtain the number of time intervals in each bin after the Gaussian smoothing;
步骤S10242：对高斯平滑后的每个分箱中的时间间隔的数量进行归一化，获得每个分箱对应的归一化结果；Step S10242: Normalize the number of time intervals in each bin after Gaussian smoothing, and obtain a normalization result corresponding to each bin;
步骤S10243：将每个分箱对应的数值作为横坐标，将每个分箱对应的归一化结果作为纵坐标，建立当前图像采集设备对(ci，cj)的直方图。StepS10243 : Taking the value corresponding to each bin as the abscissa, and taking the normalization result corresponding to each bin as the ordinate, establish a histogram of the current image acquisition device pair (ci,cj ).
在本实施方式中，高斯平滑的目的是避免各个分箱中时间间隔的数量过于离散。归一化的目的是使得所有分箱对应的归一化结果的总和为1。可以使用最大值最小值归一化方法对高斯平滑后的每个分箱中的时间间隔的数量进行归一化。本领域技术人员可以根据实际应用的需要设置预设的分箱数的取值。In this embodiment, the purpose of Gaussian smoothing is to prevent the number of time intervals in each bin from being too discrete. The purpose of normalization is to make the sum of the normalized results corresponding to all bins equal to 1. The number of time intervals in each bin after Gaussian smoothing can be normalized using the max-min normalization method. Those skilled in the art can set the value of the preset number of bins according to the needs of practical applications.
一个实施方式中，对高斯平滑后的每个分箱中的时间间隔的数量进行归一化的过程可以使用0均值标准化(Z-score standardization)方法，也可以使用概率归一化方法。In one embodiment, the process of normalizing the number of time intervals in each bin after Gaussian smoothing may use a Z-score standardization method or a probability normalization method.
步骤S1025：根据每个图像采集设备对的直方图，确定每个图像采集设备对中两个图像采集设备之间的空间关系特征和时间关系特征。Step S1025: According to the histogram of each image capturing device pair, determine the spatial relationship feature and the temporal relationship feature between the two image capturing devices in each image capturing device pair.
在本实施方式中，步骤S1025可以进一步包括以下步骤S10251至步骤S10252：In this embodiment, step S1025 may further include the following steps S10251 to S10252:
步骤S10251：根据图像采集设备对的直方图的纵坐标确定图像采集设备对中两个图像采集设备之间的时间关系特征。Step S10251: Determine, according to the ordinate of the histogram of the pair of image capture devices, a time relationship feature between two image capture devices in the pair of image capture devices.
步骤S10252：根据以下公式(1)确定两个图像采集设备之间的空间关系特征：Step S10252: Determine the spatial relationship feature between the two image acquisition devices according to the following formula (1):
其中，kij指匹配样本对中的两个人体特征样本xi和xj各自对应的图像采集设备ci和cj之间的空间关系特征，nij指图像采集设备对(ci，cj)的时间间隔列表中时间间隔的总数，
在本发明实施例的一个实施方式中，步骤S104可以进一步包括以下步骤S1041至步骤S1044：In an implementation manner of the embodiment of the present invention, step S104 may further include the following steps S1041 to S1044:
步骤S1041：根据人体特征信息确定不同行人运动轨迹之间的特征相似度。Step S1041: Determine the feature similarity between different pedestrian movement trajectories according to the human body feature information.
一个实施方式中，可以根据人体特征信息确认人体特征向量，根据不同行人轨迹对应的人体特征向量，获取人体特征向量之间的向量内积，来确定不同行人轨迹之间的特征相似度。In one embodiment, the human body feature vector can be confirmed according to the human body feature information, and the vector inner product between the human body feature vectors can be obtained according to the human body feature vectors corresponding to different pedestrian trajectories to determine the feature similarity between different pedestrian trajectories.
一个实施方式中，可以根据以下公式(2)计算两个行人运动轨迹之间的特征相似度eij：In one embodiment, the feature similarity eij between two pedestrian motion trajectories can be calculated according to the following formula (2):
eij＝eiT·ej (2)eij = eiT · ej (2)
其中，ei表示一个行人运动轨迹的中心向量，ej表示另一个行人运动轨迹的中心向量，T表示中心向量ei的向量转置；Among them, ei represents the center vector of a pedestrian movement trajectory, ej represents the center vector of another pedestrian movement trajectory, and T represents the vector transpose of the center vector ei ;
步骤S1042：根据特征相似度和关系特征，计算不同行人运动轨迹之间的关联分。Step S1042: Calculate the correlation score between different pedestrian movement trajectories according to the feature similarity and the relationship feature.
