Disclosure of Invention
The application provides a target analysis method and electronic equipment, which can integrate a plurality of target analysis results of different types and meet complex analysis requirements.
In order to achieve the above purpose, the application adopts the following technical scheme:
A method of target analysis, comprising:
inputting image data to be processed into at least two different types of target analysis modules;
Each target analysis module respectively processes the image data to obtain respective target analysis results, wherein each target analysis result comprises at least one target and track information thereof, and at least one target analysis result comprises structural information of the target;
performing track information matching on targets in target analysis results obtained by different types of target analysis modules, and determining a first target corresponding to the successfully matched track information in the corresponding target analysis results;
and outputting a result based on the structural information of the first target.
The target analysis module is preferably a target detection alarm unit and a structured target detection unit, wherein the target detection alarm unit is used for detecting an alarm event and an alarm target related to the alarm event, outputting alarm information for indicating the occurrence of the alarm event and the alarm target and track information thereof, and the structured target detection unit is used for detecting structured information of the target and outputting the structured target, the structured information thereof and the track information thereof;
The matching of track information of the targets in the target analysis results obtained by the target analysis modules of different types comprises the following steps:
And for each alarm target output by the target detection alarm unit, matching the track information of the alarm target with the track information of each structured target output by the structured target detection unit, wherein the structured targets are targets in the target analysis result output by the structured target detection unit.
Preferably, the means for determining whether the match is successful comprises:
and determining the track information of the structural targets matched with the track of the alarm target in the same time period as the track information successfully matched in the track information of all the structural targets based on the time information in the image data.
Preferably, the outputting the result based on the structured information of the first object includes:
performing secondary analysis based on the structural information of the first target, and outputting a result based on a secondary analysis result;
Or directly outputting the structural information of the first target as a result.
Preferably, the performing the secondary analysis based on the structured information of the first object includes:
performing target modeling of the output first target based on the structural information of the first target output by the structural target detection unit;
Searching in a target model library based on the target modeling result to obtain a target searching result of the first target, and storing the target modeling result in the target model library;
and correspondingly outputting each alarm information output by the target detection alarm unit, the associated alarm target and the target retrieval result of the first target.
Preferably, the target detection alarm units are multiple, and alarm events of different target detection alarm units are different;
Matching the alarm target track information output by each indoor target detection alarm unit with the target track information output by the structured target detection unit, and determining a first target successfully matched;
the method further comprises the steps of:
Matching and matching target retrieval results of the first targets corresponding to different target detection alarm units, and outputting the best matched target retrieval results and corresponding alarm information.
Preferably, the target detection alarm unit is an indoor target detection alarm unit, the determined alarm information is a close-range contact alarm, the alarm targets corresponding to the alarm information are two detected targets meeting the close-range contact condition, and the target modeling is a face modeling;
When the track information of the alarm targets is matched with the track information of each structured target, matching is carried out for each of the two targets;
Performing face modeling and retrieval aiming at each of the two output structured targets to obtain face retrieval results of the two output structured targets;
and correspondingly outputting the face retrieval results of each alarm information and the two corresponding output structural targets.
Preferably, the determining, as the track information of the successful matching, the track information of the structured target with the highest matching degree with the track of the alarm target in the same time period includes:
And comparing the track information of the alarm targets with the track information of each structural target, and selecting the track of the structural target matched with the track of the alarm target from the tracks of each structural target based on the curve fitting effect of the track, the coincidence degree of the positions of the tracks in the whole frame of image and the coincidence degree of the timestamp information corresponding to the track points at the same positions.
Preferably, when the target detection alarm unit does not output alarm information, the method further comprises:
selecting one structured object from all structured objects output by the structured object detection unit;
and carrying out target modeling based on the structural information of the selected structured target, and storing the result of the target modeling in the target model library.
An electronic device comprising at least a computer readable storage medium, further comprising a processor;
The processor is configured to read executable instructions from the computer readable storage medium and execute the instructions to implement the target analysis method of any one of the above.
