Early warning method and device of commercial concrete vehicle water injection detection system

Abstract

The invention provides an early warning method and device for a commercial concrete vehicle water injection detection system. The early warning method comprises the steps that a camera device is positioned and arranged on a commercial concrete vehicle; collecting video stream data in real time according to a positioning arrangement result and transmitting the video stream data to the vehicle-mounted edge equipment for image processing; performing water injection area calibration on the processed image data according to the current working environment; performing water injection behavior identification analysis in the water injection area, and performing alarm operation when the water injection behavior exists; and when the alarm operation is triggered, an uploading mechanism is automatically started to upload data. The accuracy of water injection behavior recognition is effectively improved, an uploading mechanism is automatically started to upload data when alarm operation is triggered, the traceability of the data can be effectively guaranteed, therefore, data management is effectively achieved, and the effectiveness and accuracy of commercial concrete vehicle water injection detection are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a warning method and device for a water injection detection system of a commercial concrete truck. Background Art

[0002] At present, commercial concrete trucks play a key role in transporting commercial concrete to the construction site in construction sites and are an indispensable part of building construction. As an important building material, the quality and use of concrete must strictly comply with relevant standards. These standards clearly stipulate that it is prohibited to add water privately during the transportation and pouring of concrete to ensure the accuracy of the concrete mix ratio, thereby guaranteeing the strength, durability, and overall quality of the concrete. However, in actual operations, due to the influence of various factors, some non-standard operating behaviors are difficult to eliminate and effectively supervised. For example, workers add water privately to facilitate pumping or improve fluidity, which will change the original mix ratio of the concrete, thereby causing quality problems and affecting the safety of commercial concrete. Therefore, how to effectively supervise and prevent such non-standard behaviors has become an urgent problem to be solved. Based on the above requirements, the invention patent "A method and system for discriminating illegal water addition behavior of a concrete tanker" with the patent number CN115830503A discloses a method based on the installation of fixed cameras, constructing an object list according to the detection results of multiple targets, and using three different behavior judgment modules to confirm whether water is injected. Although these methods have achieved certain effects, there are still some deficiencies in the actual process: First, the environmental adaptability is insufficient. The installation position and angle of the cameras in the existing solutions are fixed, resulting in some key areas that cannot be effectively monitored, making it difficult to adapt to different vehicle types and complex construction scenarios, and being easily affected by light and occlusion interference. Once the vehicle leaves the fixed construction site, the water injection behavior of the commercial concrete truck cannot be effectively monitored. Second, based on the detection of multiple targets, there are some problems in the implementation of the three different water injection behavior discriminations. The first method is to judge the water injection behavior based on the coincidence relationship between multiple target objects using a single-frame image without considering the misdetection situation. The second method is to record the license plate characteristics of the commercial concrete truck and judge the water injection behavior by detecting the same vehicle twice within 10 minutes without considering the misrecognition of the license plate. The third method is to use a behavior recognition model to judge whether there is a water injection behavior. When processing a large amount of video data, the behavior recognition model not only has a high computational complexity but also consumes a large amount of memory, resulting in a significant high resource occupancy problem. Summary of the Invention

[0003] The present invention provides an early warning method and device for a concrete truck water injection detection system, which can realize real-time monitoring of the water injection area of the concrete truck in different scenarios by selecting a suitable on-vehicle camera and its installation position, effectively improve the accuracy of water injection behavior recognition by analyzing and identifying water injection behavior in the water injection area and performing an alarm operation when there is a water injection behavior, and automatically start an upload mechanism to upload data when the alarm operation is triggered, which can effectively ensure the traceability of data, thus effectively realizing the management of data and improving the effectiveness and accuracy of the water injection detection of the concrete truck.

[0004] The present invention provides an early warning method for a concrete truck water injection detection system, including: Step 1: Position and arrange a camera device on the concrete truck; Step 2: Real-time collect video stream data according to the positioning and arrangement results and transmit it to the on-vehicle edge device for image processing; Step 3: Calibrate the water injection area according to the current working environment for the processed image data; Step 4: Analyze and identify water injection behavior in the water injection area, and perform an alarm operation when there is a water injection behavior; Step 5: Automatically start an upload mechanism to upload data when the alarm operation is triggered.

[0005] Preferably, for an early warning method of a concrete truck water injection detection system, in Step 1, positioning and arranging a camera device on the concrete truck includes: Obtain the structural characteristics of the concrete truck, and determine the water injection area of the concrete truck based on the structural characteristics; Determine the overall position distribution of the water injection area on the concrete truck, and determine the effective monitoring points for image acquisition of the water injection area based on the overall position distribution and structural characteristics; Pre-collect the shooting angles and ranges of each effective monitoring point based on the camera device, and determine the effective recording range and position of the water injection area of the concrete truck under each effective monitoring point in the collected image based on the pre-collection results; Screen the effective monitoring points according to the preset arrangement quantity based on the effective recording range and position, and position and arrange the camera device based on the screening results.

[0006] Preferably, for an early warning method of a concrete truck water injection detection system, positioning and arranging a camera device based on the screening results includes: Allocate a communication route for the positioned and arranged camera device in the on-vehicle edge device, and determine the bandwidth requirement of the camera device for the communication route based on the service function of the camera device; Adapt the communication parameters of the communication route based on the bandwidth requirement, and perform a collaborative communication test on the camera device and the on-vehicle edge device after the successful adaptation of the communication parameters; After the collaborative communication test is passed, the personalized working parameters of the camera device are adapted based on the image acquisition requirements, and the camera device is authorized to take effect after the adaptation is successful.

