Visual sharing system of vehicle-mounted network and working method of visual sharing system

Abstract

The invention discloses a visual sharing system of a vehicle-mounted network and a working method of the visual sharing system, and belongs to the field of vehicle-mounted networks and visual sharing. The system recognizes a license plate by capturing an image, and searches a network camera number corresponding to the license plate in a database. For the searched vehicle, extracting a single frame from the image acquired in a certain time period; coding compression is carried out on the frame of image, and wireless transmission of data is achieved through the WiFi technology. After a receiving end receives the transmitted data, the transmitted data are restored into continuous frames through decoding, so that a real-time video picture is formed; and the obtained image is cut and segmented, so that the image is only displayed at a specified position. According to the system, real-time and high-quality video transmission is realized, and image preprocessing, license plate positioning, character segmentation and front vehicle distance and vehicle speed measurement are completed. According to the system, more comprehensive and accurate road condition information is provided for a driver, safer and more intelligent driving experience is assisted to be achieved, the road driving safety is improved, and traffic accidents are reduced.

Description

Technical Field

[0001] The present invention relates to the field of vehicle-mounted networks and vision sharing, and in particular to a vehicle-mounted network vision sharing system based on machine vision, a vehicle-mounted network and image processing. Background Art

[0002] In modern traffic, the limited vision of drivers is one of the important causes of traffic accidents. When there are large vehicles or other obstructions in front, it is difficult for drivers to observe the road conditions ahead, which increases the risk of driving.

[0003] In response to the above problems, traditional driver assistance systems, such as radar and lidar, can provide a certain degree of obstacle detection and warning in front of the car, but cannot provide detailed visual information of the road ahead. In addition, although camera-based driver assistance systems can capture and process images, they are usually limited to the camera range of the vehicle itself and cannot solve the problem of the vehicle ahead blocking the view. The effect is very limited, and it can only solve the problem of limited view caused by blind spots caused by the vehicle's own design, and cannot solve the problem of limited view caused by the vehicle ahead blocking the view during driving.

[0004] To sum up, limited field of vision has always been a problem that affects driving safety, and traditional driving assistance systems provide very limited help. It is necessary to invent a new system to better address the problems caused by limited field of vision. Summary of the invention

[0005] The present invention aims to solve the problem of visual blind spots caused by obstruction of the vehicle in front during traditional driving, and provides an in-vehicle network visual sharing system based on WiFi image transmission. The system can not only transmit the complete field of view of the vehicle in front in real time, but also combines license plate recognition, YOLO vehicle detection, and transparent gradient processing technology to further enhance the driver's ability to perceive road conditions. The system can not only provide real-time images in front of the driver's vehicle, but also flexibly transmit data within the range of wireless communication, eliminating the limitations of wired connections. This enables the driver to observe the road conditions blocked by the body of the vehicle in front in real time, increasing driving safety and road driving reliability.

[0006] To achieve the above-mentioned purpose, the scheme adopted by the present invention is: a visual sharing system for an in-vehicle network, comprising: an image capture module, used to capture images from a hardware device; a license plate recognition module, used to recognize the license plate in the image, and search the corresponding network camera number in the database to realize directional network communication; a frame extraction module, used to extract a single frame image within a certain time period from the captured image; a coding and compression module, based on the OpenCV library and using the extracted single frame image to encode and compress it to reduce the amount of data and maintain the image quality; a data transmission module, using WiFi hotspots to realize wireless transmission of encoded image data; a decoding module, used to decode the received encoded image data at the receiving end, and form the decoded image data into a continuous frame video playback; a cropping and segmentation module, used to crop and segment the decoded image so that it is only displayed at a specified position; a YOLO vehicle recognition module, used to identify the vehicle in front and detect its distance and speed; a transparent gradient processing module, used to calculate the Euclidean distance of each pixel point from the central point of the image, and apply a transparent gradient effect according to the distance change.