在本实施方式中，可以根据以下公式(3)计算不同行人轨迹之间的关联分：In this embodiment, the correlation score between different pedestrian trajectories can be calculated according to the following formula (3):
zij＝eij-decay+decay×sij (3)zij =eij -decay+decay×sij (3)
其中，zij表示两个行人运动轨迹之间的关联分，eij表示两个行人运动轨迹之间的特征相似度，sij表示两个行人运动轨迹各自对应的图像采集设备之间的关系特征。Among them, zij represents the correlation score between the two pedestrian motion trajectories, eij represents the feature similarity between the two pedestrian motion trajectories, and sij represents the relationship between the image acquisition devices corresponding to the two pedestrian motion trajectories. .
一个实施方式中，sij可以表示两个行人运动轨迹各自对应的图像采集设备之间的空间关系特征与时间关系特征的乘积。可以根据两个行人运动轨迹对应的人体特征样本采集的时刻之间的时间间隔，查询根据上述步骤S1021至步骤S1024建立的直方图，获取该时间间隔所在的分箱对应的纵坐标，将该纵坐标作为两个行人运动轨迹各自对应的图像采集设备之间的时间关系特征，并根据公式(1)计算两个行人运动轨迹各自对应的图像采集设备之间的空间关系特征，将时间关系特征和空间关系特征相乘就能够获得sij。decay表示预设的衰减因子。In one embodiment, sij may represent the product of the spatial relationship feature and the temporal relationship feature between the image acquisition devices corresponding to each of the two pedestrian movement trajectories. According to the time interval between the moments when the human body feature samples corresponding to the two pedestrian motion trajectories are collected, the histogram established according to the above steps S1021 to S1024 can be queried, and the ordinate corresponding to the bin where the time interval is located can be obtained, and the ordinate can be obtained. The coordinates are used as the temporal relationship feature between the corresponding image acquisition devices of the two pedestrian motion trajectories, and the spatial relationship feature between the corresponding image acquisition devices of the two pedestrian motion trajectories is calculated according to formula (1). The sij can be obtained by multiplying the spatial relationship features. decay represents a preset decay factor.
步骤S1043：若两个行人运动轨迹之间的关联分大于预设的分数阈值，则判定两个行人运动轨迹属于同一行人的行人运动轨迹。Step S1043: If the correlation score between the two pedestrian movement trajectories is greater than the preset score threshold, it is determined that the two pedestrian movement trajectories belong to the pedestrian movement trajectory of the same pedestrian.
步骤S1044：对属于同一行人的行人运动轨迹进行关联生成最终对应各行人的行人运动轨迹。Step S1044: Correlate pedestrian movement trajectories belonging to the same pedestrian to generate pedestrian movement trajectories corresponding to each pedestrian finally.
本领域技术人员可以根据实际应用的具体情况设置预设的分数阈值的取值。Those skilled in the art can set the value of the preset score threshold according to the specific situation of the actual application.
针对属于同一行人的行人运动轨迹，可以按照行人运动轨迹的获取时间进行排序，根据排序后的行人运动轨迹生成对应各行人最终的行人运动轨迹。Pedestrian movement trajectories belonging to the same pedestrian can be sorted according to the acquisition time of the pedestrian movement trajectories, and a final pedestrian movement trajectory corresponding to each pedestrian is generated according to the sorted pedestrian movement trajectories.
在本发明实施例的一个实施方式中，人体特征信息包括由图像采集设备采集到的行人图像提取到的人体特征，行人图像的采集时刻，图像采集设备的设备编号和人体特征所属的行人运动轨迹的轨迹编号，步骤S104还可以包括以下步骤S1045至步骤S1047：In one implementation of the embodiment of the present invention, the human body feature information includes the human body features extracted from the pedestrian images collected by the image acquisition device, the acquisition time of the pedestrian images, the device number of the image acquisition device, and the pedestrian movement trajectory to which the human body features belong. The track number of , step S104 may also include the following steps S1045 to S1047:
步骤S1045：针对每个人体特征信息，从所有人体特征信息中获取与当前人体特征信息的人体特征相似的相似人体特征信息，并将当前人体特征信息分别与每个相似人体特征信息组成一个匹配特征对。Step S1045: For each human body feature information, obtain similar human body feature information that is similar to the human body feature of the current human body feature information from all the human body feature information, and form a matching feature with the current human body feature information and each similar human body feature information respectively. right.