[0007] Preferably, in the early warning method of a ready-mixed concrete truck water injection detection system, in step 4, when identifying and analyzing the water injection behavior in the water injection area and performing an alarm operation when there is a water injection behavior, it includes: Obtain the video frame data of each timestamp in the video stream data; Construct a detection model. At the same time, process the detection model in the vehicle-mounted edge device based on the weight parameters and activation values, and deploy the detection model in the vehicle-mounted edge device according to the processing results; Based on the vehicle-mounted edge device, sequentially transmit the video frame data of each timestamp to the detection model for water injection personnel identification; When the detection model detects that there are water injection personnel in the water injection area, perform an alarm operation.

[0008] Preferably, in the early warning method of a ready-mixed concrete truck water injection detection system, constructing a detection model includes: Capture the operation video stream data of the ready-mixed concrete truck at the construction site of multiple environmental categories, and create a training data set of personnel in the ready-mixed concrete truck water injection scenario according to the operation video stream data; Mark the personnel targets in the training data set of personnel in the ready-mixed concrete truck water injection scenario. Among them, the marking content is the position information and category of the visible area of the personnel target; Input the training data set of personnel in the ready-mixed concrete truck water injection scenario into the backbone network to extract image features, and input the image features into the detection head of the target algorithm model to obtain detection targets; According to the detection targets and the marked personnel targets, call the target loss function to optimize the parameters of the backbone network and the detection head of the target algorithm model to obtain the final detection model.

[0009] Preferably, in the early warning method of a ready-mixed concrete truck water injection detection system, when performing water injection personnel identification, it includes: When the detection model detects the existence of target personnel, mark the position information of the existing target personnel to determine the target boxes of all target personnel; Obtain the marked water injection area; Calculate the target area ratio according to the target box and the water injection area; ; Among them, represents the target area ratio; represents the coordinate data of the th target box; represents the coordinate data of the Indicates the overlapping area of the target box and the water injection area; Indicates the width of the th target box; Indicates the height of the i-th target box; Obtain a preset reference threshold, and compare the target area ratio with the preset reference threshold to determine whether the target person is a water injection person; When the target area ratio is greater than the preset reference threshold, it is determined that the target person is a water injection person; Otherwise, it is determined that the target person is an irrelevant person, and the irrelevant person is filtered out in the video data.

[0010] Preferably, for an early warning method of a commercial concrete truck water injection detection system, when the detection model detects a water injection person in the water injection area, an alarm operation is performed, including: When the detection model detects a water injection person in the water injection area, read the current frame video image in which the water injection person is detected, and use the current frame video image as the starting record frame; Obtain a first preset time interval, and perform frame-by-frame statistics on the detection results of the video data within the first preset time interval according to the starting record frame; Determine the first total number of times the water injection person appears in each video frame within the first preset time interval according to the statistical results, and when the first total number of times is equal to or greater than the first preset threshold, trigger the first alarm operation; When the first alarm operation is triggered, obtain a second preset time interval, where the second preset time interval contains multiple first preset time intervals; Record the second total number of times the first alarm operation is triggered within the second preset time interval according to the starting record frame, and when the second total number of times is equal to or greater than the second preset threshold, it is determined that there is a water injection behavior, and a second alarm operation is performed according to the determination result, and the determination result is fed back to the monitoring terminal.

[0011] Preferably, for an early warning method of a commercial concrete truck water injection detection system, in step 5, when the alarm operation is triggered, an upload mechanism is automatically started for data upload, including: When the alarm operation is triggered, read the data to be uploaded, where the data to be uploaded is: a video segment that saves the water injection behavior in the video stream data; record the time point when the water injection behavior occurs, and use the time point as the time stamp of the video segment; locate the real-time geographical location of the commercial concrete truck to determine the geographical information, and at the same time, determine the license plate number of the commercial concrete truck; Start the upload mechanism to upload the data to be uploaded to the management terminal.

[0012] Preferably, for an early warning method of a water injection detection system for a commercial concrete truck, in step 2, video stream data is collected in real time according to the positioning layout result and transmitted to an in-vehicle edge device for image processing, including: Based on the positioning layout result, control the camera device to collect video stream data in real time, and add a data source to the video stream data based on the physical position of the camera device; Transmit the video stream data with the added data source to the in-vehicle edge device, and frame the video stream data based on the in-vehicle edge device to obtain an image frame sequence; At the same time, trace the time of the video stream data, and obtain the time stamp of each image frame based on the time tracing result; Add the time stamp to the corresponding image frame to obtain a target image frame sequence; Traverse the pixels of each image frame in the target image frame sequence in sequence according to the time development order, obtain the pixel features of each image frame, and perform multi-dimensional difference analysis on the pixel features based on a preset processing standard to obtain the preprocessing parameters of each image frame in each dimension; Perform preprocessing on each image frame in sequence based on the preprocessing parameters, where the preprocessing includes illumination equalization, haze removal enhancement, and color correction optimization.

[0013] The present invention provides an early warning device for a water injection detection system for a commercial concrete truck, including: A device positioning and layout module for positioning and arranging a camera device on a commercial concrete truck; An image processing module for collecting video stream data in real time according to the positioning layout result and transmitting it to an in-vehicle edge device for image processing; A calibration module for calibrating the water injection area of the processed image data according to the current working environment; An alarm module for performing identification and analysis of water injection behavior in the water injection area and performing an alarm operation when there is a water injection behavior; A data upload module for automatically starting an upload mechanism to upload data when an alarm operation is triggered.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: By selecting a suitable in-vehicle camera and installation location, real-time monitoring of the water injection area of a commercial concrete truck in different scenarios is realized. By performing identification and analysis of water injection behavior in the water injection area and performing an alarm operation when there is a water injection behavior, the accuracy of water injection behavior identification is effectively improved. When an alarm operation is triggered, an upload mechanism is automatically started to upload data, which can effectively ensure the traceability of data, thereby effectively realizing data management and improving the effectiveness and accuracy of water injection detection for commercial concrete trucks.