[0007] As a further improved technical solution of the present invention, the image capture module includes a camera for capturing video data;

[0008] As a further improved technical solution of the present invention, the image capture module includes a camera for capturing video data, and the camera supports real-time video capture under high-speed motion and can adapt to different lighting conditions and complex road environments;

[0009] As a further improved technical solution of the present invention, the license plate recognition module is based on image processing technology, including license plate positioning and character segmentation, and can automatically match the license plate with the network camera number to form a real-time video link with the preceding vehicle.

[0010] As a further improved technical solution of the present invention, the operation based on the database includes the establishment and maintenance of the database, and the database has the function of automatic mapping of license plates and camera numbers, and supports real-time dynamic updating of license plate information.

[0011] As a further improved technical solution of the present invention, the time interval for extracting frames changes in real time according to network conditions and computing load conditions, and when the network bandwidth is insufficient, the frame rate can be automatically adjusted to maintain the continuity of the video.

[0012] As a further improved technical solution of the present invention, the encoding compression module further includes settings for selecting different compression parameters to optimize image quality and data volume, and can dynamically adjust the encoding parameters according to the scene to ensure complete presentation of image details.

[0013] As a further improved technical solution of the present invention, the data transmission module also includes a data encryption and identity authentication mechanism to ensure the security of data transmission.

[0014] As a further improved technical solution of the present invention, the cropping and segmentation module includes how to locate the image to display only the specified position, and can adaptively segment the image of the vehicle in front according to the driver's perspective and accurately locate the blocked area of ​​the vehicle in front.

[0015] As a further improved technical solution of the present invention, the transparent gradient processing module calculates the Euclidean distance between each pixel point of the image and the center point of the image, and applies a gradient effect of increasing transparency as the distance increases, and combines vehicle detection technology to ensure that the transparent area accurately corresponds to the actual contour of the preceding vehicle.

[0016] As a further improved technical solution of the present invention, the main purpose of the system is to help the driver observe the road information blocked by the vehicle ahead.

[0017] As a further improved technical solution of the present invention, the display module further includes a display device for displaying real-time video images, and overlays the processed image on the parking space of the vehicle in front, so that the driver can see the obstructed road, and the image presentation has the characteristics of high resolution and low latency, adapting to high-speed driving environments.

[0018] As a further improved technical solution of the present invention, the system can create a "perspective" effect by overlapping the transparent gradient processing with the obstruction area of ​​the preceding vehicle, thereby significantly improving the driver's ability to perceive road conditions and avoiding traffic accidents.

[0019] A working method of a visual sharing system of a vehicle network comprises the following steps:

[0020] S1. Image Capture:

[0021] After the system is started, the driving camera of the vehicle in front starts to capture video images in real time; the image capture module obtains real-time video data from the camera; the video file is opened and a video encoder is created to save the output video; then, each frame of the video is read through a frame extraction block loop and written to the output video file after processing the frame; in the loop, the required processing is performed on each frame, and the encoder, frame rate, and resolution of the output video are specified;

[0022] S2. Image compression:

[0023] The image is compressed by the encoding and compression module. During the image compression process, the image is cut into small blocks of size nXn. Each small block is processed separately during the compression process. Then the color space is converted from RGB to YCbCr, where Y represents brightness, Cb and Cr represent the color difference values ​​of green and red respectively. The formula for converting RGB to YCbCr is as follows:

[0024]

[0025] During the compression process, the image is converted into color space, each n×n image block is represented as three n×n matrices, and then the matrices are subjected to two-dimensional discrete cosine transform respectively;

[0026]

[0027] Among them, G(u, v) represents the elements of the transformed coefficient matrix, u and v represent the coordinates in the frequency domain, x and y represent the coordinates in the spatial domain; N is the size of the image;

[0028] After color space conversion and discrete cosine transform, the image block is converted into a floating point matrix; the floating point is then quantized and converted into an integer value; JPEG provides a quantization algorithm to quantize the image. The quantization algorithm is as follows:

[0029]