在本实施方式中，可以使用上述步骤S10221至步骤S10223所述的方法获取与当前人体特征信息的人体特征相似的相似人体特征信息，其中，步骤S10221至步骤S10223中的目标人体特征为人体特征信息。In this embodiment, similar human body feature information similar to the human body feature of the current human body feature information can be obtained by using the methods described in the above steps S10221 to S10223, wherein the target human body feature in steps S10221 to S10223 is human body feature information .
步骤S1046：针对每个匹配特征对，判断当前匹配特征对中两个人体特征信息对应的轨迹编号是否相同。Step S1046: For each matching feature pair, determine whether the track numbers corresponding to the two human body feature information in the current matching feature pair are the same.
在本实施方式中，步骤S1046可以进一步包括以下步骤S10461和步骤S10462：In this embodiment, step S1046 may further include the following steps S10461 and S10462:
步骤S10461：若相同，则判定当前匹配特征对中两个人体特征信息属于的行人运动轨迹是同一个行人的行人运动轨迹；Step S10461: If they are the same, then determine that the pedestrian motion trajectories to which the two human body feature information in the current matching feature pair belong are the pedestrian motion trajectories of the same pedestrian;
步骤S10462：若不同，则获取当前匹配特征对中两个人体特征信息对应的设备编号对应的图像采集设备对的直方图，根据直方图获取当前匹配特征对对应的两个图像采集设备之间的空间关系特征；Step S10462: If different, obtain the histogram of the image acquisition device pair corresponding to the device numbers corresponding to the two human body feature information in the current matching feature pair, and obtain the current matching feature pair according to the histogram corresponding to the two image acquisition devices between the two image acquisition devices. Spatial relationship characteristics;
获取由当前匹配特征对中两个人体特征信息对应的采集时刻确定的时间间隔，以时间间隔作为横坐标查询直方图得到相应的纵坐标，将纵坐标作为时间关系特征；Acquire the time interval determined by the acquisition moments corresponding to the two human body feature information in the current matching feature pair, use the time interval as the abscissa to query the histogram to obtain the corresponding ordinate, and use the ordinate as the time relationship feature;
根据人体特征信息、空间关系特征与时间关系特征，判断当前匹配特征对中两个人体特征信息所属的行人运动轨迹是否属于同一个行人的行人运动轨迹；若是，则将两个人体特征信息对应的轨迹编号修改成同一轨迹编号。According to the human body feature information, the spatial relationship feature and the time relationship feature, it is judged whether the pedestrian motion trajectory to which the two human body feature information in the current matching feature pair belong belong to the pedestrian motion trajectory of the same pedestrian; The track number is modified to the same track number.
步骤S1047：按照人体特征信息对应的采集时间由先至后的顺序，对具有相同轨迹编号的人体特征信息进行排列，根据排列后的人体特征信息生成对应各行人最终的行人运动轨迹。Step S1047: Arrange the human body feature information with the same track number according to the first-to-last order of the collection time corresponding to the human body feature information, and generate the final pedestrian motion track corresponding to each pedestrian according to the arranged human body feature information.
一个实施方式中，参阅附图2，图2是根据本发明实施例的一个实施方式的行人运动轨迹确定方法的主要步骤流程示意图行人运动轨迹确定方法可以包括以下步骤S201至步骤S206：In one embodiment, referring to FIG. 2, FIG. 2 is a schematic flowchart of the main steps of a method for determining a pedestrian motion trajectory according to an embodiment of the present invention. The pedestrian motion trajectory determining method may include the following steps S201 to S206:
步骤S201：组成匹配样本对。Step S201: Compose matching sample pairs.
在本实施方中，步骤S201与前述步骤S1021和步骤S1022所述的方法类似，为了描述简单，在此不再赘述。In this embodiment, step S201 is similar to the methods described in the foregoing steps S1021 and S1022, and for the sake of simplicity of description, details are not repeated here.