As can be seen from the above technical solution, in the present application, first, image data to be processed is input to at least two different types of target analysis modules, and each target analysis module is utilized to process the image data respectively, so as to obtain respective target analysis results. Thus, the image data can be processed by different types of target analysis modules, so that target analysis results of different angles of the same image data can be obtained. Wherein each target analysis result comprises at least one target and track information thereof, and at least one target analysis result comprises structural information of the target. And then, matching the track information of the targets in the target analysis results obtained by the target analysis modules, and determining the first target corresponding to the successfully matched track information in the target analysis modules. Therefore, the analysis results of different angles can be linked together through track matching, the target analysis results obtained from different angles are integrated, and the target successfully matched is determined. And finally, outputting a result of target analysis based on the structural information of the first target successfully matched. Therefore, integration of a plurality of different types of target analysis results based on track information can be realized, and more complex analysis requirements are met.
Detailed Description
The present application will be described in further detail with reference to the accompanying drawings, in order to make the objects, technical means and advantages of the present application more apparent.
FIG. 1 is a schematic diagram of the basic flow of the target analysis method of the present application. As shown in fig. 1, the method includes:
step 101, inputting image data to be processed into at least two different types of target analysis modules.
The image data to be processed may be single-frame image data or may be a set of image frame data composed of a plurality of image frames, such as a piece of video or animation, or the like. At least two different types of target analysis modules are introduced into the application, and different types of target analysis processing is carried out on the same image data to be processed. The same set of image data is then input to a plurality of target analysis modules of different types.
And 102, each target analysis module processes the image data to obtain respective target analysis results.
Each target analysis result comprises at least one target and track information thereof, and the at least one target analysis result comprises structural information of the target.
Specifically, each target analysis module processes the input image data independently, and internal processing of the specific target analysis module can be performed according to an existing processing mode, and a corresponding target analysis result is obtained. In general, in this step, for a single target analysis module, processing and outputting image data are performed according to an existing manner, so as to obtain a corresponding target analysis result.
And step 103, matching track information of targets in target analysis results obtained by different types of target analysis modules, and determining a first target corresponding to the successfully matched track information in the corresponding target analysis results.
In the application, the target analysis results of different types of target analysis modules obtained in the step 102 are subjected to matching processing of track information. And determining a corresponding target in the corresponding target analysis result according to the successfully matched track information, and enabling the target to be called a first target so as to be distinguished from other targets in the target analysis result.
And 104, outputting a result of target analysis based on the structural information of the first target.
After the first target is obtained, the structured information of the first target may be directly output, or the next stage of processing may be performed based on the structured information of the first target, and then the target analysis result may be output.
Thus, the flow of the target analysis method in the application is ended. Through the flow, different types of target analysis are carried out on the image data, the obtained different target analysis results are matched according to the track information, a first target successfully matched is determined, and finally, result output is carried out based on the structural information of the first target. Therefore, a plurality of target analysis processing at different angles can be integrated, and different target analysis results are integrated from the track information angle, so that the target analysis results successfully matched with the track information are output or processed later, and the more complex target analysis requirements can be met.
For complex target analysis requirements, matching and integration of different target analysis results can be generally performed based on the specific analysis requirements, application scenes and the like, and subsequent processing can be performed based on the specific requirements. For example, one type of common complex requirement may be to detect and alert objects by monitoring images and to obtain detailed information about alert objects. However, for general target analysis processing, a target analysis module generally used for realizing target detection alarm can only realize simple target detection, but cannot acquire detailed information of a target, and a target analysis module used for acquiring detailed information of a target can be, for example, a structured target detection module, and can only usually perform target detection and acquire corresponding structured information, but cannot alarm according to requirements. Therefore, the requirement of performing target alarm and acquiring detailed information of an alarm target cannot be met by utilizing any single target analysis and processing. For such a composite analysis requirement, it is possible to implement the target analysis method of the present application.
Specifically, in order to acquire an alarm target and detailed information of the target, the method can select a target detection alarm unit capable of realizing target detection alarm and a structured target detection unit capable of realizing structured detection of the target as two different types of target analysis modules, wherein the target detection alarm unit can be used for detecting an alarm event and an alarm target associated with the alarm event and outputting alarm information for indicating the occurrence of the alarm event and the alarm target and track information thereof, and the structured target detection unit can be used for detecting the structured information of the target and outputting the target and the structured information and track information thereof.