[0015] Other features and advantages of the present invention will be set forth in the following description, and in part will be obvious from the description, or can be learned by practice of the present invention. The objectives and other advantages of the present invention can be realized and attained by the structure particularly pointed out in this application document.

[0016] The technical solution of the present invention will be further described in detail below through the drawings and embodiments. Description of the Drawings

[0017] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention, and do not constitute a limitation to the present invention. In the drawings: Figure 1 It is a flowchart of an early warning method for a water injection detection system of a commercial concrete truck in an embodiment of the present invention; Figure 2 It is a logic diagram of an early warning method for a water injection detection system of a commercial concrete truck in an embodiment of the present invention; Figure 3 It is a schematic diagram of the installation position of a camera device in an embodiment of the present invention; Figure 4 It is a schematic diagram of the annotation of a water injection area in an embodiment of the present invention; Figure 5 It is a schematic diagram of alarm statistics in an embodiment of the present invention; Figure 6 It is a structural diagram of an early warning device for a water injection detection system of a commercial concrete truck in an embodiment of the present invention. Detailed Embodiments

[0018] The following describes the preferred embodiments of the present invention with reference to the drawings. It should be understood that the preferred embodiments described herein are only for the purpose of illustrating and explaining the present invention, and are not used to limit the present invention.

[0019] Embodiment 1: This embodiment provides an early warning method for a water injection detection system of a commercial concrete truck, as Figure 1 - Figure 2 shown, including: Step 1: Position and arrange a camera device on the commercial concrete truck; Select a high-resolution camera, install the camera at a position that can fully cover the water injection port area of the commercial concrete truck, and adjust it according to different vehicle models and operating environments, providing a solid foundation for the next step of real-time image acquisition and preprocessing, and water injection area calibration.

[0020] Step 2: Real-time collect video stream data according to the positioning and arrangement results and transmit it to an in-vehicle edge device for image processing; Real-time image acquisition and preprocessing are key steps in the water injection detection system for commercial concrete mixer trucks. The monitoring data of real-time vehicles is obtained through the camera installed in Step 1, and the video stream captured by the camera is transmitted to the low-power on-vehicle edge device stably and with low latency. The on-vehicle edge device is used to preprocess each frame of the image to ensure the quality of the input data.

[0021] Step 3: Calibrate the water injection area for the processed image data according to the current working environment; After installing the camera, the water injection detection system predefines and calibrates the water injection areas on the left and right sides of the commercial concrete mixer truck. After obtaining the real-time input data through Step 2, according to the current working environment, the cloud platform is used to remotely adjust the calibration of the water injection area, and dynamic setting of polygon vertices is supported to improve the accuracy of water injection behavior recognition. The camera installation is as Figure 3 shown.

[0022] Step 4: Conduct water injection behavior recognition and analysis in the water injection area, and perform an alarm operation when there is a water injection behavior; After obtaining the real-time video stream through Step 2, conduct water injection behavior recognition and analysis to determine whether there is a water injection behavior. The water injection behavior recognition first detects the water injection personnel from a single frame of image through a target detection algorithm (an improved YOLOv5 detection algorithm), and then statistically analyzes the detection results of multiple frames within a short period of time. After confirming that there is a water injection behavior within a short period of time, trigger an alarm once. Finally, statistically accumulate the alarm results of multiple short periods of time within a long period of time to output the final water injection behavior recognition result. The example diagram of the water injection area annotation is as Figure 4 shown.

[0023] Step 5: Automatically start the upload mechanism to upload data when the alarm operation is triggered.

[0024] After confirming the existence of a water injection behavior, immediately trigger the warning mechanism and automatically activate the background upload mechanism. Not only upload the video clip of the water injection behavior, but also record key data such as time information, geographical information, and license plate number to achieve a comprehensive record of the uploaded data; these information are integrated into a structured data packet for efficient transmission and storage. Upload the data to the cloud platform for classification and archiving, which is convenient for subsequent retrieval and review. The administrator can view the historical records, export reports, and conduct data analysis through the cloud platform, so as to provide deeper support and optimization for monitoring and management. After processing for a time interval of T seconds, the system waits for the arrival of the next frame of image, and then returns to the steps of image acquisition and preprocessing to continue executing the above process. The whole process forms a closed-loop system to ensure continuous and efficient monitoring and warning capabilities, and maintain the real-time response characteristics of the system.

[0025] The beneficial effects of the above technical solution are as follows: By selecting appropriate vehicle-mounted cameras and installation positions, real-time monitoring of the water injection areas of commercial concrete trucks in different scenarios is achieved. Through the identification and analysis of water injection behaviors in the water injection areas and alarm operations when water injection behaviors exist, the accuracy of water injection behavior identification is effectively improved. When the alarm operation is triggered, the upload mechanism is automatically started for data uploading, which can effectively ensure the traceability of data, thereby effectively realizing the management of data and improving the effectiveness and accuracy of water injection detection for commercial concrete trucks.

[0026] Embodiment 2: Based on Embodiment 1, this embodiment provides an early warning method for a water injection detection system of a commercial concrete truck. In step 1, the positioning and arrangement of the imaging device on the commercial concrete truck include: Obtain the structural characteristics of the commercial concrete truck and determine the water injection area of the commercial concrete truck based on the structural characteristics; Determine the overall position distribution of the water injection area on the commercial concrete truck, and determine the effective monitoring points for image acquisition of the water injection area based on the overall position distribution and structural characteristics; Pre-collect the shooting angles and shooting ranges of each effective monitoring point based on the imaging device, and determine the effective recording ranges and positions of the water injection areas of the commercial concrete truck under each effective monitoring point in the collected images based on the pre-collection results; Screen the effective monitoring points according to the preset arrangement quantity based on the effective recording ranges and positions, and position and arrange the imaging device based on the screening results.