步骤S202：建立图像采集设备对的直方图。Step S202: Establish a histogram of the image acquisition device pair.
在本实施方中，步骤S202与前述步骤S1023和步骤S1024所述的方法类似，为了描述简单，在此不再赘述。In this embodiment, step S202 is similar to the method described in the foregoing steps S1023 and S1024, and is not repeated here for the sake of simplicity.
步骤S203：获取待处理的行人运动轨迹。Step S203: Acquire the pedestrian motion trajectory to be processed.
在本实施方式中，可以获取需要确认的行人运动轨迹。In this embodiment, the pedestrian movement trajectory that needs to be confirmed can be acquired.
步骤S204：获取与待处理的行人轨迹对应的人体特征信息相似的人体特征信息，组成匹配特征对。Step S204: Obtain human body feature information similar to the human body feature information corresponding to the pedestrian trajectory to be processed, and form a matching feature pair.
在本实施方式中，步骤S204与前述步骤S1045所述的方法类似，为了描述简单，在此不再赘述。In this implementation manner, step S204 is similar to the method described in the foregoing step S1045, and for simplicity of description, details are not repeated here.
步骤S205：根据匹配特征对和直方图，生成最终的行人运动轨迹。Step S205: Generate a final pedestrian movement trajectory according to the matching feature pair and the histogram.
在本实施方式中，步骤S205与前述步骤S1046和步骤S1047所述的方法类似，为了描述简单，在此不再赘述。In this implementation manner, step S205 is similar to the method described in the foregoing steps S1046 and S1047, and is not repeated here for the sake of simplicity.
需要指出的是，尽管上述实施例中将各个步骤按照特定的先后顺序进行了描述，但是本领域技术人员可以理解，为了实现本发明的效果，不同的步骤之间并非必须按照这样的顺序执行，其可以同时(并行)执行或以其他顺序执行，这些变化都在本发明的保护范围之内。It should be pointed out that although the steps in the above embodiments are described in a specific sequence, those skilled in the art can understand that in order to achieve the effect of the present invention, different steps do not necessarily need to be executed in such an order. It may be performed simultaneously (in parallel) or in other sequences, and these variations are within the scope of the present invention.
进一步，本发明还提供了一种行人运动轨迹确定系统。Further, the present invention also provides a pedestrian motion trajectory determination system.
参阅附图3，图3是根据本发明的一个实施例的行人运动轨迹确定系统的主要结构框图。如图3所示，本发明实施例中的行人运动轨迹确定系统可以包括行人运动轨迹获取模块、关系特征获取模块、人体特征信息获取模块和最终行人运动轨迹生成模块。在本实施例中，行人运动轨迹获取模块可以被配置为获取位于不同图像采集设备的图像采集范围内的行人运动轨迹。关系特征获取模块可以被配置为获取不同图像采集设备之间的关系特征。人体特征信息获取模块可以被配置为获取每个行人运动轨迹各自对应的人体特征信息。最终行人运动轨迹生成模块可以被配置为根据人体特征信息和关系特征，对属于同一行人的行人运动轨迹进行关联，生成对应各行人最终的行人运动轨迹。Referring to FIG. 3 , FIG. 3 is a main structural block diagram of a pedestrian motion trajectory determination system according to an embodiment of the present invention. As shown in FIG. 3 , the pedestrian motion trajectory determination system in the embodiment of the present invention may include a pedestrian motion trajectory acquisition module, a relationship feature acquisition module, a human body feature information acquisition module, and a final pedestrian motion trajectory generation module. In this embodiment, the pedestrian motion trajectory acquisition module may be configured to acquire pedestrian motion trajectories located within image acquisition ranges of different image acquisition devices. The relationship feature acquisition module may be configured to acquire relationship features between different image capture devices. The human body feature information acquisition module may be configured to acquire the human body feature information corresponding to each pedestrian movement trajectory. The final pedestrian motion trajectory generation module may be configured to associate pedestrian motion trajectories belonging to the same pedestrian according to the human body feature information and relationship characteristics, and generate final pedestrian motion trajectories corresponding to each pedestrian.
