CN115830503A - Method and system for judging illegal water adding behavior of concrete tank truck - Google Patents
Method and system for judging illegal water adding behavior of concrete tank truckDownload PDFInfo
- Publication number
- CN115830503A CN115830503ACN202211549594.2ACN202211549594ACN115830503ACN 115830503 ACN115830503 ACN 115830503ACN 202211549594 ACN202211549594 ACN 202211549594ACN 115830503 ACN115830503 ACN 115830503A
- Authority
- CN
- China
- Prior art keywords
- rectangular frame
- tanker
- behavior
- judging
- license plate
- 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.)
- Pending
Links
Images
Landscapes
- Traffic Control Systems (AREA)
Abstract
Description
Translated fromChinese技术领域technical field
本发明属于建筑工程及人工智能技术领域，尤其涉及一种混凝土罐车违规加水行为判别方法和系统。The invention belongs to the technical fields of construction engineering and artificial intelligence, and in particular relates to a method and system for judging illegal water addition behavior of concrete tank trucks.
背景技术Background technique
混凝土材料作为当前土木工程建设领域使用范围最广且用量最大的建筑材料之一，占据了建筑材料的半壁江山，其质量优劣对整个工程项目的质量和安全而言至关重要。然而，在商品混凝土工业化程度逐步提升的同时，我国工程项目混凝土质量事故却层出不穷，不仅造成大量原材料浪费，且严重影响工程质量和安全，对项目质量管理工作带来了严峻考验。As one of the most widely used and used building materials in the field of civil engineering construction, concrete accounts for half of the building materials, and its quality is crucial to the quality and safety of the entire engineering project. However, while the industrialization of ready-to-use concrete is gradually improving, accidents of concrete quality in my country's engineering projects are emerging one after another, which not only causes a lot of waste of raw materials, but also seriously affects the quality and safety of projects, and brings a severe test to project quality management.
对大量事故的情况进行分析后发现主要原因有两点：1、工人为方便浇筑私自给混凝土加水，导致混凝土水灰比增加、强度降低，使得浇筑构建出现质量薄弱点。2、混凝土浇筑时，振捣不充分导致混凝土疏松不密实，出现蜂窝状孔洞或形成局部孔洞。After analyzing the situation of a large number of accidents, it is found that there are two main reasons: 1. Workers add water to the concrete privately for the convenience of pouring, resulting in an increase in the water-cement ratio and a decrease in the strength of the concrete, which makes the quality of the pouring structure weak. 2. When concrete is poured, insufficient vibration leads to loose and not dense concrete, and honeycomb holes or local holes are formed.
据了解，上述违规行为在我国的施工项目屡见不鲜，由此导致的工程质量问题更是常有报道，但至今仍没有很好的解决方案。通常采用视频监控方式无法全面覆盖施工场地，给作业人员违规操作留下了可乘之机。而采用现场监督和巡视进行管控，需要花费大量人力，且效果不佳。客观上讲，私自加水这种违规行为大多是在作业人员的主观意愿下进行的，具有明知故犯的特点，违规人员逃避检查的思想强烈，使得违规行为隐蔽性很强，最终导致相应监管措施收效甚微。另外，传统监管方式存在明显的滞后性，无法在私自加水违规行为出现时予以及时制止，往往只能在质量问题出现后进行事后处理并追责，这变相增加了质量管理工作的任务量和难度。It is understood that the above violations are not uncommon in my country's construction projects, and the resulting engineering quality problems are often reported, but there is still no good solution. Usually, the video surveillance method cannot fully cover the construction site, leaving opportunities for operators to operate illegally. The use of on-site supervision and inspection for control requires a lot of manpower, and the effect is not good. Objectively speaking, most of the violations such as adding water without authorization are carried out under the subjective will of the operators, which has the characteristics of knowingly committing crimes. The violators have a strong idea of evading inspection, which makes the violations very concealed, and ultimately leads to the corresponding regulatory measures with little effect. micro. In addition, there is an obvious lag in the traditional supervision method, which cannot be stopped in time when illegal water addition occurs, and often can only be dealt with after the quality problem occurs and held accountable, which in disguise increases the task volume and difficulty of quality management. .
发明内容Contents of the invention
本发明的目的是提供一种混凝土罐车违规加水行为判别方法和系统，解决了混凝土私自加水行为传统方式人工监督难以发现，混凝土水灰比增加、强度降低等问题，变事后管理为事中管理，有助于改变传统现场管理工作模式，提升工程项目施工管理水平。The purpose of the present invention is to provide a method and system for judging illegal water addition behavior of concrete tank trucks, which solves the problems that the unauthorized water addition behavior of concrete is difficult to be found by traditional manual supervision, the increase of concrete water-cement ratio, and the decrease of strength. It is helpful to change the traditional on-site management work mode and improve the construction management level of engineering projects.
本发明所采用的技术方案是：The technical scheme adopted in the present invention is:
一种混凝土罐车违规加水行为判别方法，其包括：A method for judging the behavior of illegally adding water to a concrete tank truck, which includes:
采集施工现场监控视频流，对视频画面进行预处理；Collect the monitoring video stream of the construction site and preprocess the video images;
对预处理后的视频画面中的目标对象进行YOLOv5目标检测，所述目标对象包括罐车、水管、人、车牌；Perform YOLOv5 target detection on the target objects in the preprocessed video screen, the target objects include tankers, water pipes, people, license plates;
根据YOLOv5目标检测结果，构建检测到的目标对象的对象列表；According to the YOLOv5 target detection results, construct an object list of detected target objects;
根据对象列表以及目标对象间区域重合关系判断加水行为；Judging the water addition behavior according to the object list and the overlapping relationship between the target objects;
根据罐车车牌特征记录判断加水行为；Judging the behavior of adding water according to the characteristic record of the license plate of the tanker;
根据行为识别模型判断加水行为。Judging the behavior of adding water based on the behavior recognition model.