In more detail, a set of image frames (e.g., a video or a moving picture) may be utilized as image data to be processed, and the corresponding image data may be input to the object detection alarm unit and the structured object detection unit, respectively. The target detection alarm unit can detect and track one or more targets in the image data, and when the targets meet alarm conditions, the target detection alarm unit determines that alarm events are detected and generates alarm information, an output target analysis result comprises the alarm information and targets (hereinafter referred to as alarm targets) and tracks thereof associated with the alarm information, the structured target detection unit can detect and track the targets in the image data, and acquire structured information of the detected targets, such as height, gesture, head information and the like, and the output target analysis result comprises the detected targets (hereinafter referred to as structured targets) and target tracks and target structured information. In order to meet the requirement of acquiring the alarm target and the detailed structural information thereof, the target analysis result of the target detection alarm unit and the target analysis result of the structural target detection unit can be matched according to the track, and the specific matching mode can include:
And in the tracks of all the structured targets, determining the track of the structured target with the best matching degree with the track of the alarm target in the same time period, and taking the determined track of the structured target as the track successfully matched, namely, considering that the determined structured target is consistent with the alarm target, the structured target can be taken as the structured target corresponding to the alarm information and is distinguished from other structured targets, and the structured target corresponding to the alarm information is called as an output structured target hereinafter. Based on the above, the structural information of the structural target successfully matched can be directly output or the structural information can be utilized to output after being subjected to subsequent processing, for example, the structural information can be utilized to carry out target modeling, the structural information can be searched in a target model library based on the result of the target modeling, the target search result of the structural target is output, each alarm information output by the target detection alarm unit and the corresponding target search result of the structural target are correspondingly output, and thus the alarm information and the related target information can be displayed more intuitively. In addition, in order to improve the accuracy of the subsequent target retrieval, the result of target modeling of the structured target (i.e. the target model obtained by the target modeling) can be output and stored in a target model library. The object modeling herein may be modeling of whole or partial members of a structured object, e.g., face modeling, etc.
In addition, the target detection alarm unit can generate a plurality of alarm information, each alarm information is associated with at least one alarm target, the structural information of each alarm target can be obtained according to the mode aiming at each alarm information and the associated alarm target, and the target retrieval result of the alarm target can be determined and output corresponding to the alarm information. The plurality of alarm information can obtain a plurality of corresponding target retrieval results, and each target retrieval result can be correspondingly output with the corresponding alarm information.
The method of analyzing an object in the present application will be described with reference to specific examples. FIG. 2 is a flow chart of a target analysis method according to an embodiment of the application. In the embodiment, the specific analysis requirement is that the indoor personnel are subjected to close-range contact detection alarm, and corresponding alarm targets are determined to perform face retrieval, namely, the target modeling is specifically face modeling, and the target retrieval is specifically face retrieval. As shown in fig. 2, the method includes:
Step 201, inputting a section of monitoring video into an indoor target detection alarm unit and a structured target detection unit respectively.
In this embodiment, a section of monitoring video is used to perform target analysis. The monitoring video is respectively input to two different types of target analysis modules, namely an indoor target detection alarm unit and a structured target detection unit. The indoor target detection alarm unit is a common target detection alarm unit, can identify various targets in a picture through monitoring videos, and generates alarm information when the targets meet alarm conditions, for example, the alarm conditions can usually alarm when people are identified.
Step 202, an indoor target detection alarm unit processes an input monitoring video, generates alarm information after detecting an alarm target, and outputs the alarm information and a track of the alarm target to a post-processing module.
The common indoor target detection alarm unit can detect and track targets of the monitoring video, distinguish different targets appearing in the monitoring video, and track and acquire the moving track of the targets one by one. Generating alarm information when the target meeting the alarm condition is found according to the preset alarm condition, and giving the alarm target and the track of the alarm target. The alarm condition may be, for example, the detection of a person in a picture.