一个实施方式中，参阅附图4，图4是根据本发明实施例的一个实施方式的行人运动轨迹确定系统的主要结构框图。如图4所示，行人轨迹确定系统可以包括时空关系学习模块和轨迹关联模块，时空关系学习模块可以包括匹配样本对子模块和直方图统计子模块，轨迹关联模块可以包括匹配特征对子模块和最终行人轨迹生成子模块。在本实施方式中，匹配样本对子模块可以被配置为获取通过每个图像采集设备采集到的行人图像得到的人体特征样本，针对每个人体特征样本，从所有人体特征样本中获取与当前人体特征样本的人体特征相似的相似人体特征样本，并将当前人体特征样本分别与每个相似人体特征样本组成一个匹配样本对。直方图统计子模块可以被配置为针对每个匹配样本对，获取当前匹配样本对中的两个人体特征样本xi和xj各自对应的图像采集设备的设备编号ci和cj，根据设备编号ci和cj建立图像采集设备对(ci,cj)；获取当前匹配样本对(ci,cj)中的两个人体特征样本xi和xj各自对应的行人图像的采集时刻tsi与tsj，确定两个采集时刻tsi与tsj之间的时间间隔|tsi-tsj|，将时间间隔|tsi-tsj|添加到当前图像采集设备对(ci,cj)的时间间隔列表中，针对每个图像采集设备对，根据预设的分箱数对当前图像采集设备对(ci,cj)的时间间隔列表进行分箱处理，根据分箱处理后落入每个分箱中的时间间隔的数量建立当前图像采集设备对(ci,cj)的直方图。匹配特征对子模块可以被配置为针对每个人体特征信息，从所有人体特征信息中获取与当前人体特征信息的人体特征相似的相似人体特征信息，并将当前人体特征信息分别与每个相似人体特征信息组成一个匹配特征对。最终行人轨迹生成子模块可以被配置为针对每个匹配特征对，判断当前匹配特征对中两个人体特征信息对应的轨迹编号是否相同；若相同，则判定当前匹配特征对中两个人体特征信息属于的行人运动轨迹是同一个行人的行人运动轨迹；若不同，则获取当前匹配特征对中两个人体特征信息对应的设备编号对应的图像采集设备对的直方图，根据直方图获取当前匹配特征对对应的两个图像采集设备之间的空间关系特征；获取由当前匹配特征对中两个人体特征信息对应的采集时刻确定的时间间隔，以时间间隔作为横坐标查询直方图得到相应的纵坐标，将纵坐标作为时间关系特征；根据人体特征信息、空间关系特征与时间关系特征，判断当前匹配特征对中两个人体特征信息所属的行人运动轨迹是否属于同一个行人的行人运动轨迹；若是，则将两个人体特征信息对应的轨迹编号修改成同一轨迹编号；按照人体特征信息对应的采集时间由先至后的顺序，对具有相同轨迹编号的人体特征信息进行排列，根据排列后的人体特征信息生成最终的行人运动轨迹。In one embodiment, referring to FIG. 4 , FIG. 4 is a main structural block diagram of a pedestrian motion trajectory determination system according to an embodiment of the present invention. As shown in FIG. 4 , the pedestrian trajectory determination system may include a spatiotemporal relationship learning module and a trajectory association module, the spatiotemporal relationship learning module may include a matching sample pair submodule and a histogram statistics submodule, and the trajectory association module may include a matching feature pair submodule and The final pedestrian trajectory generation submodule. In this embodiment, the matching sample pair sub-module may be configured to acquire human body feature samples obtained from pedestrian images collected by each image acquisition device, and for each human body feature sample, obtain from all human body feature samples the same as the current human body Similar human body feature samples with similar human body features of the feature samples, and the current human body feature sample and each similar human body feature sample respectively form a matching sample pair. The histogram statistics sub-module may be configured to, for each pair of matching samples, obtain the device numbers c i and c j of the image acquisition devices corresponding to the two human feature samples xi and xj in the current matching sample pair, respectively, according to the device numbers ci and cj . Numbers ci and cj to establish an image acquisition device pair (ci , c j) ; obtain the acquisition of pedestrian images corresponding to the two human feature samples xi and xj in the current matching sample pair (ci , cj ) time tsi and tsj , determine the time interval |tsi -tsj | between the two acquisition moments tsi and tsj , and add the time interval |tsi -tsj | to the current image acquisition device pair (ci , cj ) in the time interval list, for each image acquisition device pair, perform binning processing on the time interval list of the current image acquisition device pair (ci , cj ) according to the preset number of bins, according to the binning process The number of time intervals that fall into each bin after processing builds a histogram of the current image acquisition device pair (ci ,cj ). The matching feature pair sub-module can be configured to obtain similar human body feature information that is similar to the human body feature of the current human body feature information from all the human body feature information for each human body feature information, and compare the current human body feature information with each similar human body respectively. The feature information forms a matching feature pair. The final pedestrian trajectory generation sub-module may be configured to, for each matching feature pair, determine whether the track