[0030] in, Represents the elements in the quantized frequency domain coefficient matrix, Represents the elements in the frequency domain coefficient matrix after DCT transformation, Represents the elements in the quantization matrix, and the round function represents the rounding operation;

[0031] S3. Image transmission:

[0032] The data transmission module performs image transmission, sends the image frame to the receiving end, specifies the identifier of the sending end, specifies the connection symbol and the image frame data to be connected; the decoding module is used to cyclically receive the image frame at the receiving end, and obtain the identifier and the image frame data of the sending end;

[0033] S4. License plate recognition

[0034] Load the vehicle detection model, then use the vehicle detection model to detect the input image to obtain the vehicle's position and bounding box; the cropping and segmentation model is used to crop and segment the decoded image so that it is only displayed at the specified position; for each vehicle, extract its license plate area; adjust the license plate area to a standard rectangular shape by using perspective transformation; recognize the adjusted license plate image to obtain the license plate number; finally, draw the recognition result on the original image, including drawing the license plate number, vehicle bounding box and license plate area; after the whole process is completed, return the image of the drawing result;

[0035] In the transparent gradient processing model, the code loops through each pixel of the image to obtain the color value of the current pixel; the transparency value is calculated by calling according to the distance between the position of the pixel in the image and the center of the image; the calculated transparency value is achieved by modifying the alpha channel in the color tuple of the pixel, and then the modified color is applied to the pixel; the Euclidean distance formula is used to calculate the distance between the pixel point and the center of the image. The Euclidean distance formula is:

[0036]

[0037] Among them, d represents the distance result between two points, and the x and y coordinate values ​​of point A and point B are m;

[0038] Achieve an axis gradient effect that gradually changes color or transparency from the center to the outer edge of the image;

[0039] S5. Display positioning

[0040] The license plate number is located through image processing to determine the parking space of the vehicle in front, and then the image after axis gradient is placed above the center of the license plate. The specific position is (x-size_w / 2+w / 2,y-size_h / 2), where x indicates the x-axis pixel position of the center point of the license plate, y indicates the y-axis pixel position of the center point of the license plate, size_w indicates the widest width of the image after axis gradient, and size_h indicates the highest height of the image after axis gradient.

[0041] The beneficial effects of the present invention are:

[0042] (1) The present invention uses a camera to collect images. In the collected image information, not only can the position information of the target vehicle relative to the signal source vehicle be effectively extracted, but also the license plate number can be recognized through the collected depth image, so as to query the corresponding network communication address according to the recognized license plate number. In addition, the speed of the vehicle in front and the distance to the vehicle in front can be obtained by calculation. Through the unique transparent gradient processing technology, the driver can "see through" the vehicle in front and clearly observe the road information blocked by it, which significantly improves driving safety and reaction speed.

[0043] (2) The present invention utilizes the OpenCV library for image encoding and the H.264 standard for compression, and realizes efficient transmission of WiFi hotspots through the PyWifi library, thereby ensuring real-time transmission and high-quality display of encoded image data. This combination of technologies ensures data integrity and stability during image transmission, provides drivers with clear, smooth real-time video images, and enhances driving experience and information acquisition capabilities. In addition, the present invention dynamically optimizes video quality and transmission efficiency according to network bandwidth and computing power through an adaptive compression parameter adjustment mechanism, ensuring that stable video transmission can be provided even under complex road conditions and signal fluctuations.

[0044] (3) The present invention combines the YOLO algorithm to perform real-time identification and distance speed detection of the vehicle ahead, providing accurate vehicle dynamic information. This not only helps the driver to obtain the motion status of the vehicle ahead in a timely manner, but also effectively prevents traffic accidents caused by changes in vehicle speed, further improving driving safety and driving comfort. Through transparent gradient processing technology, the road information of the vehicle-blocked area is accurately projected into the driver's field of view, so that the system can better help the driver perceive obstacles and potential dangers ahead.

[0045] (4) The present invention realizes intelligent monitoring and real-time sharing of road information ahead through the vehicle network vision sharing system. The driver can obtain real-time images captured by other vehicles through the system, as well as relevant information obtained through database matching, which effectively improves the perception and response capabilities of road condition changes, helps optimize traffic flow control and improve the overall management efficiency of urban roads. In particular, the system can automatically match and connect to the camera of the vehicle ahead according to the vehicle license plate number, and transmit the road information in the front field of view in real time, effectively solving the problem of limited driver field of view.

[0046] (5) As an intelligent driving assistance system, the present invention not only meets the driver's needs for safety and information acquisition, but also promotes the development of intelligent driving technology. By combining advanced image processing, wireless transmission and vehicle identification technology, the present invention demonstrates a feasible intelligent vehicle traffic management solution, laying the foundation for the construction of future intelligent transportation systems. The unique transparent gradient processing of the present invention combined with vehicle positioning provides an innovative "see-through driving" experience, which not only improves the driver's perception of road conditions, but also lays a new foundation for visual sharing and collaborative driving technology of intelligent transportation systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a schematic diagram of the structure of the present invention.

[0048] Figure 2 It is a schematic diagram of the workflow of the present invention.

[0049] Figure 3 It is an interactive schematic diagram of the present invention.

[0050] Figure 4 This is a comparison chart before and after the axis gradient effect is processed.

[0051] Figure 5 This is a comparison chart before and after positioning processing.

[0052] Figure 6 Scenes showing the driving perspective. DETAILED DESCRIPTION

[0053] The specific embodiments of the present invention are further described below according to the accompanying drawings:

[0054] like Figure 1 As shown, a visual sharing system for an in-vehicle network includes: an image capture module, which is used to capture images from a hardware device; a license plate recognition module, which is used to recognize the license plate in the image and search the corresponding network camera number in the database to realize directional network communication; a frame extraction module, which is used to extract a single frame image within a certain time period from the captured image; a coding and compression module, which is based on the OpenCV library and uses the extracted single frame image to encode and compress it to reduce the data volume and maintain the image quality; a data transmission module, which uses WiFi hotspots to realize wireless transmission of coded image data; a decoding module, which is used to decode the received coded image data at the receiving end and form the decoded image data into a continuous frame video playback; a cropping and segmentation module, which is used to crop and segment the decoded image so that it is only displayed at a specified position; a YOLO vehicle recognition module, which is used to identify the vehicle in front and detect its distance and speed; a transparent gradient processing module, which is used to calculate the Euclidean distance of each pixel point from the central point of the image and apply a transparent gradient effect according to the distance change.

[0055] like Figure 1 As shown, the front camera of vehicle B in this embodiment continuously identifies the road ahead until vehicle A appears in image B within the field of view of the camera of vehicle B. After the camera of vehicle B identifies vehicle A, the information in the image is processed by DSP, and the speed and distance information of vehicle A are summarized as the front vehicle information A, which is fed back to the driver through the screen.

[0056] like Figure 1 As shown, in this embodiment, after the front camera of car B recognizes the license plate number of car A, the camera number A of car A is obtained by comparison and search through the DSP module. After establishing network communication with car A through the network module, the image transmission image A of car A is obtained, and the image transmission image A is fed back to the DSP module. After the DSP module processes the image, a real-time warning image is obtained, and the real-time warning image is fed back to the driver through the screen.

[0057] like Figure 1 As shown, in this embodiment, car A obtains image A from its front camera and then passes it to DSP for processing. The processed image is passed to the network module. After the network module establishes a connection with the network module of the rear car B, the processed image A is transmitted via WIFI.

[0058] After obtaining image B, the DSP module of this embodiment identifies the license plate number of vehicle A appearing in image B, compares the character string of license plate number A in the database, queries the vehicle camera number A corresponding to vehicle A, calculates the distance and speed of vehicle A in front through image B, and summarizes the distance and speed information as the front vehicle information. At the same time, image B is compressed and encoded to obtain processed image B, and the processed image B is sent to the rear vehicle via the network module.