[0027] In this embodiment, the structural characteristics refer to the distribution of the structural components of the commercial concrete truck, so as to facilitate determining the water injection area of the commercial concrete truck through the structural characteristics.

[0028] In this embodiment, the effective monitoring points refer to the position points for image acquisition of the water injection area, that is, the points that can comprehensively and effectively acquire images of the water injection area, and are not unique.

[0029] In this embodiment, the pre-collection refers to determining the shooting angles and shooting ranges of each effective monitoring point through the imaging device.

[0030] In this embodiment, screening the effective monitoring points according to the preset arrangement quantity based on the effective recording ranges and positions means screening out the position points with the best effective recording ranges and positions from the effective monitoring points, that is, screening out the points that can comprehensively collect the water injection area and whose acquisition angles meet the requirements.

[0031] The beneficial effects of the above technical solution are as follows: By analyzing the structural characteristics of the commercial concrete truck, the water injection area can be effectively determined, and then the effective monitoring points can be determined according to the water injection area, and the determined effective monitoring points are screened, ensuring the effective deployment of the camera device and providing guarantee for the detection of the water injection area.

[0032] Embodiment 3: Based on Embodiment 2, this embodiment provides an early warning method for the water injection detection system of the commercial concrete truck. Positioning and arranging the camera device based on the screening result includes: Allocating a communication route for the positioned and arranged camera device in the vehicle-mounted edge device, and determining the bandwidth requirement of the camera device for the communication route based on the service function of the camera device; Adapting the communication parameters of the communication route based on the bandwidth requirement, and conducting a collaborative communication test between the camera device and the vehicle-mounted edge device after the successful adaptation of the communication parameters; After the collaborative communication test passes, determining the personalized working parameter adaptation for the camera device based on the image acquisition requirements, and granting the effective authorization to the camera device after the successful adaptation.

[0033] In this embodiment, the service function refers to the types and requirements of the services executed by the camera device, so as to facilitate determining the bandwidth requirement of the communication route according to the service function.

[0034] In this embodiment, the collaborative communication test refers to testing the communication situation between the camera device and the vehicle-mounted edge device, that is, ensuring that effective communication can be carried out between the two.

[0035] In this embodiment, the personalized working parameter adaptation refers to configuring the resolution, etc. of the camera device according to the camera requirements or image acquisition requirements.

[0036] The beneficial effects of the above technical solution are as follows: By adapting the communication parameters and personalized working parameters of the camera device, the working reliability of the camera device is ensured, and the comprehensiveness and effectiveness of image acquisition in the water injection area are also improved.

[0037] Embodiment 4: Based on Embodiment 1, this embodiment provides an early warning method for the water injection detection system of the commercial concrete truck. In step 4, when identifying and analyzing the water injection behavior in the water injection area and performing an alarm operation when there is a water injection behavior, it includes: Obtaining the video frame data of each timestamp in the video stream data; Constructing a detection model. At the same time, processing the detection model in the vehicle-mounted edge device based on the weight parameters and activation values, and deploying the detection model in the vehicle-mounted edge device according to the processing results; Based on the vehicle-mounted edge device, the video frame data with each timestamp is sequentially transmitted to the detection model for water injection personnel recognition; When the detection model detects a water injection personnel in the water injection area, an alarm operation is performed.

[0038] In this embodiment, the detection model is constructed as follows: capture the operation video stream data of the commercial concrete truck under construction sites of multiple environmental categories, and create a training data set of personnel in the commercial concrete truck water injection scenario based on the operation video stream data; label the personnel targets in the training data set of personnel in the commercial concrete truck water injection scenario, where the labeling content is the position information and category of the visible area of the personnel target; input the training data set of personnel in the commercial concrete truck water injection scenario into the backbone network to extract image features, and input the image features into the detection head of the target algorithm model to obtain detection targets; according to the detection targets and the labeled personnel targets, call the target loss function to optimize the parameters of the backbone network and the detection head of the target algorithm model to obtain the final detection model.

[0039] In this embodiment, the labeling content is the position information and category of the visible area of the personnel target, where the position information of the visible area of the personnel target is: the coordinate information of the minimum bounding rectangle of the visible area of the personnel target. The category is the target label category, and its name is watering_personnel.

[0040] Based on the above: After analyzing the input image using the improved YOLOv5 algorithm, the real water injection target of the current video frame is identified through the calibration area, and the water injection personnel information is output. Before using the improved YOLOv5 algorithm, it is necessary to collect real vehicle scene samples and train the improved YOLOv5 algorithm model. After training, the model is deployed on the edge device using the asymmetric quantization method. The image data in step 3.2 is sent to the edge device, and the quantized model is used to analyze and detect the image data to initially obtain the position information of all detected personnel, and further filter the targets in the non-calibration area. The specific steps are as follows: Collect the data set: Since there are significant differences between the personnel on the commercial concrete truck and the pedestrian data in the public data set, and there is a phenomenon that personnel are partially blocked by the concrete body, it is necessary to collect data in this scenario for model fine-tuning. In the collection stage, select real construction sites, etc. as the data collection locations, covering extreme conditions such as day and night, rain, fog, and dust, to ensure the diversity and representativeness of the data set. Use the vehicle-mounted camera to capture the video stream during the operation of the commercial concrete truck in a specific manner, covering at least 336 hours of construction site operations, and create a detection data set of personnel in the commercial concrete truck water injection scenario. In the collected video stream, all the appeared personnel targets are manually labeled, with a total of 45100 targets labeled, and approximately 90% of the personnel have varying degrees of occlusion. Each label records the position information and category of the visible area of the personnel target.

