A safety control system and method for automobile environmental wind tunnel rainfall test

Abstract

The present invention belongs to the technical field of automobile environmental testing, and in particular, relates to a safety control system and method for automobile environmental wind tunnel rainfall test, wherein the automobile environmental wind tunnel comprises an environmental adjustment module and a wind tunnel operating system, wherein the environmental adjustment module communicates with the wind tunnel operating system, and the wind tunnel operating system controls the running state of the environmental adjustment module; the safety control system comprises a monitoring module, an alarm module and a safety control module; the wind tunnel operating system is connected in communication with the safety control module; the monitoring module and the alarm module are connected in communication with the safety control module, the safety control module receives the monitoring signal of the monitoring module, and analyzes the monitoring signal according to a preset safety risk monitoring model to generate a risk result, and the safety control module is also used to control the wind tunnel operating system according to the risk result to adjust the running state of the environmental adjustment module; the safety control module is also used to control the running state of the alarm module according to the risk result. The present invention can solve the problem that the existing environmental wind tunnel A-pillar overflow test and wheel water throwing test lack safety assurance.

Description

Technical Field

[0001] The invention belongs to the technical field of automobile environmental testing, and in particular relates to a safety control system and method for automobile environmental wind tunnel rainfall testing. Background Art

[0002] The automotive environmental wind tunnel is a very important test room in automobile research and development. It can simulate the real climate environment of the outside world. The temperature can be simulated at -40℃-60℃, the humidity can be simulated at 5%-95% relative humidity, the wind speed can be simulated at 0-200km / h, and rain, snow and sunshine can be simulated.

[0003] With the development of the automobile industry, people are paying attention to the safety, power economy, and comfort of vehicles while also paying attention to the cleanliness of the vehicle body surface. Therefore, based on the rainfall capacity of the environmental wind tunnel, the A-pillar overflow test was introduced in terms of vehicle driving safety, and the wheel water-throwing test was introduced in terms of vehicle surface cleanliness. In the A-pillar overflow test and the wheel water-throwing test, fluorescent agents are dissolved in water for easy observation, and fluorescent lamps are used for development. Therefore, the entire test process is in a darkroom environment in the laboratory, accompanied by high wind speed, high vehicle speed and humid environment of rain. Because the test darkroom environment is too complicated, how to ensure the safety of personnel and equipment during the test has become a major problem facing the A-pillar overflow test and the wheel water-throwing test.

[0004] At present, the safety control of the A-pillar overflow test and the wheel water-throwing test process mainly adopts the subjective control of the test personnel. Since the test process is in a dark room environment and the ambient noise is relatively large, it brings great obstacles to the subjective control of the personnel and is easy to cause misjudgment by the personnel, which will not only affect the progress of the test but also increase the safety risks of personnel and equipment. Therefore, there is an urgent need for a safety control system to supplement the safety monitoring of the test process and ensure the safety of personnel and equipment during the test. Summary of the invention

[0005] The technical problem solved by the present invention is to provide a safety control system and method for an automobile environmental wind tunnel rainfall test, so as to solve the problem that the existing environmental wind tunnel A-pillar overflow test and wheel water throwing test lack safety guarantee.

[0006] The basic scheme provided by the present invention is: a safety control system for automobile environmental wind tunnel rainfall test, wherein the automobile environmental wind tunnel comprises an environmental adjustment module and a wind tunnel operating system, wherein the environmental adjustment module communicates with the wind tunnel operating system, and the wind tunnel operating system controls the running state of the environmental adjustment module; the safety control system comprises a monitoring module, an alarm module and a safety control module; and the wind tunnel operating system is in communication connection with the safety control module;

[0007] The monitoring module and the alarm module are communicatively connected to the safety control module. The safety control module receives the monitoring signal of the monitoring module, analyzes the monitoring signal according to a preset safety risk monitoring model, and generates a risk result. The safety control module is also used to control the wind tunnel operating system according to the risk result to adjust the operating state of the environmental adjustment module; the safety control module is also used to control the operating state of the alarm module according to the risk result.

[0008] Furthermore, the monitoring module includes a monitoring camera unit and a circuit monitoring unit. The monitoring camera unit covers the flow channel area of ​​the wind tunnel, and the circuit monitoring unit monitors the real-time current and voltage of the circuit.