在本发明的实施例中，当YOLOv5目标检测结果检测到罐车但没有检测到水管时，对罐车附近的截图进行基于学习神经网络的伪装目标检测，识别水管并加入到对象列表。In the embodiment of the present invention, when the YOLOv5 target detection result detects the tanker but does not detect the water pipe, the screenshot near the tanker is detected based on the learning neural network, and the water pipe is identified and added to the object list.
在本发明的实施例中，根据对象列表以及目标对象间区域重合关系判断加水行为，包括：In an embodiment of the present invention, judging the behavior of adding water according to the object list and the area overlap relationship between target objects includes:
当对象列表包含罐车、人和水管，且人和水管区域有重合，判定为加水行为；When the object list contains tankers, people and water pipes, and the areas of people and water pipes overlap, it is judged as water adding behavior;
当对象列表包含罐车和人，根据罐车矩形框确定进料口区域矩形框，根据进料口区域矩形框与人的关系，判定加水行为。When the object list contains tank cars and people, determine the rectangular frame of the feed inlet area according to the rectangular frame of the tank car, and determine the behavior of adding water according to the relationship between the rectangular frame of the feed port area and people.
在本发明的实施例中，根据罐车矩形框确定进料口区域矩形框，包括：In an embodiment of the present invention, the rectangular frame of the feed opening area is determined according to the rectangular frame of the tank car, including:
若对象列表包含罐车车牌，对于罐车左后视角的监控画面，进料口区域矩形框的长为罐车矩形框长的30％，宽为罐车矩形框宽的30％和车牌底部与罐车矩形框顶部距离的较大者，进料口区域矩形框的右上顶点与罐车矩形框右上顶点对齐；If the object list contains the license plate of the tanker, for the monitoring screen of the left rear view of the tanker, the length of the rectangular frame of the feed opening area is 30% of the length of the rectangular frame of the tanker, the width is 30% of the width of the rectangular frame of the tanker and the bottom of the license plate is the same as the top of the rectangular frame of the tanker The larger the distance, the upper right vertex of the rectangular frame of the feed inlet area is aligned with the upper right vertex of the rectangular frame of the tanker;
若对象列表包含罐车车牌，对于罐车右后视角的监控画面，进料口区域矩形框的长为罐车矩形框长的30％，宽为罐车矩形框宽的30％和车牌底部与罐车矩形框顶部距离的较大者，进料口区域矩形框的左上顶点与罐车矩形框左上顶点对齐；If the object list contains the license plate of the tanker, for the monitoring screen of the right rear view of the tanker, the length of the rectangular frame of the feed opening area is 30% of the length of the rectangular frame of the tanker, the width is 30% of the width of the rectangular frame of the tanker and the bottom of the license plate is the same as the top of the rectangular frame of the tanker For the larger distance, the upper left vertex of the rectangular frame of the feed inlet area is aligned with the upper left vertex of the rectangular frame of the tanker;
若对象列表包含罐车车牌，对于罐车后中视角的监控画面，进料口区域矩形框的长为罐车矩形框长，宽为罐车矩形框宽的30％和车牌底部与罐车矩形框顶部距离的较大者，进料口区域矩形框的左上顶点与罐车矩形框左上顶点对齐；If the object list contains the license plate of the tanker, for the monitoring screen of the tanker's rear center perspective, the length of the rectangular frame of the feed opening area is the length of the tanker's rectangular frame, and the width is 30% of the width of the tanker's rectangular frame and the distance between the bottom of the license plate and the top of the tanker's rectangular frame. For the larger one, the upper left vertex of the rectangular frame of the feed opening area is aligned with the upper left vertex of the rectangular frame of the tanker;
若对象列表没有罐车车牌，进料口区域矩形框的长为罐车矩形框长，宽为罐车矩形框宽的20％，进料口区域矩形框的左上顶点与罐车矩形框左上顶点对齐。If there is no license plate of the tanker in the object list, the length of the rectangular frame of the feed opening area is the length of the rectangular frame of the tanker, and the width is 20% of the width of the rectangular frame of the tanker. The upper left vertex of the rectangular frame of the feed inlet area is aligned with the upper left vertex of the rectangular frame of the tanker.
在本发明的实施例中，根据行为识别模型判断加水行为，包括：In an embodiment of the present invention, the behavior of adding water is judged according to the behavior recognition model, including:
当对象列表包含罐车、车牌和人，根据罐车矩形框确定警戒区域矩形框；When the object list contains tank cars, license plates and people, determine the warning area rectangle according to the tank car rectangle;
确定警戒区域矩形框与人的关系；Determine the relationship between the rectangular frame of the warning area and the person;
截取视频传入行为识别模型判断是否有加水行为。The intercepted video is passed into the behavior recognition model to judge whether there is water adding behavior.
在本发明的实施例中，根据罐车矩形框确定警戒区域矩形框，包括：In an embodiment of the present invention, the rectangular frame of the warning area is determined according to the rectangular frame of the tank car, including:
若罐车车牌安装在罐车后上方，警戒区域矩形框的长为罐车矩形框长的40％，宽为罐车矩形框宽的50％，警戒区域矩形框的底边与罐车矩形框底边对齐且警戒区域矩形框的垂直中线与车牌垂直中线对齐；If the license plate of the tank car is installed above the rear of the tank car, the length of the rectangular frame of the warning area is 40% of the length of the rectangular frame of the tank car, and the width is 50% of the width of the rectangular frame of the tank car. The bottom edge of the rectangular frame of the warning area is aligned with the bottom edge of the rectangular frame of the tank car The vertical centerline of the rectangular frame of the area is aligned with the vertical centerline of the license plate;
若罐车车牌安装在罐车下方左侧，警戒区域矩形框的长为罐车矩形框长的40％，宽为罐车矩形框宽的50％，警戒区域矩形框的底边与罐车矩形框底边对齐且警戒区域矩形框的左侧边与车牌右侧边对齐；If the license plate of the tank car is installed on the left side below the tank car, the length of the rectangular frame of the warning area is 40% of the length of the rectangular frame of the tank car, and the width is 50% of the width of the rectangular frame of the tank car. The bottom edge of the rectangular frame of the warning area is aligned with the bottom edge of the rectangular frame of the tank car The left side of the warning area rectangle is aligned with the right side of the license plate;
确定警戒区域矩形框与人的关系；Determine the relationship between the rectangular frame of the warning area and the person;
截取视频传入行为识别模型判断是否有加水行为；The intercepted video is passed into the behavior recognition model to judge whether there is water addition;
在本发明的实施例中，根据罐车车牌特征记录判断加水行为，包括：In an embodiment of the present invention, judging the behavior of adding water according to the characteristic record of the license plate of the tanker includes:
若对象列表包含罐车、车牌，对罐车车牌特征进行提取与匹配，若罐车离开后十分钟内再次返回，判定为加水行为。If the object list contains tankers and license plates, extract and match the license plate features of the tanker, and if the tanker returns within ten minutes after leaving, it is judged to be adding water.
一种混凝土罐车违规加水行为判别方法，其包括：A method for judging the behavior of illegally adding water to a concrete tank truck, which includes:
数据与处理模块，用于采集施工现场监控视频流，对视频画面进行预处理；The data and processing module is used to collect the monitoring video stream of the construction site and preprocess the video images;
目标识别模块，用于对预处理后的视频画面中的目标对象进行YOLOv5目标检测，根据YOLOv5目标检测结果，构建检测到的目标对象的对象列表，所述目标对象包括罐车、水管、人、车牌；Target recognition module, for carrying out YOLOv5 target detection to the target object in the preprocessed video picture, according to YOLOv5 target detection result, build the object list of the detected target object, described target object comprises tanker, water pipe, people, license plate ;
罐车违规加水行为判别模块，用于根据对象列表以及目标对象间区域重合关系判断加水行为，根据罐车车牌特征记录判断加水行为，根据行为识别模型判断加水行为。The judging module of illegal water adding behavior of tank cars is used to judge water adding behavior according to the object list and the area overlap relationship between target objects, judge water adding behavior according to tank car license plate feature records, and judge water adding behavior according to behavior recognition model.
本发明的有益效果是：The beneficial effects of the present invention are:
本发明的方法检测结果准确率高，漏报和误报情况少，以低成本的方式保证了算法的精度和效率，本方法实时对混凝土罐车私自加水违规行为进行预警，可避免材料浪费，节约人力成本，提升混凝土质量管理水平，解决了混凝土私自加水行为传统方式人工监督难以发现，混凝土水灰比增加、强度降低等问题，变事后管理为事中管理，改变了传统现场管理工作模式，提升工程项目施工管理水平和施工质量，可解决施工过程面临的产业难题，具有巨大的推广应用价值。The detection result of the method of the present invention has a high accuracy rate, less false positives and false positives, and ensures the accuracy and efficiency of the algorithm at low cost. Labor costs, improve the level of concrete quality management, solve the problems of adding water to concrete without manual supervision in the traditional way, the increase of concrete water-cement ratio, and the decrease of strength. The level of construction management and construction quality of engineering projects can solve the industrial problems faced in the construction process, and has great promotion and application value.
附图说明Description of drawings
为了更清楚地说明本发明实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings that need to be used in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those skilled in the art, other drawings can also be obtained based on these drawings without creative effort.
图1是本发明一种违规加水行为判别方法的结构示意图。Fig. 1 is a structural schematic diagram of a method for judging illegal water addition in the present invention.
图2是本发明一种违规加水行为判别方法中设计的程序流程图。Fig. 2 is a flow chart of a program designed in a method for judging illegal water addition in the present invention.
具体实施方式Detailed ways
下面结合附图对本发明的具体实施方式作进一步说明。在此需要说明的是，对于这些实施方式的说明用于帮助理解本发明，但并不构成对本发明的限定。此外，下面所描述的本发明各个实施方式中所涉及的技术特征只要彼此之间未构成冲突就可以相互组合。The specific embodiments of the present invention will be further described below in conjunction with the accompanying drawings. It should be noted here that the descriptions of these embodiments are used to help understand the present invention, but are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not constitute a conflict with each other.
如图1所示，本实例提供了一种违规加水行为判别系统，具体模块及功能如下：As shown in Figure 1, this example provides a system for judging illegal water addition. The specific modules and functions are as follows:
(一)数据与处理模块，用于施工现场监控视频流，对视频画面进行预处理；(1) The data and processing module is used to monitor the video stream at the construction site and preprocess the video images;
预处理过程一般有数字化、几何变换、归一化、平滑、复原和增强等步骤。是将每一个文字图像分检出来交给识别模块识别，这一过程称为图像预处理。在图像分析中，对输入图像进行特征抽取、分割和匹配前所进行的处理。图像预处理的主要目的是消除图像中无关的信息，恢复有用的真实信息，增强有关信息的可检测性和最大限度地简化数据，从而改进特征抽取、图像分割、匹配和识别的可靠性。The preprocessing process generally includes steps such as digitization, geometric transformation, normalization, smoothing, restoration, and enhancement. It is to sort out each text image and hand it over to the recognition module for recognition. This process is called image preprocessing. In image analysis, the input image is processed before feature extraction, segmentation and matching. The main purpose of image preprocessing is to eliminate irrelevant information in the image, restore useful real information, enhance the detectability of relevant information and minimize data, thereby improving the reliability of feature extraction, image segmentation, matching and recognition.