numbers corresponding to the two human body feature information in the current matching feature pair are the same; if they are the same, determine whether the two human body feature information in the current matching feature pair are the same. The pedestrian movement track belongs to the pedestrian movement track of the same pedestrian; if they are different, obtain the histogram of the image acquisition device pair corresponding to the device number corresponding to the two human body feature information in the current matching feature pair, and obtain the current matching feature according to the histogram. For the spatial relationship features between the corresponding two image acquisition devices; obtain the time interval determined by the acquisition moments corresponding to the two human body feature information in the current matching feature pair, and use the time interval as the abscissa to query the histogram to obtain the corresponding ordinate , taking the ordinate as the temporal relationship feature; according to the human body feature information, spatial relationship feature and time relationship feature, determine whether the pedestrian motion trajectories to which the two human body feature information in the current matching feature pair belong belong to the pedestrian motion trajectory of the same pedestrian; if so, Then, modify the track numbers corresponding to the two human body feature information to the same track number; according to the first-to-last order of the collection time corresponding to the human body feature information, arrange the human body feature information with the same track number, according to the arranged human body features. information to generate the final pedestrian motion trajectory.
上述行人运动轨迹确定系统以用于执行图1和图2所示的行人运动轨迹确定方法实施例，两者的技术原理、所解决的技术问题及产生的技术效果相似，本技术领域技术人员可以清楚地了解到，为了描述的方便和简洁，行人运动轨迹确定系统的具体工作过程及有关说明，可以参考行人运动轨迹确定方法的实施例所描述的内容，此处不再赘述。The above-mentioned pedestrian motion trajectory determination system is used to perform the embodiment of the pedestrian motion trajectory determination method shown in FIG. It is clearly understood that, for the convenience and brevity of description, the specific working process and related description of the pedestrian motion trajectory determination system may refer to the content described in the embodiment of the pedestrian motion trajectory determination method, and will not be repeated here.
本领域技术人员能够理解的是，本发明实现上述一实施例的方法中的全部或部分流程，也可以通过计算机程序来指令相关的硬件来完成，所述的计算机程序可存储于一计算机可读存储介质中，该计算机程序在被处理器执行时，可实现上述各个方法实施例的步骤。其中，所述计算机程序包括计算机程序代码，所述计算机程序代码可以为源代码形式、对象代码形式、可执行文件或某些中间形式等。所述计算机可读介质可以包括：能够携带所述计算机程序代码的任何实体或装置、介质、U盘、移动硬盘、磁碟、光盘、计算机存储器、只读存储器、随机存取存储器、电载波信号、电信信号以及软件分发介质等。需要说明的是，所述计算机可读介质包含的内容可以根据司法管辖区内立法和专利实践的要求进行适当的增减，例如在某些司法管辖区，根据立法和专利实践，计算机可读介质不包括电载波信号和电信信号。Those skilled in the art can understand that all or part of the process in the method for implementing the above-mentioned embodiment of the present invention can also be completed by instructing relevant hardware through a computer program, and the computer program can be stored in a computer-readable In the storage medium, when the computer program is executed by the processor, the steps of the foregoing method embodiments can be implemented. Wherein, the computer program includes computer program code, and the computer program code may be in the form of source code, object code, executable file or some intermediate form, and the like. The computer-readable medium may include: any entity or device capable of carrying the computer program code, medium, U disk, removable hard disk, magnetic disk, optical disk, computer memory, read-only memory, random access memory, electrical carrier signal , telecommunication signals, and software distribution media. It should be noted that the content contained in the computer-readable media may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer-readable media Electric carrier signals and telecommunication signals are not included.