[0041] Model Training: The training samples are sent to the ShuffleNetV2 backbone network to extract image feature information. Then, the feature information is sent to the detection head of the improved YOLOv5 algorithm model to obtain the detection targets. Finally, the improved Focal-EIoU loss function is used to optimize the model parameters, and the final trained model is obtained. IoU is a basic metric for measuring the overlap between the predicted bounding box and the ground truth bounding box, representing the ratio of the intersection and union between the predicted bounding box and the ground truth bounding box. EIoU and Focal-EIoU are both extended forms of IoU. The formula of the Focal-EIoU loss function is as follows: Where represents the loss function of Focal-EIoU, is a hyperparameter used to control the curve radian, represents the loss function of EIoU; and represent the coordinates of the detected bounding box and the ground truth bounding box respectively; represents the Euclidean distance, represent the centers of the predicted bounding box and the ground truth bounding box respectively, represents the diagonal distance of the minimum bounding rectangle, represent the widths of the predicted bounding box and the ground truth bounding box respectively, represents the width of the minimum bounding rectangle, represent the heights of the predicted bounding box and the ground truth bounding box respectively, represents the height of the minimum bounding rectangle.

[0042] In this embodiment, when deploying on edge devices, it includes: by obtaining the original trained model, asymmetric quantization processing is performed on the weight parameters and activation values during deployment on edge devices to further improve the computing efficiency and reduce resource occupancy while maintaining high precision. Specifically, the asymmetric quantization method allows for more flexible mapping of floating-point values to the integer range during quantization, thus effectively reducing quantization errors and maximizing the preservation of the accuracy of the original model; in the forward inference stage, 12-bit integer operations are retained to replace the traditional 8-bit quantization scheme, avoiding the problem of a large drop in accuracy caused by 8-bit quantization. The 12-bit quantization strategy provides higher numerical representation accuracy, controlling the accuracy loss after model quantization within a very small range and ensuring high accuracy of the model in practical applications. This scheme provides a high-cost-effective algorithm support for the water injection detection of commercial concrete trucks while minimizing the accuracy loss after model quantization and still meeting the real-time requirements.

[0043] In this embodiment, the timestamp video frame data refers to the frame data obtained by splitting the video stream data according to the time series, and the frame data corresponds one-to-one with the time series.

[0044] In this embodiment, the activation value refers to the configuration parameter that needs to be performed when the detection model is activated.

[0045] The beneficial effects of the above technical solution are: by splitting the video stream data into each timestamp video frame data, and inputting each timestamp video frame data into the constructed detection model for water injection personnel identification, thereby realizing accurate and effective identification of whether there are water injection personnel in the water injection area, and performing corresponding alarm operations when there are water injection personnel, improving the accuracy and reliability of the water injection detection of the commercial concrete truck.

[0046] Embodiment 5: Based on Embodiment 4, this embodiment provides a warning method for a water injection detection system of a commercial concrete truck. When identifying water injection personnel, it includes: When the detection model detects the existence of a target person, mark the location information of the existing target person and determine the target boxes of all target persons; Obtain the marked water injection area; Calculate the target area ratio according to the target box and the water injection area; ; Wherein, represents the target area ratio; represents the coordinate data of the th target box; represents the coordinate data of the th marked area; represents the overlapping area of the intersection of the target box and the water injection area; represents the width of the th target box; Obtain a preset reference threshold, and compare the target area ratio with the preset reference threshold to determine whether the target person is a water injection person; When the target area ratio is greater than the preset reference threshold, it is determined that the target person is a water injection person; Otherwise, it is determined that the target person is an irrelevant person, and the irrelevant person is filtered out in the video data.

[0047] In this embodiment, the water injection personnel filtering is a method of filtering out irrelevant personnel by using the threshold judgment of the overlapping area ratio calculated by the calibration area after obtaining the detection result based on the improved object detection. In the picture taken by the camera, often when the water injection personnel are photographed, other pedestrians on the road surface will also be in the picture, so it is necessary to filter out the irrelevant personnel.

[0048] In this embodiment, the target box refers to the marked box obtained after marking the target person.

[0049] In this embodiment, the preset reference threshold is set in advance and is used as the measurement basis for measuring whether the target area ratio meets the requirements.

[0050] The beneficial effects of the above technical solution are as follows: By marking the target persons existing in the water injection area and judging whether the target persons are water injection persons according to the marking results, and filtering out the irrelevant persons when the target persons are irrelevant persons, the accuracy and reliability of the water injection detection of the commercial concrete truck are ensured.

[0051] Embodiment 6: Based on Embodiment 4, this embodiment provides an early warning method for a water injection detection system of a commercial concrete truck. When the detection model detects that there are water injection persons in the water injection area, an alarm operation is performed, including: When the detection model detects that there are water injection persons in the water injection area, read the current frame video image in which the water injection persons are detected, and use the current frame video image as the starting record frame; Obtain the first preset time interval, and perform frame-by-frame statistics on the detection results of the video data within the first preset time interval according to the starting record frame; Determine the first total number of times the water injection persons appear in each video frame within the first preset time interval according to the statistical results, and when the first total number of times is equal to or greater than the first preset threshold, trigger the first alarm operation; When the first alarm operation is triggered, obtain the second preset time interval, where the second preset time interval includes multiple first preset time intervals; Record the second total number of times the first alarm operation is triggered within the second preset time interval according to the starting record frame, and when the second total number of times is equal to or greater than the second preset threshold, determine that there is a water injection behavior, perform a second alarm operation according to the determination result, and feedback the determination result to the monitoring terminal.

[0052] In this embodiment, multi-frame detection and statistics are performed in a short time (i.e., the first preset time interval). The short-time multi-frame detection and statistics is a method of performing multi-frame statistics based on the results of single-frame detection to obtain a stable result output. When the number of frames containing water injection personnel detected reaches the set threshold, it is considered that there is a water injection behavior this time. In a short time (e.g., 1 second), the system sequentially inputs 6 frames of data into the system at equal intervals, and counts the number of frames detected containing water injection personnel. When the number of detected frames reaches or exceeds 4 frames (i.e., the first preset threshold), it is considered that there is a water injection behavior in a short time, and an alarm is triggered. Long-time (i.e., the second preset time interval) cumulative alarm statistics: In a longer time period, the number of alarms in multiple consecutive short times is cumulatively counted. When the set threshold is reached or exceeded, the system determines that a water injection behavior has been detected and issues a final water injection behavior alarm. For example, if the long time is set to 5 seconds, and the number of alarms per 1 second in the short time is cumulatively counted. When the detected long time reaches 5 seconds (i.e., the second preset threshold), and the number of alarms reaches or exceeds 4 times, the system will finally confirm the existence of a water injection behavior and output a reliable water injection detection result. The alarm statistics are specifically as Figure 5 shown.

[0053] In this embodiment, the first preset time interval is set in advance.

[0054] In this embodiment, the first target total number refers to the total number of times the water injection personnel appear in all video frames within the first preset time zone.

[0055] In this embodiment, the first preset threshold is set in advance and is used as a reference for measuring whether the first alarm is required.

[0056] In this embodiment, the second preset time interval is set in advance.

[0057] In this embodiment, the second preset threshold is set in advance and is used to measure whether there is a water injection behavior of the water injection personnel.

[0058] The beneficial effects of the above technical solution are as follows: By determining the number of times the detected water injection personnel appear in the water injection area, when the number of appearances is greater than the preset threshold, the first alarm operation is performed, and at the same time, it is judged whether the water injection personnel have a water injection behavior. Furthermore, when there is a water injection behavior, the second alarm is issued, which improves the rigor and reliability of the water injection detection of the commercial concrete truck.

[0059] Embodiment 7: On the basis of Embodiment 1, this embodiment provides an early warning method for a commercial concrete truck water injection detection system. In step 5, when the alarm operation is triggered, the upload mechanism is automatically started for data upload, including: When an alarm operation is triggered, read the data to be uploaded. Among them, the data to be uploaded is: a video clip of the water injection behavior saved in the video stream data; record the time point when the water injection behavior occurs, and use the time point as the timestamp of the video clip; locate the real-time geographical location of the commercial concrete truck, determine the geographical information, and at the same time, determine the license plate number of the commercial concrete truck; Start the upload mechanism to upload the data to be uploaded to the management terminal.

[0060] In this embodiment, after confirming the existence of the water injection behavior, the early warning mechanism is immediately triggered, and the background upload mechanism is automatically started. Not only the video clip of the water injection behavior is uploaded, but also the key data such as time information, geographical information and license plate number are recorded, realizing a comprehensive record of the uploaded data, specifically including the following: (1) Saving the video clip of the water injection behavior not only provides intuitive visual evidence, but also provides rich materials for subsequent detailed review and analysis; (2) Record the exact time when the event occurs to ensure that the timestamp of each video clip is accurate; (3) Store geographical information, obtain and mark the real-time geographical location information of the commercial concrete truck through GPS or other positioning technologies, which is convenient for tracking and managing the behavior of the vehicle at different locations; (4) Identify the license plate number of the commercial concrete truck, which is convenient for subsequent management and responsibility tracing.

[0061] The beneficial effects of the above technical solutions are: ensuring that the data to be uploaded corresponding to the alarm operation is accurately and effectively uploaded to the management terminal, so as to facilitate the traceability of the alarm information and ensure the rigor of the water injection detection of the commercial concrete truck.

[0062] Embodiment 8: On the basis of Embodiment 1, this embodiment provides an early warning method for a water injection detection system of a commercial concrete truck. In step 2, the video stream data is collected in real time according to the positioning layout result and transmitted to the vehicle-mounted edge device for image processing, including: Control the camera device to collect video stream data in real time based on the positioning layout result, and add a data source to the video stream data based on the physical position of the camera device; Transmit the video stream data with the added data source to the vehicle-mounted edge device, and frame the video stream data based on the vehicle-mounted edge device to obtain an image frame sequence; At the same time, perform time traceability on the video stream data, and obtain the timestamp of each image frame based on the time traceability result; Add the timestamp to the corresponding image frame to obtain a target image frame sequence; Traverse the pixels of each image frame in the target image frame sequence in sequence according to the time development order to obtain the pixel features of each image frame, and perform multi-dimensional difference analysis on the pixel features based on a preset processing standard to obtain the preprocessing parameters of each image frame in each dimension; Perform preprocessing on each image frame in sequence based on the preprocessing parameters, where the preprocessing includes illumination equalization, defogging enhancement, and color correction optimization.

[0063] In this embodiment, adding a data source to the video stream data based on the physical position of the imaging device refers to differentiating the terminals of the video stream data, that is, being able to determine the imaging devices corresponding to different video stream data.

[0064] In this embodiment, the target image frame sequence refers to the result obtained after adding the specific time corresponding to each frame of image to the corresponding image frame.

[0065] In this embodiment, the pixel features refer to the pixel values on each image frame and the corresponding resolution, etc.

[0066] In this embodiment, the preset processing standard is set in advance, including the analysis dimensions of the pixel features and the corresponding analysis standards.

[0067] In this embodiment, the preprocessing parameters refer to the specific parameters that need to be processed for each image frame in each dimension, such as the adjustment amount of the image brightness and the adjustment amount of the resolution, etc.

[0068] The beneficial effects of the above technical solution are: by transmitting the collected video stream data to the vehicle-mounted edge device and analyzing and processing the video stream data through the vehicle-mounted edge device, accurate and effective determination of the preprocessing parameters of each image frame in each dimension is achieved, so as to realize reliable preprocessing of each image frame, providing convenience and basis for the water injection detection of the commercial concrete truck.

[0069] Embodiment 9: Based on Embodiment 1, this embodiment provides an early warning method for a water injection detection system of a commercial concrete truck. In step 3, calibrate the water injection area for the processed image data according to the current working environment, including: Obtain historical image data and analyze the historical image data to obtain the vehicle type set and working environment set of the commercial concrete truck; Determine the corresponding reference water injection area from the preset instruction manual based on the vehicle type of the commercial concrete truck. At the same time, determine the behavioral characteristics of the user when there is water injection based on the water injection detection requirements of the commercial concrete truck, and determine the activity range of the user under the reference water injection area based on the behavioral characteristics; Perform a first correction on the reference water injection area based on the activity range, and determine the first water injection area in the image data based on the first correction result; Meanwhile, analyze the working environment to obtain the interference factors for the first water injection area under different working environments, and determine the interference range of the interference factors on the first water injection area; Perform a second correction on the first water injection area based on the interference range, and obtain a second water injection area based on the second correction; Determine the relative positions between different object targets in the second water injection area, and determine the effective monitoring range in the second water injection area based on the relative positions; Determine the dynamic configuration strategy of polygon vertices under different working environments based on the effective monitoring range; Lock the water injection area in the processed image data according to the current working environment based on the dynamic configuration strategy of polygon vertices, and perform calibration based on the locking result.

[0070] In this embodiment, a preset instruction manual is known in advance and is used to record the water injection areas corresponding to commercial concrete trucks of different vehicle models.

[0071] In this embodiment, the water injection detection requirements for commercial concrete trucks are known in advance.

[0072] In this embodiment, the behavioral characteristics refer to the specific behavioral actions of the user when the user performs water injection.

[0073] In this embodiment, the interference factors refer to the influencing factors existing when determining the first water injection area under different working environments, such as obstacles, etc.

[0074] In this embodiment, the dynamic configuration strategy of polygon vertices refers to the strategy or scheme for determining the water injection area under different working environments, considering the enclosed range. For example, when there are obstacles interfering with the water injection area, the monitoring range of the water injection area can be adjusted according to the shape of the obstacles.

[0075] The beneficial effects of the above technical solution are: By analyzing historical image data, the vehicle model set and working environment set of commercial concrete trucks are determined, and then different vehicle models and working environments are analyzed to effectively formulate the dynamic configuration strategy of polygon vertices of commercial concrete trucks under different working environments. Finally, the water injection area in the processed image data is locked according to the dynamic configuration strategy of polygon vertices for the current working environment, ensuring the accuracy and reliability of the locking of the water injection area.

[0076] Embodiment 10: This embodiment provides an early warning device for a water injection detection system of a commercial concrete truck, as Figure 6 shown, including: A device positioning and arrangement module for positioning and arranging a camera device on a commercial concrete truck; An image processing module, configured to collect video stream data in real time according to the positioning layout result and transmit it to the vehicle-mounted edge device for image processing; A calibration module, configured to calibrate the water injection area for the processed image data according to the current working environment; An alarm module, configured to perform water injection behavior recognition and analysis in the water injection area, and perform an alarm operation when there is a water injection behavior; A data upload module, configured to automatically start an upload mechanism to upload data when an alarm operation is triggered.

[0077] The beneficial effects of the above technical solution are as follows: By selecting appropriate vehicle-mounted cameras and installation positions, real-time monitoring of the water injection area of the commercial concrete truck in different scenarios is achieved. By performing water injection behavior recognition and analysis in the water injection area and performing an alarm operation when there is a water injection behavior, the accuracy of water injection behavior recognition is effectively improved. When an alarm operation is triggered, the upload mechanism is automatically started to upload data, which can effectively ensure the traceability of the data, thereby effectively realizing the management of the data and improving the effectiveness and accuracy of the water injection detection of the commercial concrete truck.

[0078] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these changes and modifications.

Claims

1. An early warning method for a water injection detection system of a commercial concrete truck, characterized in that, Including: Step 1: Position and arrange the camera device on the commercial concrete truck; Step 2: Real-time collect video stream data according to the positioning and arrangement results and transmit it to the vehicle-mounted edge device for image processing; Step 3: Calibrate the water injection area of the processed image data according to the current working environment; Step 4: Conduct water injection behavior recognition and analysis in the water injection area, and perform an alarm operation when there is a water injection behavior; Step 5: Automatically start the upload mechanism to upload data when the alarm operation is triggered.

2. The warning method of a water injection detection system for a commercial concrete truck according to claim 1, characterized in that, In Step 1, positioning and arranging the camera device on the commercial concrete truck includes: Obtain the structural characteristics of the commercial concrete truck, and determine the water injection area of the commercial concrete truck based on the structural characteristics; Determine the overall position distribution of the water injection area on the commercial concrete truck, and determine the effective monitoring points for image acquisition of the water injection area based on the overall position distribution and structural characteristics; Pre-collect the shooting angles and shooting ranges of each effective monitoring point based on the camera device, and determine the effective recording range and position of the water injection area of the commercial concrete truck under each effective monitoring point in the collected image based on the pre-collection results; Screen the effective monitoring points according to the preset arrangement quantity based on the effective recording range and position, and position and arrange the camera device based on the screening results.

3. The warning method of a concrete truck water injection detection system according to claim 2, characterized in that, Positioning and arranging the camera device based on the screening results includes: Allocate a communication route for the positioned and arranged camera device in the vehicle-mounted edge device, and determine the bandwidth requirement of the camera device for the communication route based on the business function of the camera device; Adapt the communication parameters of the communication route based on the bandwidth requirement, and conduct a collaborative communication test between the camera device and the vehicle-mounted edge device after the successful adaptation of the communication parameters; After the collaborative communication test passes, determine the personalized working parameter adaptation for the camera device based on the image acquisition requirements, and perform an effective authorization for the camera device after the successful adaptation.