[0009] Furthermore, the wind tunnel operating system is also used to transmit the parameters of the environment adjustment module when it is in operation to the safety control module. The safety control module is also used to receive the parameters and determine whether the parameters are within a preset range. If not, the safety control module controls the wind tunnel operating system to adjust the operation state of the environment adjustment module.

[0010] Furthermore, the security risk monitoring model preset in the security control module is specifically:

[0011] Obtain historical data from automobile wind tunnel rainfall tests, perform model training after preprocessing, and generate a safety risk monitoring model suitable for automobile wind tunnel environments;

[0012] The security risk monitoring model obtains the video data collected by the monitoring camera unit and performs risk analysis on the video data.

[0013] Furthermore, the security risk monitoring model obtains the video data collected by the monitoring camera unit, and performs risk analysis on the video data as follows:

[0014] Read the video data, set the frame number threshold according to the stability characteristics of the video data in the flow channel area of ​​the automobile wind tunnel before the rainfall test, and establish the test background image;

[0015] Obtain the difference image after subtracting the current frame of the video data from the background, generate multiple connected regions after performing morphological operations, and exclude fixed facilities in the automobile wind tunnel according to the area of ​​the connected regions to generate a binary image of the region;

[0016] The remaining area is stored in a suspicious object queue, and the suspicious objects in the suspicious object queue are continuously tracked in subsequent frames;

[0017] By calculating the distance between different object coordinates, the matching degree between grayscale images and the matching degree between Rain moments, the number of frames in which each suspicious object in the suspicious object queue exists and the number of frames in which it disappears are counted;

[0018] If a new object is detected in a blank area, it will be added to the suspicious object queue and tracked continuously. If the object persists and its coordinates change, it will be judged as an area intrusion, and the type of the intruder will be determined based on the preset AI algorithm model.

[0019] If the number of frames where the suspicious object disappears continuously exceeds the preset frame number threshold, it will be removed from the suspicious object queue;

[0020] If the number of frames in which a suspicious object exists exceeds the preset frame number threshold, the suspicious object is judged to be a relic, and the preset AI algorithm model is called to identify the relic.

[0021] Furthermore, the distance calculation formula of the different object coordinates is:

[0022]

[0023] The calculation formula for the matching degree between the grayscale images is:

[0024]

[0025] The calculation formula for the matching degree between Rain moments is:

[0026]

[0027] Where N is the number of objects, C j is the horizontal or vertical coordinate of the jth object, C i is the horizontal or vertical coordinate of the i-th pixel of the j-th object, n is the number of pixels of the j-th object, p and q are the two objects being compared; k represents the number of Rain moment components, represents the kth component of the Rain moment of object p, represents the kth component of the Rain moment of object q, m represents the range of the grayscale histogram, s represents the number of pixels in the grayscale histogram, Indicates the number of pixels with S grayscale in the grayscale histogram of object p, Represents the number of pixels with S grayscale in the grayscale histogram of object q.

[0028] Furthermore, the preset AI algorithm model adopts a wind tunnel object feature recognition model, specifically:

[0029] Divide the dataset of feature maps of objects to be identified into a training set and a test set;

[0030] A wind tunnel object feature recognition model is constructed based on the YOLO model, and a residual network and a WCA module are added to the wind tunnel object feature recognition model; the WCA module performs a global average pooling operation on the input object feature map; then performs a one-dimensional convolution operation with a convolution kernel size of k, and obtains the weight of each channel through an activation function;

[0031] The EOUC loss function is introduced, and the wind tunnel object feature recognition model is trained using the training set;

[0032] After the training is completed, the test set is input into the wind tunnel object feature recognition model and the results are output.

[0033] Furthermore, the EOUC loss function is expressed as:

[0034]

[0035] Among them, L dis is the center distance loss between the predicted box and the real box, L IOU is the overlap loss between the predicted box and the true box, L asp is the width and height loss of the predicted box and the real box. IOU is the accuracy standard for detecting corresponding objects in a specific data set. w, h, d are weights, and p is the activation function. and are the minimum bounding rectangle height and width of the real box and the predicted box, respectively, C 2 for and The sum of squares, d gt Indicates the coordinates of the center point of the target box, w gt Indicates the target box width information, h gt Indicates the target box height information.

[0036] A safety control method for an automobile environmental wind tunnel rainfall test is applied to the above-mentioned safety control system for an automobile environmental wind tunnel rainfall test.

[0037] The principle and advantage of the present invention is that the safety control module receives the monitoring signal in the automobile wind tunnel, including video data and circuit information data, and the circuit information data can be obtained by judging whether the circuit is abnormal. The video data is identified according to the image recognition algorithm. The image recognition algorithm of the present application identifies and judges the detected suspicious objects, and can effectively obtain the recognition results. Finally, according to the recognition results and the judgment results, an alarm is issued for the abnormality in the automobile wind tunnel, so as to remind the staff to deal with it in time. Therefore, the advantages of the present application are:

[0038] 1. The safety control module of this application is controlled by the signal of the monitoring module, which can effectively prevent the system from mistakenly starting the safety control module during other tests in the wind tunnel, and false alarms will affect the conduct of other tests;

[0039] 2. Before and after the wind tunnel operating system is started, the wind tunnel test environment risks are checked through the safety control module, and risks are eliminated in advance and continuously in advance to reduce the probability of test interruption. The wind tunnel operating system can be controlled in time after the risks are found, effectively preventing test personnel from starting the wind tunnel operating system without eliminating the risk items and causing test safety accidents;

[0040] 3. When the wind tunnel operating system controls the operation of the environment adjustment module, the operating parameters are transmitted to the safety control module to determine whether the operating parameters are abnormal and the risk factor of the abnormality. If the risk factor is small, the equipment does not need to be started or stopped. This can reduce test interruptions caused by the system determining wind tunnel flow abnormalities under low-risk conditions and save test resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a structural schematic diagram of an embodiment of the present invention;

[0042] Figure 2 Schematic diagram of a flow chart of an embodiment of the present invention. DETAILED DESCRIPTION

[0043] The following is further described in detail through specific implementation methods:

[0044] The symbols in the drawings of the specification include: monitoring camera unit 1, rainfall water supply system 2, alarm module 3, circuit monitoring unit 4, rainfall sprinkler 5, fan 6, safety control module 7, wind tunnel operating system 8.

[0045] The embodiment is basically as follows Figure 1 and Figure 2 As shown: a safety control system for an automobile environmental wind tunnel rainfall test, wherein the automobile environmental wind tunnel includes an environmental adjustment module and a wind tunnel operating system 8. In this embodiment, the environmental adjustment module includes a rainfall water supply system 2, a rainfall nozzle 5 and a fan 6, wherein the rainfall water supply system 2 is connected to the rainfall nozzle 5, and the rainfall nozzle 5 is installed at the outlet of the fan 6. The rainfall water supply system 2 provides the rainfall nozzle 5 with simulated rainwater required for the rainfall test, and then the fan 6 simulates the production of wind to achieve a simulated automobile environmental wind tunnel rainfall test.

[0046] The rainfall water supply system 2 and the fan 6 in the environmental adjustment module are communicatively connected with the wind tunnel operating system 8, and the operating states of the rainfall water supply system 2 and the fan 6 can be adjusted through the wind tunnel operating system 8.

[0047] The safety control system includes a monitoring module, an alarm module 3 and a safety control module 7. For an automotive environmental wind tunnel, the monitoring module is installed in the flow field of the wind tunnel, specifically at the top of the wind tunnel, the alarm module 3 is installed outside the wind tunnel flow field, and the safety control module 7 and the wind tunnel operating system 8 are located outside the wind tunnel.

[0048] The monitoring module and the alarm module 3 are connected to the safety control module 7, and the wind tunnel operating system 8 is connected to the safety control module 7. The safety control module 7 receives the monitoring signal of the monitoring module, and analyzes the monitoring signal according to the preset safety risk monitoring model to generate a risk result. The safety control module 7 is also used to control the wind tunnel operating system 8 according to the risk result to adjust the operating state of the environment adjustment module; the safety control module 7 is also used to control the operating state of the alarm module 3 according to the risk result; specifically:

[0049] The monitoring module includes a monitoring camera unit 1 and a circuit monitoring unit 4. The monitoring camera unit 1 covers the wind tunnel flow channel area and is used to monitor whether there are suspicious objects in the wind tunnel flow channel and generate video data. The circuit monitoring unit 4 is used to monitor the real-time current and voltage of the circuit in the wind tunnel and generate a circuit monitoring signal.

[0050] The video data of the monitoring camera unit 1 is transmitted to the safety control module 7 for risk analysis. The safety control module 7 is preset with a safety risk monitoring model, which is constructed using a YOLO network model. The historical data during the automobile wind tunnel rainfall test is obtained and transmitted to the initial model for training after preprocessing to generate a safety risk monitoring model suitable for the automobile wind tunnel environment;

[0051] The security risk monitoring model then obtains the video data collected by the monitoring camera unit 1 and performs risk analysis on the video data. The specific analysis process is as follows:

[0052] Read the video data, set the frame number threshold according to the stability characteristics of the video data in the flow channel area of ​​the automobile wind tunnel before the rainfall test, and establish the test background image;

[0053] Obtain the difference image after subtracting the current frame of the video data from the background, generate multiple connected regions after performing morphological operations, and exclude fixed facilities in the automobile wind tunnel according to the area of ​​the connected regions to generate a binary image of the region;

[0054] The remaining area is stored in a suspicious object queue, and the suspicious objects in the suspicious object queue are continuously tracked in subsequent frames;

[0055] By calculating the distance between different object coordinates, the matching degree between grayscale images and the matching degree between Rain moments, the number of frames in which each suspicious object in the suspicious object queue exists and the number of frames in which it disappears are counted; the distance calculation formula between different object coordinates is:

[0056]

[0057] The calculation formula for the matching degree between the grayscale images is:

[0058]

[0059] The calculation formula for the matching degree between Rain moments is:

[0060]

[0061] Where N is the number of objects, C j is the horizontal or vertical coordinate of the jth object, C i is the horizontal or vertical coordinate of the i-th pixel of the j-th object, n is the number of pixels of the j-th object, p and q are the two objects being compared, and k represents the number of Rain moment components. represents the kth component of the Rain moment of object p, represents the kth component of the Rain moment of object q, m represents the range of the grayscale histogram, s represents the number of pixels in the grayscale histogram, Indicates the number of pixels with S grayscale in the grayscale histogram of object p, Represents the number of pixels with S grayscale in the grayscale histogram of object q.

[0062] If a new object is detected in a blank area, it will be added to the suspicious object queue and tracked continuously. If the object persists and its coordinates change, it will be judged as an area intrusion, and the type of the intruder will be determined based on the preset AI algorithm model.

[0063] If the number of frames where the suspicious object disappears continuously exceeds the preset frame number threshold, it will be removed from the suspicious object queue;

[0064] If the number of frames in which a suspicious object exists exceeds the preset frame number threshold, the suspicious object is judged to be a relic, and the preset AI algorithm model is called to identify the relic.

[0065] Among them, the preset AI algorithm model adopts the wind tunnel object feature recognition model, specifically:

[0066] Divide the dataset of feature maps of objects to be identified into a training set and a test set;

[0067] A wind tunnel object feature recognition model is constructed based on the YOLO model, and a residual network and a WCA module are added to the wind tunnel object feature recognition model; the WCA module performs a global average pooling operation on the input object feature map; then performs a one-dimensional convolution operation with a convolution kernel size of k, and obtains the weight of each channel through an activation function;

[0068] The EOUC loss function is introduced, and the wind tunnel object feature recognition model is trained through the training set; the EOUC loss function expression formula is:

[0069]

[0070] Among them, L dis is the center distance loss between the predicted box and the real box, L IOU is the overlap loss between the predicted box and the true box, L asp is the width and height loss of the predicted box and the real box. IOU is the accuracy standard for detecting corresponding objects in a specific data set. w, h, d are weights, and p is the activation function. and are the minimum bounding rectangle height and width of the real box and the predicted box, respectively, C 2 for and The sum of squares, d gt Indicates the coordinates of the center point of the target box, w gt Indicates the target box width information, h gt Indicates the target box height information;

[0071] After the training is completed, the test set is input into the wind tunnel object feature recognition model and the results are output.

[0072] In the present application, in addition to risk identification of suspicious objects in the automobile wind tunnel, the wind tunnel operating system 8 is also used to transmit the parameters of the environment adjustment module when it is in operation to the safety control module 7. The safety control module 7 is also used to receive the parameters and determine whether the parameters are within a preset range. If not, the safety control module 7 controls the wind tunnel operating system 8 to adjust the operation state of the environment adjustment module. For example, for the wind speed parameter of the fan 6, if the wind speed is ≤30kph, the risk factor is determined to be small. After being reminded, it is manually determined whether the test needs to be stopped and checked. If the wind speed is >30kph, the risk factor is determined to be high, and the test needs to be stopped for investigation.