进一步，本发明还提供了一种控制装置。在根据本发明的一个控制装置实施例中，控制装置包括处理器和存储装置，存储装置可以被配置成存储执行上述方法实施例的行人运动轨迹确定方法的程序，处理器可以被配置成用于执行存储装置中的程序，该程序包括但不限于执行上述方法实施例的行人运动轨迹确定方法的程序。为了便于说明，仅示出了与本发明实施例相关的部分，具体技术细节未揭示的，请参照本发明实施例方法部分。该控制装置可以是包括各种电子设备形成的控制装置设备。Further, the present invention also provides a control device. In an embodiment of the control device according to the present invention, the control device includes a processor and a storage device, the storage device may be configured to store a program for executing the pedestrian motion trajectory determination method of the above method embodiment, and the processor may be configured to The program in the storage device is executed, and the program includes, but is not limited to, a program for executing the method for determining the pedestrian movement trajectory of the above method embodiment. For the convenience of description, only the parts related to the embodiments of the present invention are shown, and the specific technical details are not disclosed, please refer to the method part of the embodiments of the present invention. The control device may be a control device device formed including various electronic devices.
进一步，本发明还提供了一种计算机可读存储介质。在根据本发明的一个计算机可读存储介质实施例中，计算机可读存储介质可以被配置成存储执行上述方法实施例的行人运动轨迹确定方法的程序，该程序可以由处理器加载并运行以实现上述行人运动轨迹确定方法。为了便于说明，仅示出了与本发明实施例相关的部分，具体技术细节未揭示的，请参照本发明实施例方法部分。该计算机可读存储介质可以是包括各种电子设备形成的存储装置设备，可选的，本发明实施例中计算机可读存储介质是非暂时性的计算机可读存储介质。Further, the present invention also provides a computer-readable storage medium. In one embodiment of the computer-readable storage medium according to the present invention, the computer-readable storage medium may be configured to store a program for executing the method for determining a pedestrian movement trajectory of the above method embodiment, and the program may be loaded and executed by a processor to implement The above-mentioned pedestrian motion trajectory determination method. For the convenience of description, only the parts related to the embodiments of the present invention are shown, and the specific technical details are not disclosed, please refer to the method part of the embodiments of the present invention. The computer-readable storage medium may be a storage device device formed by including various electronic devices. Optionally, the computer-readable storage medium in this embodiment of the present invention is a non-transitory computer-readable storage medium.
进一步，应该理解的是，由于各个模块的设定仅仅是为了说明本发明的装置的功能单元，这些模块对应的物理器件可以是处理器本身，或者处理器中软件的一部分，硬件的一部分，或者软件和硬件结合的一部分。因此，图中的各个模块的数量仅仅是示意性的。Further, it should be understood that since the setting of each module is only for describing the functional units of the apparatus of the present invention, the physical device corresponding to these modules may be the processor itself, or a part of software in the processor, a part of hardware, or Part of the combination of software and hardware. Therefore, the numbers of the various modules in the figures are merely schematic.
本领域技术人员能够理解的是，可以对装置中的各个模块进行适应性地拆分或合并。对具体模块的这种拆分或合并并不会导致技术方案偏离本发明的原理，因此，拆分或合并之后的技术方案都将落入本发明的保护范围内。Those skilled in the art can understand that, each module in the device can be adaptively split or combined. Such splitting or merging of specific modules will not cause the technical solutions to deviate from the principles of the present invention, and therefore, the technical solutions after splitting or combining will fall within the protection scope of the present invention.
至此，已经结合附图所示的优选实施方式描述了本发明的技术方案，但是，本领域技术人员容易理解的是，本发明的保护范围显然不局限于这些具体实施方式。在不偏离本发明的原理的前提下，本领域技术人员可以对相关技术特征作出等同的更改或替换，这些更改或替换之后的技术方案都将落入本发明的保护范围之内。So far, the technical solutions of the present invention have been described with reference to the preferred embodiments shown in the accompanying drawings, however, those skilled in the art can easily understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.