[0059] After receiving the processed image A from the DSP, the network module of this embodiment will broadcast its own signal until it successfully finds a network information source that matches its own camera number, and after connecting with it, it sends the processed image A.

[0060] After receiving the image A sent by the preceding vehicle through the network module, the DSP module of this embodiment performs graphic transformation to cover part of the image content around the license plate number in the image B with the image A to form a real-time warning image.

[0061] After receiving the camera number F sent by the DSP module connected to it, the network module of this embodiment immediately searches for the network source corresponding to the number F in the surrounding area, connects to the network source corresponding to the number F, and obtains the image transmission image F through other vehicle network modules.

[0062] like Figure 2 As shown, the license plate recognition part receives the license plate image taken by the camera of the front vehicle and performs license plate recognition processing. The recognized license plate number is passed to the front vehicle information query module for query. The front vehicle information query module queries the camera number of the front vehicle through the database module. The camera calling part calls the corresponding camera according to the obtained camera number, obtains the image of the front vehicle, and then decodes the image, cuts and dilutes it. Finally, the obtained image is merged with the image taken by the front vehicle camera, and the front vehicle distance and speed information calculated by the image taken by the front vehicle camera are fed back to the driver through the screen. The self-image data is transmitted from the front vehicle camera module to the image transmission module, and after compression encoding, it is transmitted to the rear vehicle receiving module through the hotspot for decoding and real-time video playback.

[0063] like Figure 3As shown, the system allows the driver of the following vehicle to connect to the hotspot and display the image transmitted by real-time video. The driver can view the image information transmitted in real time. The system can compare and query from the database and add new vehicle information to the database. The system makes relevant driving judgments based on the existing information and supports function expansion, hazard warning, traffic sign detection, pedestrian detection, etc.

[0064] The operation process of this embodiment is as follows:

[0065] (1) Classify and list several situations where the whistle needs to be blown:

[0066] Based on common sense, we have summarized several typical scenarios where honking is required on highways or in urban-rural fringe areas. On highways, vehicles travel at a high speed, and the vehicle in front blocks the driver's field of vision, which will cause the driver to be unable to grasp the road conditions ahead in time, increasing the risk of rear-end collisions; on urban roads, the traffic volume is heavy, and large vehicles in front (such as buses and trucks) often block the driver's field of vision, affecting the driver's observation of important information such as traffic lights and road signs; when driving at night, the field of vision is limited, and the taillights of the vehicle in front may affect the driver's vision, especially on bends and complex sections; in severe weather conditions such as rain, snow and fog, visibility is reduced, and it is difficult for the driver to judge the road conditions ahead, increasing the difficulty and danger of driving; after a traffic accident occurs, the road ahead may be blocked, and the driver needs to obtain road information ahead in time in order to choose a suitable detour route.

[0067] (2) Image capture:

[0068] After the system is activated, the driving camera of the vehicle in front begins to capture video images in real time.

[0069] The image capture section acquires live video data from the camera. This section opens a video file and creates a video encoder to save the output video. Then, a loop reads each frame of the video and writes it to the output video file after processing the frame. In the loop, the required processing is performed on each frame, and the encoder, frame rate, and resolution of the output video are specified.

[0070] (3) Image compression:

[0071] During the image compression process, the image is cut into small blocks of 8X8, and each small block is processed separately during the compression process. Then the color space is converted from RGB->YCbCr, where Y represents luminance, Cb and Cr represent the color difference values ​​of green and red respectively. The formula for converting RGB to YCbCr is as follows:

[0072]

[0073] During the compression process, the image is converted in color space, each 8×8 image block is represented as three 8x8 matrices, and then the three matrices are subjected to two-dimensional discrete cosine transform respectively.

[0074]

[0075] Among them, G(u, v) represents the elements of the coefficient matrix after transformation, u and v represent the coordinates in the frequency domain, x and y represent the coordinates in the spatial domain, and N is the size of the image.

[0076] After color space conversion and discrete cosine transform, the image block is converted into a floating point matrix. In order to save space, these floating point numbers are quantized and converted into integer values. JPEG provides a quantization algorithm that uses a specific quantization table to quantize the image to reduce the use of storage space. The quantization algorithm is as follows:

[0077]

[0078] in, Represents the elements in the quantized frequency domain coefficient matrix, Represents the elements in the frequency domain coefficient matrix after DCT transformation, Represents the elements in the quantization matrix, and the round function represents the rounding operation. After quantization, a large part of the data becomes 0, which is very beneficial for subsequent compression and transmission.

[0079] (3) Image transmission:

[0080] The image frame is sent to the receiving end. The identifier of the sending end, the connector to be connected and the image frame data can be specified. The receiving end receives the image frame cyclically and obtains the identifier and image frame data of the sending end.

[0081] (4) License plate recognition

[0082] First, the vehicle detection model is loaded, and then the input image is detected using the vehicle detection model to obtain the location and bounding box of the vehicle. For each vehicle, its license plate area is extracted. The license plate area is adjusted to a standard rectangular shape by using perspective transformation. The adjusted license plate image is recognized to obtain the license plate number. Finally, the recognition result is drawn on the original image, including drawing the license plate number, vehicle bounding box, and license plate area. After the whole process is completed, the image of the drawing result is returned.

[0083] The axis gradient effect of the image gradually changes the color or transparency from the center of the image to the outer edge, making the image look closer to the perspective effect. The code loops through each pixel of the image and obtains the color value of the current pixel. According to the distance between the position of the pixel in the image and the center of the image, the transparency value is calculated by calling. The calculated transparency value is achieved by modifying the alpha channel in the color tuple of the pixel, and then the modified color is applied to the pixel. Figure 4 This is a comparison chart before and after the axis gradient effect. The Euclidean distance formula is used to calculate the distance between the pixel and the center of the image. The Euclidean distance formula is:

[0084]

[0085] Among them, d represents the distance result between two points, and x and y are the n coordinate values ​​of point A and point B.

[0086] (6) Display positioning

[0087] The license plate number is located through image processing to determine the parking space of the vehicle in front, and then the image after axis gradient is placed above the center of the license plate. The specific position is (x-size_w / 2+w / 2,y-size_h / 2), where x indicates the x-axis pixel position of the center point of the license plate, y indicates the y-axis pixel position of the center point of the license plate, size_w indicates the widest width of the image after axis gradient, and size_h indicates the highest height of the image after axis gradient. Figure 5 This is a comparison chart before and after positioning processing.

[0088] (7) Final effect

[0089] Figure 6 The scene showing the driving perspective obtains the road information ahead through the on-board camera and uses transparent gradient processing technology to "project" the road image onto the rear windshield of the vehicle ahead, allowing the driver to see the road ahead through the vehicle ahead. This visual effect not only improves the driver's safety and reaction speed, but also reduces traffic accidents caused by obstructed vision, greatly optimizing the driving experience.

[0090] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit the present invention. Although the above embodiments describe the present invention in detail, it should be understood that a person skilled in the art can still modify the technical solutions or replace some of the technical features with equivalents. These modifications or replacements do not change the essence of the corresponding technical solutions, and still fall within the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A visual sharing system for an in-vehicle network, characterized in that: The system includes: An image capture module, used to capture images from a hardware device, the module including a camera for capturing video data; The license plate recognition module is used to identify the license plate in the image and find the corresponding network camera number in the database; the license plate recognition module is based on image processing technology, including license plate positioning and character segmentation; A frame extraction module is used to extract a single frame image within a certain period of time from the captured image; The encoding and compression module is based on the OpenCV library and uses the extracted single frame image to encode and compress it to reduce the amount of data and maintain image quality; Data transmission module, using WiFi hotspot to achieve wireless transmission of encoded image data; A decoding module is used to decode the received coded image data at the receiving end, and form the decoded image data into a continuous frame video for playback; The cropping and segmentation module is used to crop and segment the decoded image so that it is displayed only at the specified position; YOLO vehicle recognition module, used to identify the vehicle ahead and detect its distance and speed; A transparent gradient processing module is used to calculate the Euclidean distance between each pixel and the center point of the image, and apply a transparent gradient effect according to the distance change; The display module is a display device for displaying real-time video images and overlaying the processed image at the parking space of the vehicle in front so that the driver can see the blocked road.