通过在施工现场布置监控摄像头可以实时获取施工现场监控视频流，从中截取视频画面的图像，进行图像预处理。By arranging monitoring cameras at the construction site, the monitoring video stream of the construction site can be obtained in real time, and the image of the video screen can be intercepted from it for image preprocessing.
(二)目标识别模块，用于对预处理后的视频画面中的目标对象进行YOLOv5目标检测，根据YOLOv5目标检测结果，构建检测到的目标对象的对象列表，其中，目标对象包括但不限于罐车、水管、人、车牌等。(2) Target recognition module, for performing YOLOv5 target detection on the target object in the preprocessed video picture, according to the YOLOv5 target detection result, constructing the object list of the detected target object, wherein, the target object includes but not limited to tanker , water pipes, people, license plates, etc.
其中，YOLOv5目标检测是现有技术，主要通过改进的YOLOv5模型实现。Among them, YOLOv5 target detection is an existing technology, which is mainly realized through the improved YOLOv5 model.
YOLOv5是一种单阶段目标检测算法，该算法在YOLOv4的基础上添加了一些新的改进思路，使其速度与精度都得到了极大的性能提升。主要的改进思路如下：YOLOv5 is a single-stage target detection algorithm. Based on YOLOv4, this algorithm has added some new improvement ideas, so that its speed and accuracy have been greatly improved. The main improvement ideas are as follows:
输入端：在模型训练阶段，提出了一些改进思路，主要包括Mosaic数据增强、自适应锚框计算、自适应图片缩放；Input: In the model training phase, some improvement ideas were proposed, mainly including Mosaic data enhancement, adaptive anchor frame calculation, and adaptive image scaling;
基准网络：融合其它检测算法中的一些新思路，主要包括：Focus结构与CSP结构；Benchmark network: Integrating some new ideas in other detection algorithms, mainly including: Focus structure and CSP structure;
Neck网络：目标检测网络在BackBone与最后的Head输出层之间往往会插入一些层，Yolov5中添加了FPN+PAN结构；Neck network: The target detection network often inserts some layers between the BackBone and the final Head output layer, and the FPN+PAN structure is added to Yolov5;
Head输出层：输出层的锚框机制与YOLOv4相同，主要改进的是训练时的损失函数GIOU_Loss，以及预测框筛选的DIOU_nms。Head output layer: The anchor frame mechanism of the output layer is the same as that of YOLOv4. The main improvement is the loss function GIOU_Loss during training and DIOU_nms for prediction frame screening.
关于YOLOv5目标检测，更多可参考中国专利CN115223009A中公开的一种基于改进型YOLOv5的小目标检测方法及装置。For YOLOv5 target detection, more reference can be made to an improved YOLOv5-based small target detection method and device disclosed in Chinese patent CN115223009A.
进一步的，当YOLOv5目标检测结果检测到罐车但没有检测到水管时，则将罐车附近的人截图传入伪装物体识别模型，进行基于学习神经网络的伪装目标检测，识别水管，将结果加入到对象列表。Further, when the YOLOv5 target detection result detects the tanker but does not detect the water pipe, the screenshot of the person near the tanker is passed to the camouflage object recognition model, and the camouflage target detection based on the learning neural network is performed to identify the water pipe, and the result is added to the object list.
基于学习神经网络的伪装目标检测是现有技术，主要通过伪装目标检测模型实现，可参考中国专利CN114842324A公开的一种基于学习神经网络的伪装目标检测方法及系统。能够对YOLOv5模型未检测到的水管进一步进行查漏，从而更准确地识别水管。The detection of camouflaged targets based on learning neural networks is an existing technology, which is mainly realized by a camouflaged target detection model. For reference, a method and system for detecting camouflaged targets based on learning neural networks disclosed in Chinese patent CN114842324A. It is possible to further detect leaks in water pipes not detected by the YOLOv5 model, so as to identify water pipes more accurately.
(三)罐车违规加水行为判别模块，用于判定加水行为，具体如下：(3) The judging module for illegal water filling behavior of tank trucks, which is used to judge the water adding behavior, the details are as follows:
判定方法一：根据对象列表以及目标对象间区域重合关系判断加水行为，包括：Judgment method 1: Judging the water adding behavior based on the object list and the area overlap relationship between target objects, including:
(1)若检测结果包含罐车、人和水管(桶)，且人和水管(桶)临近，判定为加水行为；(1) If the test results include tankers, people and water pipes (barrels), and the people and water pipes (barrels) are close to each other, it is determined to be adding water;
(2)若检测结果包含罐车和人，且人在罐车进料口区域，判定为加水行为，具体如下：(2) If the test results include tank trucks and people, and the people are in the area of the tank truck feed inlet, it is determined to be water adding behavior, the details are as follows:
若检测到了罐车车牌，对于罐车左后视角的监控画面，进料口区域矩形框的长为罐车矩形框长的30％，宽为罐车矩形框宽的30％和车牌底部与罐车矩形框顶部距离的较大者，进料口区域矩形框的右上顶点与罐车矩形框右上顶点对齐；If the license plate of the tank car is detected, for the monitoring screen of the left rear view of the tank car, the length of the rectangular frame of the feed opening area is 30% of the length of the rectangular frame of the tank car, the width is 30% of the width of the rectangular frame of the tank car and the distance between the bottom of the license plate and the top of the rectangular frame of the tank car The larger of , the upper right vertex of the rectangular frame of the feed inlet area is aligned with the upper right vertex of the rectangular frame of the tank car;
若检测到了罐车车牌，对于罐车右后视角的监控画面，进料口区域矩形框的长为罐车矩形框长的30％，宽为罐车矩形框宽的30％和车牌底部与罐车矩形框顶部距离的较大者，进料口区域矩形框的左上顶点与罐车矩形框左上顶点对齐；If the license plate of the tanker is detected, for the monitoring screen of the right rear view of the tanker, the length of the rectangular frame of the feed opening area is 30% of the length of the rectangular frame of the tanker, the width is 30% of the width of the rectangular frame of the tanker and the distance between the bottom of the license plate and the top of the rectangular frame of the tanker The larger one, the upper left vertex of the rectangular frame of the feed inlet area is aligned with the upper left vertex of the rectangular frame of the tanker;
若检测到了罐车车牌，对于罐车后中视角的监控画面，进料口区域矩形框的长为罐车矩形框长，宽为罐车矩形框宽的30％和车牌底部与罐车矩形框顶部距离的较大者，进料口区域矩形框的左上顶点与罐车矩形框左上顶点对齐；If the license plate of the tanker is detected, for the monitoring screen of the tanker's rear center perspective, the length of the rectangular frame of the feed opening area is the length of the tanker's rectangular frame, and the width is 30% of the width of the tanker's rectangular frame, and the distance between the bottom of the license plate and the top of the tanker's rectangular frame is greater Or, the upper left vertex of the rectangular frame of the feed opening area is aligned with the upper left vertex of the rectangular frame of the tank car;
若没有检测到罐车车牌，进料口区域矩形框的长为罐车矩形框长，宽为罐车矩形框宽的20％，进料口区域矩形框的左上顶点与罐车矩形框左上顶点对齐。If the license plate of the tank car is not detected, the length of the rectangular frame in the area of the feed opening is the length of the rectangular frame of the tank car, and the width is 20% of the width of the rectangular frame of the tank car.
(3)根据进料口区域矩形框与人的关系，判定加水行为。(3) According to the relationship between the rectangular frame in the area of the feed inlet and the person, determine the behavior of adding water.
判定方法二：根据行为识别模型判断加水行为，具体为：若检测结果包含罐车、车牌和人，且人在警戒区域，截取视频传入行为识别模型判断是否有加水行为，具体如下：Judgment method 2: Judging the behavior of adding water according to the behavior recognition model, specifically: if the detection results include tankers, license plates and people, and the people are in the warning area, intercept the video and pass it into the behavior recognition model to judge whether there is water adding behavior, the details are as follows:
(1)根据罐车矩形框确定警戒区域矩形框，具体如下：(1) Determine the rectangular frame of the warning area according to the rectangular frame of the tanker, as follows:
若罐车车牌安装在罐车后上方，警戒区域矩形框的长为罐车矩形框长的40％，宽为罐车矩形框宽的50％，警戒区域矩形框的底边与罐车矩形框底边对齐且警戒区域矩形框的垂直中线与车牌垂直中线对齐；If the license plate of the tank car is installed above the rear of the tank car, the length of the rectangular frame of the warning area is 40% of the length of the rectangular frame of the tank car, and the width is 50% of the width of the rectangular frame of the tank car. The bottom edge of the rectangular frame of the warning area is aligned with the bottom edge of the rectangular frame of the tank car The vertical centerline of the rectangular frame of the area is aligned with the vertical centerline of the license plate;
若罐车车牌安装在罐车下方左侧，警戒区域矩形框的长为罐车矩形框长的40％，宽为罐车矩形框宽的50％，警戒区域矩形框的底边与罐车矩形框底边对齐且警戒区域矩形框的左侧边与车牌右侧边对齐；If the license plate of the tank car is installed on the left side below the tank car, the length of the rectangular frame of the warning area is 40% of the length of the rectangular frame of the tank car, and the width is 50% of the width of the rectangular frame of the tank car. The bottom edge of the rectangular frame of the warning area is aligned with the bottom edge of the rectangular frame of the tank car The left side of the warning area rectangle is aligned with the right side of the license plate;
(2)确定警戒区域矩形框与人的关系；(2) Determine the relationship between the rectangular frame of the warning area and the person;
(3)截取视频传入行为识别模型判断是否有加水行为。(3) The intercepted video is passed into the behavior recognition model to judge whether there is water adding behavior.
其中，行为识别模型为现有技术，可参考中国专利CN114627397A公开的一种行为识别模型构建方法以及行为识别方法。Wherein, the behavior recognition model is a prior art, and reference may be made to a method for constructing a behavior recognition model and a behavior recognition method disclosed in Chinese patent CN114627397A.
判定方法三：根据罐车车牌特征记录判断加水行为，具体为：若检测结果包含罐车、车牌，对罐车车牌特征进行提取与匹配，若罐车离开后十分钟内再次返回，判定为加水行为。Judgment method 3: Judging the behavior of adding water according to the characteristic records of the license plate of the tanker, specifically: if the detection result includes the tanker and the license plate, extract and match the characteristics of the license plate of the tanker, and if the tanker returns within ten minutes after leaving, it is determined to be a behavior of adding water.
配合图2，一种混凝土罐车违规加水行为判别方法，具体按以下步骤实施：With reference to Fig. 2, a method for judging illegal water addition behavior of concrete tank trucks is implemented according to the following steps:
步骤1，采集施工现场监控视频流，对视频画面进行预处理；Step 1, collect the monitoring video stream of the construction site, and preprocess the video images;
步骤2，使用改进的YOLOv5模型和伪装物体识别模型检测罐车、水管等对象；Step 2, using the improved YOLOv5 model and the camouflaged object recognition model to detect objects such as tankers and water pipes;
步骤3，判定加水行为；Step 3, determine the behavior of adding water;
本发明的特点还在于：The present invention is also characterized in that:
其中步骤2中目标检测具体按以下步骤实施：Wherein the target detection in step 2 is specifically implemented according to the following steps:
步骤2.1，对画面进行YOLOv5目标检测，检测罐车、水管等对象；Step 2.1, perform YOLOv5 target detection on the screen, and detect objects such as tankers and water pipes;
步骤2.2，如果没有检测到水管(桶)对象，则将罐车附近的人截图传入伪装物体识别模型识别水管；Step 2.2, if no water pipe (barrel) object is detected, the screenshot of the person near the tanker is passed into the camouflaged object recognition model to identify the water pipe;
其中步骤3中使用罐车违规加水行为判别模块，判定加水行为，具体按以下步骤实施：In step 3, the judging module for illegal water addition behavior of tank trucks is used to determine the water addition behavior, which is implemented in the following steps:
步骤3.1，若检测结果包含罐车、人和水管(桶)，且人和水管(桶)临近，判定为加水行为；Step 3.1, if the detection result includes a tanker, a person and a water pipe (barrel), and the person and the water pipe (barrel) are close to each other, it is determined to be adding water;
步骤3.2，若检测结果包含罐车和人，且人在罐车进料口区域，判定为加水行为；Step 3.2, if the detection result includes tank trucks and people, and the people are in the area of the tank truck feed inlet, it is determined to be adding water;
步骤3.3，若检测结果包含罐车、车牌和人，且人在警戒区域，截取视频传入行为识别模型判断是否有加水行为；Step 3.3, if the detection result includes tank trucks, license plates and people, and the people are in the warning area, intercept the video and pass it into the behavior recognition model to determine whether there is water adding behavior;
步骤3.4，若检测结果包含罐车、车牌，对罐车车牌特征进行提取与匹配，若罐车离开后十分钟内再次返回，判定为加水行为；Step 3.4, if the detection result includes the tanker and the license plate, extract and match the license plate features of the tanker, and if the tanker returns within ten minutes after leaving, it is determined to be adding water;
其中步骤3.2对人是否在进料口区域的判别具体按以下步骤实施：Wherein, step 3.2 specifically implements the following steps to determine whether a person is in the area of the feed inlet:
步骤3.2.1，根据罐车矩形框确定进料口区域矩形框；Step 3.2.1, determine the rectangular frame of the feed inlet area according to the rectangular frame of the tank car;
步骤3.2.2，根据进料口区域矩形框与人的关系，判定加水行为；Step 3.2.2, according to the relationship between the rectangular frame in the area of the feed inlet and the person, determine the behavior of adding water;
其中步骤3.2.1对进料口区域使用矩形框标注，具体步骤如下：Among them, step 3.2.1 uses a rectangular frame to mark the area of the feed inlet, and the specific steps are as follows:
步骤3.2.1.1，若检测到了罐车车牌，对于罐车左后视角的监控画面，进料口区域矩形框的长为罐车矩形框长的30％，宽为罐车矩形框宽的30％和车牌底部与罐车矩形框顶部距离的较大者，进料口区域矩形框的右上顶点与罐车矩形框右上顶点对齐；Step 3.2.1.1, if the license plate of the tank car is detected, for the monitoring screen from the left rear perspective of the tank car, the length of the rectangular frame of the feed opening area is 30% of the length of the rectangular frame of the tank car, the width is 30% of the width of the rectangular frame of the tank car and the bottom of the license plate is The larger the distance from the top of the rectangular frame of the tank car, the upper right vertex of the rectangular frame of the feed opening area is aligned with the upper right vertex of the rectangular frame of the tank car;
步骤3.2.1.2，若检测到了罐车车牌，对于罐车右后视角的监控画面，进料口区域矩形框的长为罐车矩形框长的30％，宽为罐车矩形框宽的30％和车牌底部与罐车矩形框顶部距离的较大者，进料口区域矩形框的左上顶点与罐车矩形框左上顶点对齐；Step 3.2.1.2, if the license plate of the tank car is detected, for the monitoring screen from the right rear perspective of the tank car, the length of the rectangular frame in the area of the feed opening is 30% of the length of the rectangular frame of the tank car, the width is 30% of the width of the rectangular frame of the tank car and the bottom of the license plate is The larger the distance from the top of the rectangular frame of the tank car, the upper left vertex of the rectangular frame in the area of the feed opening is aligned with the upper left vertex of the rectangular frame of the tank car;
步骤3.2.1.3，若检测到了罐车车牌，对于罐车后中视角的监控画面，进料口区域矩形框的长为罐车矩形框长，宽为罐车矩形框宽的30％和车牌底部与罐车矩形框顶部距离的较大者，进料口区域矩形框的左上顶点与罐车矩形框左上顶点对齐；Step 3.2.1.3, if the license plate of the tank car is detected, for the monitoring screen of the rear view of the tank car, the length of the rectangular frame of the feed opening area is the length of the rectangular frame of the tank car, the width is 30% of the width of the rectangular frame of the tank car and the bottom of the license plate is in line with the rectangular frame of the tank car The larger of the top distances, the upper left vertex of the rectangular frame of the feed opening area is aligned with the upper left vertex of the rectangular frame of the tanker;
步骤3.2.1.4，若没有检测到罐车车牌，进料口区域矩形框的长为罐车矩形框长，宽为罐车矩形框宽的20％，进料口区域矩形框的左上顶点与罐车矩形框左上顶点对齐；Step 3.2.1.4, if the license plate of the tank car is not detected, the length of the rectangular frame of the tank car is as long as the length of the rectangular frame of the tank car, and the width is 20% of the width of the rectangular frame of the tank car. vertex alignment;
其中步骤3.3，对人是否在警戒区域具体按以下步骤实施：Wherein step 3.3, whether the person is in the warning area is specifically implemented according to the following steps:
步骤3.3.1，根据罐车矩形框确定警戒区域矩形框；Step 3.3.1, determine the warning area rectangular frame according to the tank car rectangular frame;
步骤3.3.2，确定警戒区域矩形框与人的关系；Step 3.3.2, determine the relationship between the warning area rectangle and the person;
步骤3.3.3，截取视频传入行为识别模型判断是否有加水行为；Step 3.3.3, intercepting the video input behavior recognition model to determine whether there is water adding behavior;
其中步骤3.3.1对警戒区域使用矩形框标注，具体步骤如下：Among them, step 3.3.1 uses a rectangular frame to mark the warning area, and the specific steps are as follows:
步骤3.3.1.1，若罐车车牌安装在罐车后上方，警戒区域矩形框的长为罐车矩形框长的40％，宽为罐车矩形框宽的50％，警戒区域矩形框的底边与罐车矩形框底边对齐且警戒区域矩形框的垂直中线与车牌垂直中线对齐；Step 3.3.1.1, if the license plate of the tanker is installed above the rear of the tanker, the length of the rectangular frame of the warning area is 40% of the length of the rectangular frame of the tanker, and the width is 50% of the width of the rectangular frame of the tanker. The bottom edge is aligned and the vertical centerline of the rectangular frame of the warning area is aligned with the vertical centerline of the license plate;
步骤3.3.1.2，若罐车车牌安装在罐车下方左侧，警戒区域矩形框的长为罐车矩形框长的40％，宽为罐车矩形框宽的50％，警戒区域矩形框的底边与罐车矩形框底边对齐且警戒区域矩形框的左侧边与车牌右侧边对齐。Step 3.3.1.2, if the license plate of the tank car is installed on the left side below the tank car, the length of the rectangular frame of the warning area is 40% of the length of the rectangular frame of the tank car, and the width is 50% of the width of the rectangular frame of the tank car. The bottom edge of the frame is aligned and the left edge of the warning area rectangular frame is aligned with the right edge of the license plate.
与现有技术相比，本发明具有以下优点：Compared with the prior art, the present invention has the following advantages:
本发明的方法检测结果准确率高，漏报和误报情况少，以低成本的方式保证了算法的精度和效率，本方法实时对混凝土罐车私自加水违规行为进行预警，可避免材料浪费，节约人力成本，提升混凝土质量管理水平，解决了混凝土私自加水行为传统方式人工监督难以发现，混凝土水灰比增加、强度降低等问题，变事后管理为事中管理，改变了传统现场管理工作模式，提升工程项目施工管理水平和施工质量，可解决施工过程面临的产业难题，具有巨大的推广应用价值。The detection result of the method of the present invention has a high accuracy rate, less false positives and false positives, and ensures the accuracy and efficiency of the algorithm at low cost. Labor costs, improve the level of concrete quality management, solve the problems of adding water to concrete without manual supervision in the traditional way, the increase of concrete water-cement ratio, and the decrease of strength. The level of construction management and construction quality of engineering projects can solve the industrial problems faced in the construction process, and has great promotion and application value.
以上结合附图对本发明的实施方式作了详细说明，但本发明不限于所描述的实施方式。对于本领域的技术人员而言，在不脱离本发明原理和精神的情况下，对这些实施方式进行多种变化、修改、替换和变型，仍落入本发明的保护范围内。The embodiments of the present invention have been described in detail above with reference to the accompanying drawings, but the present invention is not limited to the described embodiments. For those skilled in the art, without departing from the principle and spirit of the present invention, various changes, modifications, substitutions and modifications to these embodiments still fall within the protection scope of the present invention.
从上述内容可以看出，本发明很好地适用于实现上述所有目的和目标，以及其它明显的和该结构固有的优点。应当理解，某些特征和子组合是有用的，并且可以在不参考其他特征和子组合的情况下使用。这是在本发明的范围内。From the foregoing it will be seen that the invention is well adapted to carry out all of the above objects and objects, as well as other advantages which are obvious and inherent in the structure. It should be understood that certain features and subcombinations are useful and can be used without reference to other features and subcombinations. This is within the scope of the present invention.
Claims (8)
Translated fromChinesePriority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211549594.2ACN115830503A (en) | 2022-12-05 | 2022-12-05 | Method and system for judging illegal water adding behavior of concrete tank truck |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211549594.2ACN115830503A (en) | 2022-12-05 | 2022-12-05 | Method and system for judging illegal water adding behavior of concrete tank truck |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115830503Atrue CN115830503A (en) | 2023-03-21 |
Family
ID=85544055
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211549594.2APendingCN115830503A (en) | 2022-12-05 | 2022-12-05 | Method and system for judging illegal water adding behavior of concrete tank truck |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115830503A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN120318771A (en)* | 2025-06-13 | 2025-07-15 | 深圳市有为信息技术发展有限公司 | An early warning method and device for a water injection detection system for a commercial concrete truck |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103198670A (en)* | 2013-03-18 | 2013-07-10 | 陕西科技大学 | Device for detecting automobile back-up action at expressway fork and detecting method thereof |
| CN109741470A (en)* | 2018-12-12 | 2019-05-10 | 成都宜泊信息科技有限公司 | A kind of parking charging method and system based on Car license recognition |
| CN110348312A (en)* | 2019-06-14 | 2019-10-18 | 武汉大学 | A kind of area video human action behavior real-time identification method |
| CN111144232A (en)* | 2019-12-09 | 2020-05-12 | 国网智能科技股份有限公司 | Transformer substation electronic fence monitoring method based on intelligent video monitoring, storage medium and equipment |
| CN111814601A (en)* | 2020-06-23 | 2020-10-23 | 国网上海市电力公司 | A Video Analysis Method Combining Object Detection and Human Pose Estimation |
| CN113469150A (en)* | 2021-09-03 | 2021-10-01 | 中国电力科学研究院有限公司 | Method and system for identifying risk behaviors |
| CN114067283A (en)* | 2021-11-15 | 2022-02-18 | 青岛海尔工业智能研究院有限公司 | A method, device, electronic device and storage medium for identifying violations |
| CN114120234A (en)* | 2021-11-29 | 2022-03-01 | 国网宁夏电力有限公司信息通信公司 | Ladder transportation detection method and system for power operation construction and storage medium |
| CN114627286A (en)* | 2021-12-13 | 2022-06-14 | 南京理工大学 | Method for detecting wagon staff invasion based on PSPNet and improved YOLOv4 |
| CN114627397A (en)* | 2020-12-10 | 2022-06-14 | 顺丰科技有限公司 | Behavior recognition model construction method and behavior recognition method |
| CN114842324A (en)* | 2022-03-16 | 2022-08-02 | 南京邮电大学 | Method and system for detecting disguised target based on learning neural network |
| CN114973094A (en)* | 2022-06-08 | 2022-08-30 | 长沙海信智能系统研究院有限公司 | Power utilization detection method and device and electronic equipment |
- 2022
- 2022-12-05CNCN202211549594.2Apatent/CN115830503A/enactivePending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103198670A (en)* | 2013-03-18 | 2013-07-10 | 陕西科技大学 | Device for detecting automobile back-up action at expressway fork and detecting method thereof |
| CN109741470A (en)* | 2018-12-12 | 2019-05-10 | 成都宜泊信息科技有限公司 | A kind of parking charging method and system based on Car license recognition |
| CN110348312A (en)* | 2019-06-14 | 2019-10-18 | 武汉大学 | A kind of area video human action behavior real-time identification method |
| CN111144232A (en)* | 2019-12-09 | 2020-05-12 | 国网智能科技股份有限公司 | Transformer substation electronic fence monitoring method based on intelligent video monitoring, storage medium and equipment |
| CN111814601A (en)* | 2020-06-23 | 2020-10-23 | 国网上海市电力公司 | A Video Analysis Method Combining Object Detection and Human Pose Estimation |
| CN114627397A (en)* | 2020-12-10 | 2022-06-14 | 顺丰科技有限公司 | Behavior recognition model construction method and behavior recognition method |
| CN113469150A (en)* | 2021-09-03 | 2021-10-01 | 中国电力科学研究院有限公司 | Method and system for identifying risk behaviors |
| CN114067283A (en)* | 2021-11-15 | 2022-02-18 | 青岛海尔工业智能研究院有限公司 | A method, device, electronic device and storage medium for identifying violations |
| CN114120234A (en)* | 2021-11-29 | 2022-03-01 | 国网宁夏电力有限公司信息通信公司 | Ladder transportation detection method and system for power operation construction and storage medium |
| CN114627286A (en)* | 2021-12-13 | 2022-06-14 | 南京理工大学 | Method for detecting wagon staff invasion based on PSPNet and improved YOLOv4 |
| CN114842324A (en)* | 2022-03-16 | 2022-08-02 | 南京邮电大学 | Method and system for detecting disguised target based on learning neural network |
| CN114973094A (en)* | 2022-06-08 | 2022-08-30 | 长沙海信智能系统研究院有限公司 | Power utilization detection method and device and electronic equipment |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN120318771A (en)* | 2025-06-13 | 2025-07-15 | 深圳市有为信息技术发展有限公司 | An early warning method and device for a water injection detection system for a commercial concrete truck |
| CN120318771B (en)* | 2025-06-13 | 2025-09-26 | 深圳市有为信息技术发展有限公司 | Early warning method and device for commercial concrete vehicle water injection detection system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | Automatic defect detection of metro tunnel surfaces using a vision-based inspection system | |
| CN110796819B (en) | A detection method and system for intruding cross-border personnel on the yellow line of the platform | |
| CN110648364B (en) | Multi-dimensional space solid waste visual detection positioning and identification method and system | |
| CN111582188B (en) | Concrete pouring side station supervision method based on artificial intelligence | |
| CN103632158B (en) | Forest fire prevention monitor method and forest fire prevention monitor system | |
| CN102610102A (en) | Suspect vehicle inspection and control method and system | |
| CN113470374B (en) | Vehicle overspeed monitoring method and device, computer equipment and storage medium | |
| CN115691148B (en) | Intelligent charging auxiliary method, equipment and medium based on expressway | |
| CN105702047B (en) | Car license recognition mistake filter method and device in a kind of deck analysis | |
| CN101325690A (en) | Method and system for detecting human flow analysis and crowd accumulation process of monitoring video flow | |
| CN103279756A (en) | Vehicle detecting analysis system and detecting analysis method thereof based on integrated classifier | |
| CN115830503A (en) | Method and system for judging illegal water adding behavior of concrete tank truck | |
| CN108052887A (en) | A kind of doubtful illegal land automatic recognition system and method for merging SLAM/GNSS information | |
| CN111626382A (en) | Rapid intelligent identification method and system for cleanliness of vehicle on construction site | |
| CN110490150A (en) | A kind of automatic auditing system of picture violating the regulations and method based on vehicle retrieval | |
| CN105163014A (en) | Road monitoring device and method | |
| CN109614924A (en) | A kind of garbage on water detection method based on deep learning algorithm | |
| CN111931555A (en) | Method for identifying whether ship AIS is started or not by utilizing video image | |
| CN108694387A (en) | A kind of falseness car plate filter method and device | |
| CN104834932A (en) | Matlab algorithm of automobile license plate identification | |
| CN104517444B (en) | A kind of detecting system driving to use mobile phone illegal activities | |
| CN106571040A (en) | Suspicious vehicle confirmation method and equipment | |
| CN103423596B (en) | A kind of drainpipe detecting and appraisal procedure applying handheld video and closed circuit TV | |
| CN109934161B (en) | Vehicle identification and detection method and system based on convolutional neural network | |
| CN100568312C (en) | A new type of bayonet |
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 |