Claims (12)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210210798.7ACN114613005B (en) | 2022-03-04 | 2022-03-04 | Pedestrian motion trajectory determination method, system, device and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210210798.7ACN114613005B (en) | 2022-03-04 | 2022-03-04 | Pedestrian motion trajectory determination method, system, device and medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114613005Atrue CN114613005A (en) | 2022-06-10 |
| CN114613005B CN114613005B (en) | 2025-06-13 |
Family
ID=81861714
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210210798.7AActiveCN114613005B (en) | 2022-03-04 | 2022-03-04 | Pedestrian motion trajectory determination method, system, device and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114613005B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105224912A (en)* | 2015-08-31 | 2016-01-06 | 电子科技大学 | Based on the video pedestrian detection and tracking method of movable information and Track association |
| WO2018133666A1 (en)* | 2017-01-17 | 2018-07-26 | 腾讯科技（深圳）有限公司 | Method and apparatus for tracking video target |
| EP3739356A1 (en)* | 2019-05-12 | 2020-11-18 | Origin Wireless, Inc. | Method, apparatus, and system for wireless tracking, scanning and monitoring |
| CN113436229A (en)* | 2021-08-26 | 2021-09-24 | 深圳市金大智能创新科技有限公司 | Multi-target cross-camera pedestrian trajectory path generation method |
| US20220066456A1 (en)* | 2016-02-29 | 2022-03-03 | AI Incorporated | Obstacle recognition method for autonomous robots |
- 2022
- 2022-03-04CNCN202210210798.7Apatent/CN114613005B/enactiveActive
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105224912A (en)* | 2015-08-31 | 2016-01-06 | 电子科技大学 | Based on the video pedestrian detection and tracking method of movable information and Track association |
| US20220066456A1 (en)* | 2016-02-29 | 2022-03-03 | AI Incorporated | Obstacle recognition method for autonomous robots |
| WO2018133666A1 (en)* | 2017-01-17 | 2018-07-26 | 腾讯科技（深圳）有限公司 | Method and apparatus for tracking video target |
| EP3739356A1 (en)* | 2019-05-12 | 2020-11-18 | Origin Wireless, Inc. | Method, apparatus, and system for wireless tracking, scanning and monitoring |
| CN113436229A (en)* | 2021-08-26 | 2021-09-24 | 深圳市金大智能创新科技有限公司 | Multi-target cross-camera pedestrian trajectory path generation method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114613005B (en) | 2025-06-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Huang et al. | Self-supervised attentive generative adversarial networks for video anomaly detection | |
| Sun et al. | Deep affinity network for multiple object tracking | |
| CN110175549B (en) | Face image processing method, device, equipment and storage medium | |
| Lin et al. | Single shot temporal action detection | |
| Ding et al. | Violence detection in video by using 3D convolutional neural networks | |
| Shou et al. | Online detection of action start in untrimmed, streaming videos | |
| Zhang et al. | Improving triplet-wise training of convolutional neural network for vehicle re-identification | |
| CN109522814B (en) | A kind of target tracking method and device based on video data | |
| CN110163041A (en) | Video pedestrian recognition methods, device and storage medium again | |
| GB2409028A (en) | Face detection | |
| WO2022160591A1 (en) | Crowd behavior detection method and apparatus, and electronic device, storage medium and computer program product | |
| US11200683B2 (en) | Image processing device and image processing method | |
| WO2022142417A1 (en) | Target tracking method and apparatus, electronic device, and storage medium | |
| CN107153824A (en) | Across video pedestrian recognition methods again based on figure cluster | |
| CN114048344A (en) | Similar face searching method, device, equipment and readable storage medium | |
| CN109800329B (en) | Monitoring method and device | |
| CN112949382A (en) | Camera movement detection method and device, and electronic device | |
| Wong et al. | Multi-camera face detection and recognition in unconstrained environment | |
| Yoshinaga et al. | Background model based on statistical local difference pattern | |
| Parui et al. | An efficient violence detection system from video clips using ConvLSTM and keyframe extraction | |
| Xie et al. | Learning to find good correspondences of multiple objects | |
| Peng et al. | Tracklet siamese network with constrained clustering for multiple object tracking | |
| CN111563479B (en) | Concurrent person weight removing method, partner analyzing method and device and electronic equipment | |
| CN106934339B (en) | Target tracking and tracking target identification feature extraction method and device | |
| CN114613005A (en) | Pedestrian trajectory determination method, system, device and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |