Unmanned vehicle panoramic vision sensor intelligent calibration method and electronic equipment

Abstract

The embodiment of the invention provides an intelligent calibration method for a panoramic vision sensor of an unmanned vehicle and electronic equipment, and relates to the technical field of computer vision. The intelligent calibration method for the panoramic vision sensor of the unmanned vehicle comprises the following steps: S100, setting an initial state of the unmanned vehicle and initial states of fisheye cameras in four directions; s200, determining a calibration board with feature points according to the size of the unmanned vehicle and the shooting range; s300, using an angular point detection algorithm; s400, pre-identifying angular point coordinate information in combination with an angular point detection algorithm; s500, based on the detected angular points of the calibration plate and the known positions of the angular points in the world coordinate system; and in combination with an image processing algorithm, the influence of factors such as vibration and illumination change in the operation process of the unmanned vehicle on the position and direction of the camera can be accurately sensed. Through the application of the adaptive algorithm, the position and direction parameters of the camera can be corrected in real time, thereby ensuring the accuracy and consistency of the obtained image data.

Description

Technical Field

[0001] This application relates to the field of computer vision technology, and in particular, to an intelligent calibration method for a panoramic vision sensor of an autonomous vehicle and an electronic device. Background Art

[0002] In the related art, an autonomous vehicle is a self-driving vehicle that can independently complete driving tasks through functions such as perception, decision-making, and control. To achieve accurate and reliable environmental perception, an autonomous vehicle is usually equipped with a surround-view camera system, which consists of multiple fish-eye cameras and can provide an omnidirectional view to assist the autonomous vehicle in detecting the surrounding environment and making accurate decisions. However, factors such as the installation position, angle, and distortion of the surround-view cameras will affect the quality and accuracy of the images. Therefore, it is necessary to calibrate the surround-view cameras to eliminate these effects; However, the traditional camera calibration method refers to the method of using known calibration objects to obtain the internal and external parameters of the camera. Generally, it is necessary to take pictures of the calibration object at multiple different poses, which increases the time and difficulty of calibration. Moreover, manually detecting the feature points on the calibration object may have errors or failures; in some complex environments, the traditional camera calibration method cannot meet the complex environmental requirements. Therefore, there is an urgent need for an intelligent calibration method to improve the accuracy and robustness of the surround-view camera system, reduce manual intervention and error accumulation, and reduce maintenance costs while improving efficiency. Summary of the Invention

[0003] This application aims to at least solve one of the technical problems existing in the prior art. For this purpose, this application proposes an intelligent calibration method for a panoramic vision sensor of an autonomous vehicle and an electronic device.

[0004] In a first aspect, an intelligent calibration method for a panoramic vision sensor of an autonomous vehicle provided by an embodiment of the present invention includes: S100. Set the initial state of the autonomous vehicle and the initial states of the fish-eye cameras in four directions; S200. Determine a calibration board with feature points according to the size of the autonomous vehicle and the shooting range, and fix it in four directions of the autonomous vehicle so that the fish-eye cameras in four directions can all capture at least one calibration board; S300. Use a corner detection algorithm to process the calibration board in the camera video frame and pre-extract the corner coordinate information; S400. Combine the corner coordinate information pre-identified by the corner detection algorithm, adopt a feature descriptor algorithm to automatically match the corners in different images, and establish the correspondence between the calibration board and the image points; S500. Based on the detected corner points of the calibration board and their known positions in the world coordinate system, use a calibration model to calculate the internal parameters of the camera; S600. Evaluate and verify the calibration results, adjust the parameters of the corner detection model to improve the calibration results, obtain the bird's-eye view and projection matrix of each fisheye camera, and complete the panoramic calibration work.

[0005] In addition, the intelligent calibration method for the panoramic vision sensor of the unmanned vehicle according to the embodiment of the present application further has the following additional technical features: In a preferred embodiment of the present invention, in S100, setting the initial state of the unmanned vehicle and the initial states of the fisheye cameras at four orientations includes: Camera mounting points are set in advance in the front, rear, left, and right of the unmanned vehicle; Change the angle and position by rotating and moving the fisheye camera to ensure good image quality; ensure that the images captured by the fisheye camera can cover the entire field of view.