4. The warning method of a water injection detection system for a commercial concrete truck according to claim 1, characterized in that, In Step 4, conducting water injection behavior recognition and analysis in the water injection area, and performing an alarm operation when there is a water injection behavior includes: Obtain the video frame data of each timestamp in the video stream data; Construct a detection model. At the same time, process the detection model in the vehicle-mounted edge device based on the weight parameters and activation values, and deploy the detection model in the vehicle-mounted edge device according to the processing results; Based on the vehicle-mounted edge device, sequentially transmit the video frame data of each timestamp to the detection model for water injection personnel recognition; When the detection model detects that there are water injection personnel in the water injection area, perform an alarm operation.

5. The warning method of a warning system for water injection detection of a commercial concrete truck according to claim 4, wherein, Constructing a detection model includes: Capture the operation video stream data of the commercial concrete truck at construction sites with multiple environmental categories, and create a training data set of personnel in the water injection scenario of the commercial concrete truck according to the operation video stream data; Mark the personnel targets in the training data set of personnel in the water injection scenario of the commercial concrete truck. Among them, the marking content is the visible area position information and category of the personnel targets; Input the training data set of personnel in the water injection scenario of the commercial concrete truck into the backbone network to extract image features, and input the image features into the detection head of the target algorithm model to obtain detection targets; Adjust the parameters of the backbone network and the detection head of the target algorithm model according to the detection targets and the marked personnel targets by invoking the target loss function to obtain the final detection model.

6. The warning method of a water injection detection system for a commercial concrete truck according to claim 4, characterized in that, When conducting water injection personnel recognition, it includes: When the detection model detects the presence of target personnel, mark the location information of the existing target personnel to determine the target boxes of all target personnel; Obtain the marked water injection area; Calculate the target area ratio based on the target box and the water injection area; ； Among them, represents the target area ratio; represents the coordinate data of the th target bounding box; represents the coordinate data of the th labeled area; represents the overlapping area of the target bounding box and the water injection area; represents the width of the th target bounding box; represents the height of the ith target bounding box; Obtain a preset reference threshold, compare the target area ratio with the preset reference threshold, and determine whether the target personnel are water injection personnel; When the target area ratio is greater than the preset reference threshold, determine that the target personnel are water injection personnel; Otherwise, determine that the target personnel are irrelevant personnel and filter out the irrelevant personnel from the video data.

7. The warning method of a warning method for a water injection detection system of a commercial concrete truck according to claim 4, characterized in that, When the detection model detects the presence of water injection personnel in the water injection area, perform an alarm operation, including: When the detection model detects the presence of water injection personnel in the water injection area, read the current frame video image in which the water injection personnel are detected, and use the current frame video image as the starting record frame; Obtain a first preset time interval, and perform frame-by-frame statistics on the detection results of the video data within the first preset time interval based on the starting record frame; Determine the first total number of times the water injection personnel appear in each video frame within the first preset time interval according to the statistical results, and when the first total number of times is equal to or greater than the first preset threshold, trigger the first alarm operation; When the first alarm operation is triggered, obtain a second preset time interval, where the second preset time interval contains multiple first preset time intervals; Record the second total number of times the first alarm operation is triggered within the second preset time interval based on the starting record frame, and when the second total number of times is equal to or greater than the second preset threshold, determine that there is a water injection behavior, perform a second alarm operation according to the determination result, and feedback the determination result to the monitoring terminal.

8. The warning method of a water injection detection system for a commercial concrete truck according to claim 1, wherein, In step 5, when the alarm operation is triggered, automatically start the upload mechanism to upload data, including: When the alarm operation is triggered, read the data to be uploaded, where the data to be uploaded is: the video segment that saves the water injection behavior in the video stream data; record the time point when the water injection behavior occurs, and use the time point as the timestamp of the video segment; locate the real-time geographical location of the commercial concrete truck, determine the geographical information, and at the same time, determine the license plate number of the commercial concrete truck; Start the upload mechanism to upload the data to be uploaded to the management terminal.

9. The warning method of a warning system for water injection detection of a commercial concrete truck according to claim 1, characterized in that, In step 2, collect video stream data in real time according to the positioning layout result and transmit it to the vehicle-mounted edge device for image processing, including: Based on the positioning layout result, control the camera device to collect video stream data in real time, and add a data source to the video stream data based on the physical position of the camera device; Transmit the video stream data with the added data source to the vehicle-mounted edge device, and based on the vehicle-mounted edge device, frame the video stream data to obtain an image frame sequence; At the same time, perform time tracing on the video stream data, and obtain the timestamp of each image frame based on the time tracing result; Add the timestamp to the corresponding image frame to obtain a target image frame sequence; Traverse the pixels of each image frame in the target image frame sequence in sequence based on the time development order to obtain the pixel features of each image frame, and perform multi-dimensional difference analysis on the pixel features based on a preset processing standard to obtain the preprocessing parameters of each image frame in each dimension; Preprocess each image frame in sequence based on preprocessing parameters, where the preprocessing includes illumination equalization, defogging enhancement, and color correction optimization.

10. An early warning device for a water injection detection system of a commercial concrete truck, characterized in that, It includes: The device positioning and arrangement module is used to position and arrange the imaging device on the commercial concrete truck; The image processing module is used to collect video stream data in real time according to the positioning and arrangement results and transmit it to the vehicle-mounted edge device for image processing; The calibration module is used to calibrate the water injection area of the processed image data according to the current working environment; The alarm module is used to perform water injection behavior recognition and analysis in the water injection area and perform alarm operations when there is a water injection behavior; The data upload module is used to automatically start the upload mechanism for data upload when the alarm operation is triggered.

Images