In the embodiment, a new alarm condition is defined based on the detected personnel alarming, namely, if two detected personnel meet the personnel close-range contact alarm condition (for example, the distance between the two personnel is smaller than or equal to a threshold value), alarm information is generated, the two personnel meeting the personnel close-range contact alarm condition are taken as alarm targets, and the alarm information and the track of the two alarm targets are taken as target analysis results to output.
In fact, the personnel proximity alarm can be realized in the indoor target detection alarm unit, or the indoor target detection alarm unit can only realize personnel detection alarm, and the subsequent post-processing unit further realizes further determination of the proximity alarm based on the information of the personnel detection alarm.
And 203, processing the input monitoring video by the structural target detection unit, and outputting the detected targets and the tracks of the targets to a post-processing module.
As previously mentioned, structured target detection units are commonly used to perform target detection, tracking, and structured information acquisition in surveillance videos. Different targets in the monitoring video can be distinguished, the moving track of the targets can be obtained by tracking the targets one by one, and the structural information of the targets, such as height, gesture and the like, can be further obtained. The target detected by the structured target detection unit is referred to as a structured target, and in this embodiment, the structured target detection unit is used for detecting personnel and acquiring the track and the structured information of each detected personnel. The specific manner of detecting and tracking the target and obtaining the structured information can be implemented in various existing manners.
The processing of step 202 and step 203 may be performed in parallel.
In step 204, the post-processing module matches the track of the alarm target corresponding to the alarm information with the track of each structured target output by the structured target detection unit for each alarm information output by the indoor target detection alarm unit, and determines the matched track of the structured target and the corresponding output structured target.
In the step, different target analysis results output by the indoor target detection alarm unit and the structured target detection unit are matched. In the embodiment, when matching is performed, the alarm information is used as a unit to perform matching, and whether the alarm information belongs to the same target is used as a matching standard.
Specifically, the matching is performed in units of alarm information, that is, a set of matching is performed for each alarm information, a corresponding matching result is determined, and the matching processing performed for each alarm information is the same, and the matching performed for the alarm information a will be described below as an example. And when the matching processing is carried out, matching the alarm target of the alarm information A with the structural target detected by the structural target detection unit by taking whether the alarm target belongs to the same target as a matching standard. The specific judging method for the matching is as follows:
For the alarm information A, the track of the alarm target corresponding to the alarm information A is respectively compared with the track of each structured target detected by the structured target detection unit, and the track of the structured target with the best matching degree with the track of the alarm target is found in the tracks of all structured targets, so that the track of the structured target is considered to be matched with the track of the alarm target, namely, the corresponding structured target is matched with the alarm target, and then the corresponding structured target and the alarm target are considered to point to the same target, so that the matching is completed. The matched structured targets are referred to as output structured targets.
It should be noted that, in this embodiment, the alarm information corresponds to two alarm targets, so when performing the matching process, a matching operation needs to be performed for each of the two alarm targets, so as to find an output structured target that corresponds to the matching, and finally find two output structured targets that match the two alarm targets.
In addition, in the matching process, when the track of the alarm target is compared with the track of the structured target to determine the track of the structured target with the best matching degree, besides the matching similarity of the track itself, whether the positions of the tracks in the image are consistent or not and whether the timestamp information corresponding to the track points at the same positions are consistent or not need to be compared, and only when the matching degree of the track, the position similarity of the track in the image and the timestamp information corresponding to the track points at the same position are greater than or equal to the corresponding set threshold (namely, the matching similarity of the track is greater than or equal to the set track similarity threshold, and the similarity of the track points at the same position and the timestamp information is greater than or equal to the time similarity threshold), the corresponding tracks are considered to be matched.
Various existing track comparison methods can be adopted for the fitting similarity of the tracks themselves. In this embodiment, the specific comparison method may be performed based on an open source fitting comparison algorithm, a curve/shape similarity algorithm based on the Frechet distance.
In more detail, let X and Y be the Buchz space, Ω is an open convex subset of X, x0∈Ω, and f: Ω→Y. If F has an n-order F derivative operator F (n) on Ω, the following Taylor formula is established for any h εX, x0+h εΩ:
And is also provided with
If f (n) is still continuous at Ω, then there is
In mathematics, the taylor formula is a formula describing values around a certain point by using information of the function, and is a method of approximating a function f (x) having an n-th derivative at x=x0 by using an n-th polynomial on (x-x 0).
If the function f (x) has an n-th derivative over a certain closed interval [ a, b ] including x0 and an n+1-th derivative over an open interval (a, b), the following formula holds for any point x over the closed interval [ a, b ]:
Where f (x) represents the n-th derivative of f (x), the polynomial after the equal sign is called the Taylor expansion of the function f (x) at x0, and the remainder Rn (x) is the remainder of the Taylor formula, which is the higher order infinite of (x-x 0) n.
Those skilled in the art can know that the similarity between any one of the small radians and the similarity of the tangential angles on the two track curves can be determined based on the above mathematical formula, and the similarity between the two track curves can be calculated based on the similarities, so that the curve fitting degree of the two tracks can be determined. Of course, the method for determining the similarity of curve fitting of two tracks can also be performed in other existing manners, which is not limited by the present application.
The comparison and similarity calculation of the position of the whole track in the image and the timestamp information of the track points at the same position can be realized in the existing manner, and will not be repeated here.
The tracks similar in the fitting degree of the track curve, the positions of the tracks in the image and the timestamp information of the track points at the same positions mean that two targets corresponding to the two tracks have the same moving track at the same positions within the same period of time, so that the two targets can be considered to be the same target in practice. That is, the output structured target obtained by the above-described trajectory comparison method matching is directed to the same target as the alarm target.
And step 205, taking the alarm information, the alarm target and the structural detection result of the output structural target as the optimal target analysis result.
After the output structured target is determined, the structural detection result of the output structured target gives out detailed structural information of the corresponding target, namely the detailed structural information of the alarm target, and in order to meet the requirement of acquiring the detailed information of the alarm target, the alarm information, the structural detection result of the alarm target and the structural detection result of the output structured target are jointly used as an optimal target analysis result for representing the alarm information, the alarm target and the detailed information thereof.
Step 206, based on the structured detection result in the best target analysis result, performing face modeling of the output structured target.
The optimal target analysis result comprises a structured detection result of the output structured target, and is also considered as a structured detection result of the alarm target. Because the face of the alarm target needs to be searched, the face of the alarm target needs to be modeled first, and then the face model generated by modeling is used for searching the face. Therefore, in this step, face modeling of the output structured target, that is, face modeling of the alarm target, needs to be performed.
Specifically, since detailed information of the target, including detailed information of the face, is included in the structured information, face modeling can be performed using the structured information. The specific modeling method may take various existing manners, and will not be described in detail herein.
Step 207, searching in a face database based on the face modeling result to obtain a face searching result of the output structural target, and storing the face modeling result in the face database.
The face model is obtained by performing face modeling in step 206, and the model is compared with the face model stored in the face database to determine the face retrieval result. Meanwhile, in order to enlarge the scale of the face model in the face library and improve the accuracy of the subsequent retrieval of the same face, the face model obtained by face modeling in the step 206 can be stored in the face library. Therefore, when the retrieval of the same face is triggered for the second time, the information of the current target can be detected more efficiently, and the accuracy of the next detection of the same face is improved.
And step 208, outputting each alarm information output by the indoor detection alarm unit and the corresponding alarm target and the face retrieval result of the output structured target correspondingly.
The analysis requirements of this embodiment are that it is desirable to be able to detect alarms for indoor personnel, determine alarm targets, and perform face retrieval of alarm targets. Therefore, when the result is output finally, the alarm information, the alarm target and the face retrieval result of the output structured target are correspondingly output. In addition, as described above, in this embodiment, the alarm targets corresponding to the single alarm information are two closely contacted targets, so when the corresponding output structured targets perform face modeling and face retrieval, two face models are obtained by modeling, and two face retrieval results are obtained by performing face retrieval on the two face models, and correspond to the two alarm targets respectively. Therefore, in this embodiment, the face search results corresponding to the alarm information a include two alarm targets and two output structured targets.
The processing of each alarm message is performed in the manner described above in steps 204-208.
To this end, the flow of the target analysis method in the embodiment of the present application shown in fig. 2 is ended.
In the above embodiment, the specific implementation of the target analysis method is described by taking the target analysis performed by the indoor target detection alarm unit and the structured target detection unit as an example. In fact, for other complex analysis requirements, the method of the present application may also be applied, for example, if the analysis requirement is that a perimeter detection personnel alarms and obtains an alarm target and a face search result thereof, a method similar to that shown in fig. 2 above may be correspondingly adopted, specifically, a perimeter detection alarm unit may be utilized to perform personnel detection alarm to determine a track of the alarm target, for the track of the alarm target corresponding to each alarm information, a track of a matched structured target is selected from the tracks of the structured targets output by the structured target detection unit, the structured target corresponding to the track of the selected structured target is used as an output structured target corresponding to the alarm information, the face modeling and the face search are performed by using the structured detection result of the output structured target, and the face search result is output corresponding to the alarm information and the alarm target. The trajectory comparison matching process may be the same as the method shown in fig. 2. An example of an application for track comparison and matching is given below:
Fig. 3 shows a detection result of a perimeter target detection alarm unit, fig. 3 shows a track line segment 1 of an alarm target obtained from analysis of a section of monitoring video through a tracking rule, fig. 4 shows a plurality of targets detected by a structured target detection unit and corresponding target track line segments 1-5, and the optimized furcher algorithm is utilized to compare the track line segments in fig. 3 with each track line segment in fig. 4, so that track line segments matched with the track line segments in fig. 3 are obtained in all track line segments in fig. 4, structured detection result data of the targets corresponding to the track line segments are sent to face modeling, and subsequent face modeling and face retrieval processing is performed. In more detail, the detection result shown in fig. 4 has 5 track segments, the detection result shown in fig. 3 has 1 track segment, the 5 track segments in the detection result shown in fig. 4 all correspond to the face image and the human body image data of respective targets, the similarity between the track segment 5 in fig. 4 and the track segment in fig. 3 is finally determined to be the highest by comparing the track segment in fig. 3 with each track segment in fig. 4 through the optimized friechet algorithm, and then the face data and the human body data corresponding to the track segment 5 in the detection result in fig. 4 are sent to the modeling comparison algorithm to obtain the attribute of the current face, and face modeling and face retrieval are performed according to the current material data.
In addition, in the above-described flow shown in fig. 2 and the above application example, the processing is performed based on the alarm information output by the object detection alarm unit, and in fact, if the object detection alarm unit does not output the alarm information, the object search may be performed by using the structured object detection unit to expand the object model library. The method comprises the steps of selecting one structured target from all structured targets output by a structured target detection unit, carrying out target modeling based on a structured detection result of the selected structured target, and storing a target modeling result in a target model library.
The foregoing specific embodiment and application example shown in fig. 2 provide two ways of implementing the complex analysis requirements and their specific target analysis. In addition, there may be further other analysis requirements based on multiple complex analysis requirements, specifically, the number of target detection alarm units may be multiple, and alarm events of different target detection alarm units may be different, the alarm target track information output by each target detection alarm unit may be respectively matched with the target track information output by the structured target detection unit, a first target successfully matched may be determined, for each target detection alarm unit, target modeling and target retrieval may be performed on the first target to obtain respective target retrieval results, matching may be performed on target retrieval results of the first targets corresponding to each target retrieval alarm unit, and the best matched target retrieval results and corresponding alarm information may be output.
For example, for an analysis result satisfying the analysis requirement of fig. 2 to obtain an alarm information and a face search result, for an analysis result satisfying the analysis requirement of an application instance to obtain an alarm information and a face search result, it may be necessary to compare the two face search results, thereby implementing further result analysis. For this requirement, the target detection alarm unit may be configured to include an indoor target detection alarm unit and a perimeter detection alarm unit, where the indoor target detection alarm unit may be matched according to the method shown in fig. 2, to determine an output structured target (hereinafter referred to as a first output structured target), the perimeter detection alarm unit may be matched according to the method shown in the application example, to determine a corresponding output structured target (hereinafter referred to as a second output structured target), and the face search result of each first output structured target corresponding to each alarm information output by the indoor target detection alarm unit is matched with the face search result of each second output structured target corresponding to each alarm information output by the perimeter detection alarm unit, and the face search result with the best match and the alarm information corresponding thereto are output or further processed.
The target analysis method can effectively solve the problem that the single target analysis processing cannot meet the complex requirement, can match a plurality of different types of target analysis results through certain logic processing to obtain the optimal target analysis result, and directly outputs or carries out subsequent processing according to the analysis requirement to obtain the final analysis result.
The application also provides a target analysis device which can be used for implementing the target analysis method. FIG. 5 is a schematic diagram of the basic structure of the target analysis device according to the present application, as shown in FIG. 5, the device includes at least two target analysis modules, a post-processing module, and an output module.
Each target analysis module is used for processing the input image data to be processed to obtain target analysis results and outputting the target analysis results to the post-processing module, wherein each target analysis result comprises at least one target and track information thereof, and at least one target analysis result comprises structural information of the target;
The post-processing module is used for matching track information of the targets in the target analysis results obtained by the target analysis modules of different types, and determining a first target corresponding to the successfully matched track information in the corresponding target analysis results;
The output module is used for outputting the result of target analysis based on the structural information of the first target;
and each target analysis module performs different types of target analysis processing.
Alternatively, the image data to be processed may be a set of image frames;
The target analysis module can be a target detection alarm unit and a structured target detection unit respectively;
The target detection alarm unit can be used for processing the image data, detecting an alarm event and an alarm target associated with the alarm event, and outputting alarm information and alarm target and track information for indicating the occurrence of the alarm event;
the structured target detection unit is used for processing the image data, detecting structured information of the target and outputting the structured target, the structured information and track information of the structured target;
In the post-processing module, the track information matching processing of the targets in the target analysis results obtained by the target analysis modules of different types may specifically include:
and for the alarm targets corresponding to each alarm information output by the target detection alarm unit, matching the track of the alarm target with the track of each structured target output by the structured target detection unit, wherein the structured targets are targets in the target analysis result output by the structured target detection unit.
Optionally, the method for determining whether the matching is successful may specifically include:
and determining the track information of the structural targets matched with the track of the alarm target in the same time period as the track information of successful matching in the track information of all the structural targets based on the time information in the image data.
Optionally, the processing of outputting the result based on the structured information of the first target may specifically include:
performing secondary analysis based on the structured information of the first target, and outputting a result based on a secondary analysis result;
Or directly outputting the structural information of the first target as a result.
Optionally, the apparatus may further include a target modeling module and a target retrieval module;
the target modeling module can be used for carrying out target modeling of the output structured target based on the structured detection result of the output structured target;
The target retrieval module can be used for retrieving in a target model library based on the result of target modeling to obtain a target retrieval result of the output structured target, and storing the result of target modeling in the target model library;
In the output module, the processing of performing secondary analysis based on the structured information of the first target may specifically include:
and correspondingly outputting each alarm information output by the target detection alarm unit and the face retrieval results of the associated alarm target and the first target.
Alternatively, the number of the target detection alarm units may be plural, and alarm events of different target detection alarm units are different;
in the post-processing unit, matching the alarm target track information output by each target detection alarm unit with the target track information output by the structured target, determining a first target successfully matched, and respectively carrying out target modeling and target retrieval on each first target to obtain a target retrieval result;
The device may further include a second-level comparing unit, configured to match and pair target search results of the first targets corresponding to the different target detection alarm units, and output the best matched target search result and the alarm information corresponding to the best matched target search result through the output unit.
Optionally, the target detection alarm unit may be an indoor target detection alarm unit, the determined alarm information may be a close-range contact alarm, the alarm target corresponding to the alarm information may be two detected targets meeting the close-range contact condition, and the target modeling may be face modeling;
In the post-processing unit, when the track of the alarm target is matched with the track of each structured target respectively, matching can be carried out aiming at each of the two targets;
In a face modeling module, face modeling is performed for each of two output structured targets;
In the face retrieval module, retrieving is carried out aiming at each of the two output structured targets to obtain respective face retrieval results of the two output structured targets;
and in the output module, correspondingly outputting each alarm information and the face retrieval results of the two corresponding output structural targets.
Optionally, in the post-processing module, the processing of matching the track of the alarm target with the track of each structured target output by the structured target detection unit may specifically include:
And comparing the track of the alarm target with the track of each structured target, and selecting the track of the structured target matched with the track of the alarm target from the tracks of each structured target based on the curve fitting effect of the track, the coincidence degree of the positions of the tracks in the whole frame of image and the coincidence degree of the timestamp information corresponding to the track points at the same positions.
Optionally, when the target detection alarm unit does not output alarm information, the post-processing unit may be further configured to select one structured target from all the structured targets output by the structured target detection unit;
the face modeling module can be used for carrying out target modeling based on the structural detection result of the structural target selected by the post-processing unit and storing the result of the target modeling in the target model library.
The present application also provides a computer readable storage medium storing instructions that, when executed by a processor, perform steps in a method of achieving target analysis as described above. In practice, the computer readable medium may be comprised by or separate from the apparatus/device/system of the above embodiments, and may not be incorporated into the apparatus/device/system. Wherein the instructions are stored in a computer readable storage medium, which stored instructions, when executed by a processor, can perform the steps in the target analysis method as described above.
According to the disclosed embodiments of this application, the computer-readable storage medium may be a non-volatile computer-readable storage medium, which may include, for example, but is not limited to, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing, but is not intended to limit the scope of the application. In the disclosed embodiments, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
Fig. 6 is an electronic device provided by the application. As shown in fig. 6, a schematic structural diagram of an electronic device according to an embodiment of the present application is shown, specifically:
The electronic device may include a processor 601 of one or more processing cores, a memory 602 of one or more computer readable storage media, and a computer program stored on the memory and executable on the processor. The target analysis method can be implemented when the program of the memory 602 is executed.
Specifically, in practical applications, the electronic device may further include a power supply 603, an input/output unit 604, and other components. It will be appreciated by those skilled in the art that the structure of the electronic device shown in fig. 6 is not limiting of the electronic device and may include more or fewer components than shown, or may combine certain components, or a different arrangement of components. Wherein:
The processor 601 is a control center of the electronic device, connects various parts of the entire electronic device using various interfaces and lines, and performs various functions of a server and processes data by running or executing software programs and/or modules stored in the memory 602, and calling data stored in the memory 602, thereby performing overall monitoring of the electronic device.
The memory 602 may be used to store software programs and modules, i.e., the computer-readable storage media described above. The processor 601 executes various functional applications and data processing by running software programs and modules stored in the memory 602. The memory 602 may mainly include a storage program area that may store an operating system, an application program required for at least one function, and the like, and a storage data area that may store data created according to the use of a server, and the like. In addition, the memory 602 may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory 602 may also include a memory controller to provide access to the memory 602 by the processor 601.
The electronic device further comprises a power supply 603 for supplying power to the various components, which may be logically connected to the processor 601 via a power management system, so that functions of managing charging, discharging, power consumption management, etc. are achieved via the power management system. The power supply 603 may also include one or more of any components, such as a direct current or alternating current power supply, a recharging system, a power failure detection circuit, a power converter or inverter, a power status indicator, and the like.
The electronic device may also include an input output unit 604, which input unit output 604 may be used to receive input numeric or character information and to generate keyboard, mouse, joystick, optical signal inputs related to user settings and function control. The input unit output 604 may also be used to display information entered by a user or provided to a user as well as various graphical user interfaces that may be composed of graphics, text, icons, video, and any combination thereof.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather to enable any modification, equivalent replacement, improvement or the like to be made within the spirit and principles of the invention.