[0006] The entire field of view includes the edges and corners; adjust to avoid problems such as blurred images, uneven illumination, or overexposure.

[0007] In a preferred embodiment of the present invention, in S200, determining the calibration board with feature points according to the size of the unmanned vehicle and the shooting range includes: The calibration board with feature points is in the shape of a checkerboard, and its inner corner points have strong contrast and high sub-pixel accuracy; The checkerboard is composed of alternating horizontal and vertical lines to form square grids, and the side length of each square grid is 1 cm - 17 cm, where multiple horizontal lines are arranged in parallel and multiple vertical lines are arranged in parallel; The checkerboard is placed under lighting conditions; adopt appropriate lighting conditions to avoid too dark or too bright environments.

[0008] Keep the distance between the checkerboard calibration board and the camera lens, and set the camera lens angle to be adjustable. The camera lens angle changes large enough to cover different rotation and translation states.

[0009] In a preferred embodiment of the present invention, in S300, using the FAST corner detection algorithm to process the calibration board in the camera video frame and pre-extract the corner coordinate information further includes: In the camera video frame, obtain one frame of image at a certain time interval; convert each obtained frame of image into a grayscale image for calculating the brightness difference; Use the FAST algorithm to detect corners on the grayscale image, and adjust the threshold parameter of the FAST algorithm according to specific requirements to control the sensitivity and quantity of corners; Suppress non-maximum values for the corners detected by the FAST algorithm to screen out the corners with the maximum response; Save the filtered corner coordinates as the corner information of the calibration board.

[0010] In a preferred embodiment of the present invention, in S400, the SIFT feature descriptor algorithm is used to automatically match the corners in different images and establish the correspondence between the calibration board and the image points. It further includes: Calculate the local feature descriptor in the image area around the key points to capture the image gradient and direction information around the key points; Compare the feature descriptor of the key point with the feature descriptors of the key points in other images. Calculate the distance between its feature descriptor and all the feature descriptors of the key points in other images using the nearest neighbor matching, and select the matching point with the closest distance as the corresponding point; Perform matching screening to remove incorrect matches and improve the accuracy of matching. Further eliminate the incorrect matching points. After the matching screening, the key points in each image are established with the corresponding relationship with the key points in other images.

[0011] In a preferred embodiment of the present invention, in S500, the Zhang Zhengyou calibration model is used to calculate the internal parameters of the camera. It further includes: Estimate the internal parameters of the camera by the least squares method according to the Zhang Zhengyou calibration model; The internal parameters include: focal lengths (fx, fy), the focal lengths of the camera in the horizontal and vertical directions, the principal point coordinates (cx, cy), the principal point coordinates on the image plane, representing the intersection point of the camera optical axis and the image plane, and distortion parameters (k1, k2, p1, p2, k3), which are used to correct the lens distortion.

[0012] In a preferred embodiment of the present invention, in S600, the evaluation and verification of the calibration result include: Reproject the three-dimensional points on the calibration board onto the image plane and calculate the difference between the projected points and the actual corner positions; Calculate the difference between the detected corner positions and the actual corner positions; Adjust the corner detection model parameters or algorithm parameters to improve the calibration result.

[0013] The calibration process is an iterative process. After parameter adjustment and model optimization, re-run the calibration process and evaluate the result.

[0014] In a preferred embodiment of the present invention, the feature points are points with obvious features; the corners of the checkerboard shape are the vertices where the horizontal lines and vertical lines intersect.

[0015] In a second aspect, an embodiment of the present invention further provides an electronic device, including: The intelligent calibration method for the panoramic vision sensor of an unmanned vehicle described in any one of the above; and A memory for storing algorithm programs; A processor for executing the algorithm programs and machine learning programs required in the calibration process and implementing the algorithm steps.

[0016] In a preferred embodiment of the present invention, it further includes a housing for storing the memory and the processor. The housing is provided with a heat dissipation port, and the heat dissipation port is covered with a dust-proof plate, and a heat dissipation fan is installed on the surface of the dust-proof plate.

[0017] According to the intelligent calibration method for the panoramic vision sensor of an unmanned vehicle in the embodiments of the present application, through intelligent algorithms and the multi-camera information fusion in four directions, the automatic calibration of the surround-view camera system is realized. The real-time acquisition and processing of the fisheye camera data, combined with the image processing algorithm, can accurately perceive the influence of factors such as the vibration and light change of the unmanned vehicle during operation on the position and direction of the camera. Through the application of the adaptive algorithm, the position and direction parameters of the camera can be corrected in real time, so as to ensure the accuracy and consistency of the acquired image data. Through automatic calibration, the present invention can reduce the errors in the relative position and direction between cameras, and eliminate the distortion and distortion phenomena of the images. Accurate image data can provide a clearer and more real environmental view, enhancing the perception ability of the unmanned vehicle to the surrounding environment. This is crucial for functions such as obstacle detection, road recognition, and lane keeping of the unmanned vehicle, and improves the safety and reliability of the unmanned vehicle under complex road conditions. The method of the present invention has strong adaptability and can adapt to various complex operating environments and working states. When the unmanned vehicle is actually running, it may be affected by factors such as vibration, light brightness change, and camera position adjustment, which will cause deviations in the calibration of the surround-view camera system; through the automatic calibration method, the present invention can correct these deviations in real time, maintain the accuracy and stability of the surround-view camera system, and has strong applicability.

[0018] Additional aspects and advantages of the present application will be given in part in the following description, become apparent in part from the following description, or be understood through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required to be used in the embodiments of the present application. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can also be obtained based on these drawings without creative efforts.

[0020] Figure 1 It is a flowchart of the intelligent calibration method for the panoramic vision sensor of an unmanned vehicle provided by an embodiment of the present invention; Figure 2 It is a top view of the principle of the intelligent calibration method for the panoramic vision sensor of the driverless vehicle provided by the embodiment of the present invention; Figure 3 It is a schematic diagram of intelligent calibration provided by the embodiment of the present invention. Specific Embodiments

[0021] Next, the technical solutions in the embodiments of the present application will be described with reference to the accompanying drawings in the embodiments of the present application.

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Apparently, the described embodiments are some but not all of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0023] Embodiment Next, an intelligent calibration method and an electronic device for a panoramic vision sensor of a driverless vehicle according to an embodiment of the present application will be described with reference to the accompanying drawings; As Figures 1-3 shown, the intelligent calibration method for the panoramic vision sensor of the driverless vehicle according to the embodiment of the present application includes: S100. Set the initial state of the driverless vehicle and the initial states of the fisheye cameras in four directions; The four directions include the front, rear, left, and right of the driverless vehicle; the initial state of the driverless vehicle includes that the four fisheye cameras can be normally started and the image information of the four directions of the driverless vehicle can be displayed in real time.

[0024] In the embodiment of the present application, the driverless vehicle is equipped with four fisheye cameras, which are respectively installed at the front, rear, left, and right positions of the vehicle, and it is necessary to ensure that the camera power supply is correctly started and the connection lines are normal.

[0025] After starting the camera, it is necessary to detect the state of each camera. This includes checking whether the camera can work normally, whether it can capture images, and whether the image quality meets the requirements. By detecting the state of the camera, it can be ensured that the camera is in a normal working state before starting the calibration.

[0026] Once the camera is normally started and passes the state detection, the system will start to display the image information of the four directions of the driverless vehicle in real time. This means that the camera will continuously capture images and transmit them to the display device for users or the system to observe and analyze.

[0027] S200. Determine a calibration board with feature points according to the size of the driverless vehicle and the shooting range, and fix it at four positions of the driverless vehicle so that the fish-eye cameras at the four positions can all capture at least one calibration board. In the embodiment of the present application, according to the size and shooting range of the driverless vehicle, a calibration board with a suitable size and obvious feature points needs to be selected. A checkerboard pattern is used for camera calibration. The feature points of the checkerboard pattern are clear, which is more conducive to improving the accuracy of the calibration process.

[0028] Determine suitable positions in the front, rear, left, and right of the driverless vehicle to install the calibration board. These positions ensure that the fish-eye cameras at the four positions can all capture at least one calibration board.

[0029] After installing and fixing the calibration board, it is necessary to verify whether each fish-eye camera can capture at least one calibration board, and observe the real-time image of the camera or perform subsequent image processing.