2. The visual sharing system of the vehicle network according to claim 1, characterized in that: The data transmission module also includes data encryption and identity authentication mechanisms.

3. The working method of the visual sharing system of the vehicle network according to claim 1 is characterized in that: The following steps are involved: S1. Image Capture: After the system is started, the driving camera of the vehicle in front starts to capture video images in real time; the image capture module obtains real-time video data from the camera; the video file is opened and a video encoder is created to save the output video; then, each frame of the video is read through a frame extraction block loop and written to the output video file after processing the frame; in the loop, the required processing is performed on each frame, and the encoder, frame rate, and resolution of the output video are specified; S2. Image compression: The image is compressed by the encoding and compression module. During the image compression process, the image is cut into small blocks of size nXn. Each small block is processed separately during the compression process. Then the color space is converted from RGB to YCbCr, where Y represents brightness, Cb and Cr represent the color difference values ​​of green and red respectively. The formula for converting RGB to YCbCr is as follows: ； During the compression process, the image is converted into color space, each n×n image block is represented as three nxn matrices, and then the matrices are subjected to two-dimensional discrete cosine transform respectively; ； Among them, G(u, v) represents the elements of the transformed coefficient matrix, u and v represent the coordinates in the frequency domain, x and y represent the coordinates in the spatial domain; N is the size of the image; After color space conversion and discrete cosine transform, the image block is converted into a floating point matrix; the floating point is then quantized and converted into an integer value; JPEG provides a quantization algorithm to quantize the image. The quantization algorithm is as follows: ； in, Represents the elements in the quantized frequency domain coefficient matrix, Represents the elements in the frequency domain coefficient matrix after DCT transformation, Represents the elements in the quantization matrix, and the round function represents the rounding operation; S3. Image transmission: The data transmission module performs image transmission, sends the image frame to the receiving end, specifies the identifier of the sending end, specifies the connection symbol and the image frame data to be connected; the decoding module is used to cyclically receive the image frame at the receiving end, and obtain the identifier and the image frame data of the sending end; S4. License plate recognition Load the vehicle detection model, then use the vehicle detection model to detect the input image to obtain the vehicle's position and bounding box; the cropping and segmentation model is used to crop and segment the decoded image so that it is only displayed at the specified position; for each vehicle, extract its license plate area; adjust the license plate area to a standard rectangular shape by using perspective transformation; recognize the adjusted license plate image to obtain the license plate number; finally, draw the recognition result on the original image, including drawing the license plate number, vehicle bounding box and license plate area; after the whole process is completed, return the image of the drawing result; In the transparent gradient processing model, the code loops through each pixel of the image to obtain the color value of the current pixel; the transparency value is calculated by calling according to the distance between the position of the pixel in the image and the center of the image; the calculated transparency value is achieved by modifying the alpha channel in the color tuple of the pixel, and then the modified color is applied to the pixel; the Euclidean distance formula is used to calculate the distance between the pixel point and the center of the image. The Euclidean distance formula is: ； Among them, d represents the distance result between two points, and the x and y coordinate values ​​of point A and point B are m; Achieve an axis gradient effect that gradually changes color or transparency from the center to the outer edge of the image; S5. Display positioning The license plate number is located through image processing to determine the parking space of the vehicle in front, and then the image after axis gradient is placed above the center of the license plate through the display module. The specific position is (x-size_w / 2+w / 2,y-size_h / 2), where x indicates the x-axis pixel position of the center point of the license plate, y indicates the y-axis pixel position of the center point of the license plate, size_w indicates the widest width of the image after axis gradient, and size_h indicates the highest height of the image after axis gradient.

Images