[0073] As for the circuit signal monitored by the circuit monitoring unit 4, it is determined whether there is an abnormal situation of short circuit, open circuit, and abnormal situation of insulation resistance to ground. If a risk situation occurs, the test is stopped and an investigation is carried out.

[0074] The specific implementation process is as follows: at the beginning of the test, when the safety control module 7 starts the rainfall control system through the feedback of the wind tunnel operating system 8 and the wind tunnel light off signal transmitted by the monitoring camera unit 1, the safety control module 7 starts automatically and interacts with the wind tunnel operating system 8. The safety control module 7 controls the monitoring camera unit 1 to identify the safety risks in the wind tunnel flow area. If a safety risk item is identified, the safety control module 7 sends a command to the wind tunnel operating system 8 to prohibit the start of the fan 6. After the experiment personnel clear the safety risk items identified by the safety control module 7, the safety control module 7 monitors the electrical equipment, circuits, etc. through the current monitoring unit. After identifying the abnormal power consumption, the power is cut off, the lighting in the wind tunnel is turned on, and the wind tunnel operating system 8 is sent to the wind tunnel operating system 8. The command to prohibit the start of the fan 6 and the rainfall water supply system 2 is sent to the wind tunnel operating system 8, and the experiment personnel will restart the test after checking the risks.

[0075] After the fan 6 is started, the wind tunnel operating system 8 transmits the wind speed signal to the safety control module 7. The safety control module 7 collects the wind speed signal ≤30kph. The safety control module 7 determines that the risk factor is small. If the safety control module 7 receives an abnormality in the transmission of the monitoring signal, it is also necessary to stop the test to check the risk. If no abnormality is received, the test will continue until the end of the test.

[0076] After the fan 6 is started, the wind tunnel operating system 8 transmits a wind speed signal to the safety control module 7. When the safety control module 7 collects a wind speed signal greater than 30 kph, the safety control module 7 determines that the risk is high. When the monitoring module detects an intrusion or suspicious object, the test is stopped for risk investigation. After the investigation is completed, the test is continued until the test is completed.

[0077] In another embodiment of the present application, a vehicle environmental wind tunnel rainfall test safety control method is also included, which is applied to the above-mentioned vehicle environmental wind tunnel rainfall test safety control system.

[0078] The above are only embodiments of the present invention. Common knowledge such as the known specific structures and characteristics in the scheme is not described in detail here. Ordinary technicians in the relevant field know all the common technical knowledge in the technical field to which the invention belongs before the application date or priority date, can obtain all the existing technologies in the field, and have the ability to apply conventional experimental means before that date. Ordinary technicians in the relevant field can improve and implement this scheme in combination with their own abilities under the enlightenment given by this application. Some typical known structures or known methods should not become obstacles for ordinary technicians in the relevant field to implement this application. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several deformations and improvements can be made, which should also be regarded as the scope of protection of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A safety control system for an automobile environmental wind tunnel rainfall test, the automobile environmental wind tunnel comprising an environmental adjustment module and a wind tunnel operating system, the environmental adjustment module communicates with the wind tunnel operating system, and the wind tunnel operating system controls the operating state of the environmental adjustment module; characterized in that: The safety control system includes a monitoring module, an alarm module and a safety control module; the wind tunnel operating system is communicatively connected with the safety control module; The monitoring module and the alarm module are in communication connection with the safety control module. The safety control module receives the monitoring signal of the monitoring module, and analyzes the monitoring signal according to the preset safety risk monitoring model to generate a risk result. The safety control module is also used to control the wind tunnel operating system according to the risk result to adjust the operating state of the environment adjustment module; the safety control module is also used to control the operating state of the alarm module according to the risk result; The security risk monitoring model preset in the security control module is specifically: Obtain historical data from automobile wind tunnel rainfall tests, perform model training after preprocessing, and generate a safety risk monitoring model suitable for automobile wind tunnel environments; The security risk monitoring model obtains the video data collected by the monitoring camera unit and performs risk analysis on the video data; The security risk monitoring model obtains the video data collected by the monitoring camera unit and performs risk analysis on the video data as follows: Read the video data, set the frame number threshold according to the stability characteristics of the video data in the flow channel area of ​​the automobile wind tunnel before the rainfall test, and establish the test background image; Obtain the difference image after subtracting the current frame of the video data from the background, generate multiple connected regions after performing morphological operations, and exclude fixed facilities in the automobile wind tunnel according to the area of ​​the connected regions to generate a binary image of the region; The remaining area is stored in a suspicious object queue, and the suspicious objects in the suspicious object queue are continuously tracked in subsequent frames; By calculating the distance between different object coordinates, the matching degree between grayscale images and the matching degree between Rain moments, the number of frames in which each suspicious object in the suspicious object queue exists and the number of frames in which it disappears are counted; If a new object is detected in a blank area, it will be added to the suspicious object queue and tracked continuously. If the object persists and its coordinates change, it will be judged as an area intrusion, and the type of the intruder will be determined based on the preset AI algorithm model. If the number of frames where the suspicious object disappears continuously exceeds the preset frame number threshold, it will be removed from the suspicious object queue; If the number of frames in which a suspicious object exists exceeds the preset frame number threshold, the suspicious object is judged to be a relic, and the preset AI algorithm model is called to identify the relic.