[0030] S300. Use a corner detection algorithm to process the calibration board in the camera video frame and pre-extract the corner coordinate information. In an alternative embodiment, refer to Figure 3 As shown, specifically use the FAST corner detection algorithm; convert the camera video frame into a grayscale image, and select a pixel in the image as the central pixel. According to the selected threshold parameter, set a grayscale difference threshold. This threshold is used to determine whether the surrounding pixels are different enough from the central pixel to be considered as corners. Check whether there is a sufficient difference in the grayscale values between the 16 pixels around the central pixel and the central pixel. If there are more than the threshold number of consecutive pixels with a large grayscale difference from the central pixel, then the central pixel is considered as a corner.

[0031] If the central pixel is detected as a corner, perform non-maximum suppression to exclude adjacent duplicate corners.

[0032] Repeat the above steps for each pixel in the image until the entire image is traversed.

[0033] For the detected corners, extract their coordinate information. Usually, pixel coordinates can be used to represent the positions of the corners, and the extracted corners are screened and corrected.

[0034] S400. Combine the corner coordinate information pre-identified by the corner detection algorithm, adopt a feature descriptor algorithm to automatically match the corners in different images, and establish the corresponding relationship between the calibration board and the image points. Specifically adopt the SIFT feature descriptor algorithm; apply the SIFT feature extraction algorithm to the calibration board image and the image to be matched respectively, and extract their SIFT feature descriptors.

[0035] Match the SIFT feature descriptors of the calibration board image and the SIFT feature descriptors of the image to be matched using the nearest neighbor matching method.

[0036] According to the distance or distance ratio between the matched feature descriptors, filter out the corner point pairs with good matching relationships.

[0037] Pair the coordinates of the calibration board corners with the coordinates of their matched image points to form a corresponding relationship.

[0038] S500. Based on the detected calibration board corners and their known positions in the world coordinate system, use the calibration model to calculate the internal parameters of the camera; Specifically, use the Zhang Zhengyou calibration model to calculate the internal parameters of the camera; Use the camera to capture calibration board images at multiple different angles to ensure that the calibration board can be clearly detected in the images. At the same time, record the pixel coordinates of the corners in each calibration board image and know the true positions of these corners in the world coordinate system.

[0039] According to the Zhang Zhengyou calibration model, it is necessary to first estimate the initial distortion parameters. It can be done through simplified methods, such as directly setting the distortion parameters to 0, or using some empirical values as the initial distortion parameters.

[0040] Use the Levenberg - Marquardt algorithm to iteratively optimize the internal parameters of the camera. Define an objective function to measure the difference between the reprojection error of the calibration board corners on the image and their true positions in the world coordinate system, and initialize the internal parameters of the camera to be optimized.

[0041] In each iteration, according to the current internal parameter estimation, calculate the reprojection position of the calibration board corners and compare it with their true positions in the world coordinate system to calculate the error.

[0042] After reaching the maximum number of iterations, output the optimal estimation result of the camera internal parameters.

[0043] The transformation calculation formula between the world coordinate system and the pixel coordinate system is as follows: Among them, represents the internal parameters of the surround - view camera, represents the external parameters of the surround - view camera.

[0044] S600. Evaluate and verify the calibration results, adjust the corner detection model parameters to improve the calibration results, obtain the bird's - eye view and projection matrix of each fisheye camera, and complete the surround - view calibration work After calibrating the intrinsic, extrinsic, and distortion parameters of the camera, it is necessary to evaluate and verify the calibration results. This can be achieved by using known points or objects in the world coordinate system and measuring and tracking them in the calibrated camera images. By calculating the difference between the measurement results and the expected results, the accuracy and precision of the calibration can be evaluated. If deviations or errors are found in the calibration results during the evaluation and verification process, the parameters of the corner detection model or algorithm can be considered for adjustment to improve the calibration results.

[0045] After parameter adjustment and improvement, the camera calibration can be performed again, and the new calibration results can be verified. This includes using the same verification dataset or test dataset for measurement and tracking to evaluate and verify the improvement effect of the calibration results. If the results still do not meet the requirements, the process of parameter adjustment and improvement can be iterated until the expected calibration accuracy and precision are achieved.

[0046] By using accurate and stable camera intrinsic and distortion parameters, the surround-view camera system of the autonomous vehicle can achieve the environmental perception function. The surround-view camera system can capture images of the surrounding environment and perform real-time analysis and processing through computer vision algorithms. In this way, the autonomous vehicle can obtain accurate information about environmental elements such as roads, obstacles, and pedestrians, thus providing more precise environmental perception capabilities.

[0047] Based on the same inventive concept, an electronic device is also provided in an embodiment of the present application. This device may be a terminal device or an embedded system to enable the normal operation of the algorithm.

[0048] The electronic device includes a memory for storing various data and programs. The memory can be an internal memory, an external memory, or a combination of both. In the present invention, the memory is used to store the machine learning algorithm program.

[0049] The electronic device also includes a processor, also known as a central processing unit (CPU) or a graphics processing unit (GPU). The processor is the core component of the electronic device and is responsible for performing various calculations and operations. In the present invention, the processor is used to run the machine learning algorithm program.

[0050] When the device is a terminal device, a housing also needs to be provided to store the memory and the processor. A heat dissipation port is provided on the housing, and the heat dissipation port is covered with a dust-proof plate. A heat dissipation fan is installed on the surface of the dust-proof plate; it is used to protect the memory and the processor and dissipate the heat generated by the device in real time.

[0051] Specifically, the invention has: 1. Improve the accuracy of environmental perception: By applying intelligent calibration technology to the panoramic camera system of the autonomous vehicle, this invention significantly enhances the environmental perception accuracy of the autonomous vehicle by utilizing precise and stable camera internal parameters and distortion coefficients. For an autonomous vehicle, efficient environmental perception ability is crucial for achieving safe navigation and obstacle avoidance, enabling the vehicle to quickly identify surrounding obstacles and other important environmental features, providing accurate data support for path planning and obstacle avoidance.

[0052] 2. Enhance the accuracy of positioning and target recognition: Precise camera internal parameters and distortion coefficients are essential for an autonomous vehicle to perform precise positioning and target recognition. The algorithms and technologies provided by this invention enable the autonomous vehicle to more accurately estimate its own position, direction, and attitude, and achieve efficient and reliable target recognition and tracking. This is crucial for the application of autonomous vehicles in fields such as logistics distribution, environmental monitoring, traffic management, etc., improving the safety and efficiency of task execution.

[0053] 3. Expand application potential: The technologies and functions provided by this invention can significantly enhance the perception, recognition, and decision-making abilities of autonomous vehicles in various environments. This will help improve the application efficiency of autonomous vehicles in fields such as logistics distribution, environmental monitoring, agricultural automation, public safety, etc. The improvement of the autonomy and intelligence level of autonomous vehicles can reduce manual intervention, improve operation efficiency, lower operating costs, and ensure the safety of personnel in certain high-risk environments. Through these innovations, autonomous vehicles can better serve society, improve production efficiency, and create greater economic and social benefits.

[0054] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed by this application can easily think of changes or substitutions, which should all be covered within the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

Claims

1. An intelligent calibration method for an unmanned vehicle panoramic vision sensor, characterized in that: include S100, setting the initial state of the unmanned vehicle and the initial states of the fisheye cameras in four directions; S200, determining a calibration plate with characteristic points according to the size of the unmanned vehicle and the shooting range, and fixing the calibration plate at four positions of the unmanned vehicle so that the fisheye cameras at the four positions can capture at least one calibration plate; S300, using a corner point detection algorithm to process the calibration plate in the camera video frame and extract corner point coordinate information in advance; S400, combining the corner point coordinate information pre-identified by the corner point detection algorithm, using a feature descriptor algorithm, automatically matching the corner points in different images, and establishing a corresponding relationship between the calibration plate and the image points; S500, based on the detected corner points of the calibration plate and their known positions in the world coordinate system, using the calibration model to calculate the internal parameters of the camera; S600, evaluating and verifying the calibration results, adjusting the corner detection model parameters to improve the calibration results, obtaining the bird's-eye view and projection matrix of each fisheye camera, and completing the surround calibration work.

2. The method for intelligent calibration of panoramic vision sensors for unmanned vehicles according to claim 1, characterized in that: In S100, the setting of the initial state of the unmanned vehicle and the initial states of the fisheye cameras in four directions includes: Set up camera installation points in advance at the front, rear, left, and right sides of the unmanned vehicle; Ensure good image quality by rotating and moving the fisheye camera to change the angle and position; ensure that the image captured by the fisheye camera covers the entire field of view.

3. The method for intelligent calibration of panoramic vision sensors for unmanned vehicles according to claim 1, characterized in that: In S200, a calibration plate with feature points is determined according to the size of the unmanned vehicle and the shooting range, including: The calibration plate with characteristic points is in the shape of a chessboard; The chessboard is a square grid formed by the interlacing of horizontal lines and vertical lines, each of which has a side length of 1 cm to 17 cm, wherein a plurality of the horizontal lines are arranged in parallel, and a plurality of the vertical lines are arranged in parallel; The chessboard is placed under lighting conditions.

4. The method for intelligent calibration of panoramic vision sensors for unmanned vehicles according to claim 1, characterized in that: In S300, the FAST corner detection algorithm is used to process the calibration plate in the camera video frame and pre-extract the corner coordinate information, which also includes: In the camera video frame, an image frame is acquired at a certain time interval; each acquired image frame is converted into a grayscale image to calculate the brightness difference; Use the FAST algorithm to detect corners on grayscale images, adjust the threshold parameters of the FAST algorithm according to specific needs, and control the sensitivity and number of corners; By performing non-maximum suppression on the corner points detected by the FAST algorithm, the corner points with the maximum response are screened out; The filtered corner point coordinates are saved as the corner point information of the calibration plate.

5. The method for intelligent calibration of panoramic vision sensors for unmanned vehicles according to claim 1, characterized in that: In S400, SIFT feature descriptor algorithm is used to automatically match corner points in different images and establish the corresponding relationship between the calibration plate and the image points, which also includes: Calculate local feature descriptors in the image area around the key points to capture image gradient and direction information around the key points; The feature descriptor of the key point is compared with the feature descriptors of the key points in other images, and the distance between its feature descriptor and all the feature descriptors of the key points in other images is calculated by nearest neighbor matching, and the matching point with the closest distance is selected as the corresponding point; Matching screening is used to remove false matches and improve matching accuracy, and to further eliminate false matching points. After matching screening, the key points in each image establish a corresponding relationship with the key points in other images.

6. The method for intelligent calibration of panoramic vision sensors for unmanned vehicles according to claim 1, characterized in that: In S500, the Zhang Zhengyou calibration model is used to calculate the internal parameters of the camera, including: According to Zhang Zhengyou's calibration model, the camera's internal parameters are estimated by the least squares method; The internal parameters include: focal length, which is the focal length of the camera in the horizontal and vertical directions, principal point coordinates, which are the coordinates of the principal point on the image plane, which represent the intersection of the camera optical axis and the image plane, and distortion parameters, which are parameters used to correct lens distortion.

7. The method for intelligent calibration of panoramic vision sensors for unmanned vehicles according to claim 3 is characterized in that: In S600, the evaluating and verifying the calibration result includes: Reprojecting the three-dimensional points on the calibration plate onto the image plane, and calculating the difference between the projected points and the actual corner point positions; Calculating the difference between the detected corner point position and the actual corner point position; Adjust the corner detection model parameters or algorithm parameters to improve the calibration results.

8. The method for intelligent calibration of panoramic vision sensors for unmanned vehicles according to claim 7, characterized in that: The characteristic points are points with obvious characteristics; the corner points of the chessboard shape are vertices where the horizontal lines and the vertical lines intersect.

9. An electronic device, characterized in that: include The intelligent calibration method of the panoramic vision sensor of an unmanned vehicle according to any one of claims 1 to 8; as well as A memory for storing algorithm programs; The processor is used to execute the algorithm program and machine learning program required in the calibration process and implement the algorithm steps.

10. An electronic device according to claim 9, characterized in that: It also includes a shell for storing the memory and the processor. The shell is provided with a heat dissipation port, the heat dissipation port is covered with a dustproof plate, and a heat dissipation fan is installed on the surface of the dustproof plate.

Images