2. The automobile environmental wind tunnel rainfall test safety control system according to claim 1 is characterized by: The monitoring module includes a monitoring camera unit and a circuit monitoring unit. The monitoring camera unit covers the wind tunnel flow channel area, and the circuit monitoring unit monitors the real-time current and voltage of the circuit.

3. The automobile environmental wind tunnel rainfall test safety control system according to claim 2 is characterized by: The wind tunnel operating system is also used to transmit the parameters of the environment adjustment module when it is in operation to the safety control module. The safety control module is also used to receive the parameters and determine whether the parameters are within a preset range. If not, the safety control module controls the wind tunnel operating system to adjust the operation state of the environment adjustment module.

4. The automobile environmental wind tunnel rainfall test safety control system according to claim 1 is characterized by: The distance calculation formula between the coordinates of different objects is: The calculation formula for the matching degree between the grayscale images is: The calculation formula for the matching degree between Rain moments is: Where N is the number of objects, C j is the horizontal or vertical coordinate of the jth object, C i is the horizontal or vertical coordinate of the i-th pixel of the j-th object, n is the number of pixels of the j-th object, p and q are the two objects being compared; k represents the number of Rain moment components, represents the kth component of the Rain moment of object p, represents the kth component of the Rain moment of object q, m represents the range of the grayscale histogram, s represents the number of pixels in the grayscale histogram, Indicates the number of pixels with S grayscale in the grayscale histogram of object p, Represents the number of pixels with S grayscale in the grayscale histogram of object q.

5. The automobile environmental wind tunnel rainfall test safety control system according to claim 4 is characterized by: The preset AI algorithm model adopts a wind tunnel object feature recognition model, specifically: Divide the dataset of feature maps of objects to be identified into a training set and a test set; A wind tunnel object feature recognition model is constructed based on the YOLO model, and a residual network and a WCA module are added to the wind tunnel object feature recognition model; the WCA module performs a global average pooling operation on the input object feature map; then performs a one-dimensional convolution operation with a convolution kernel size of k, and obtains the weight of each channel through an activation function; The EOUC loss function is introduced, and the wind tunnel object feature recognition model is trained using the training set; After the training is completed, the test set is input into the wind tunnel object feature recognition model and the results are output.

6. The automobile environmental wind tunnel rainfall test safety control system according to claim 5, characterized in that: The EOUC loss function expression formula is: Among them, L dis is the center distance loss between the predicted box and the real box, L IOU is the overlap loss between the predicted box and the true box, L asp is the width and height loss of the predicted box and the real box. IOU is the accuracy standard for detecting corresponding objects in a specific data set. w, h, d are weights, and p is the activation function. and are the minimum bounding rectangle height and width of the real box and the predicted box, respectively, C 2 for and The sum of squares, d gt Indicates the coordinates of the center point of the target box, w gt Indicates the target box width information, h gt Indicates the target box height information.

7. A safety control method for automobile environmental wind tunnel rainfall test, characterized in that: An automobile environmental wind tunnel rainfall test safety control system applied to any one of claims 1-6 above.

Images