Automobile automatic driving system capable of analyzing surrounding road conditions

Abstract

The invention relates to the technical field of automatic driving, in particular to a surrounding road condition analysis automobile automatic driving system which comprises an environment sensing data acquisition module, a data processing and edge extraction module, an object recognition module and a safety distance early warning module. The point cloud data is projected to a two-dimensional plane, then edge information is extracted, and after extraction, the object recognition module matches and calculates object categories corresponding to the object edge information in the point cloud data and the digital signals, senses the distances of different objects between corresponding automobiles in the cloud point data, and sets a similarity threshold value and a reflection threshold value; and the edge information of the object in the data processing and edge extraction module is determined again through the similarity threshold value and the reflection threshold value, so that the safe distance early warning module can more accurately know the distances between different objects and the automobile in the new digital image.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to an automobile autonomous driving system for analyzing surrounding road conditions. Background Art

[0002] The existing car autonomous driving system is an intelligent system that integrates a variety of advanced technologies and is designed to enable the car to automatically complete driving tasks without the direct intervention of human drivers. During the autonomous driving process, the camera is responsible for collecting information about objects in the road conditions around the vehicle, and the lidar is used to accurately measure the distance between the object and the vehicle. The control system will use this data to determine the driving speed according to the distance between the vehicle and different objects.

[0003] However, when a camera or lidar is used, collection deviations may occur due to the characteristics of the object; when the camera faces highly reflective metal or glass, the image is prone to light spots and flares, which interfere with the capture of features such as shape and cause recognition errors; objects with a single texture or a color similar to the background are easily confused with the background, making it difficult to accurately extract information; when the lidar encounters a strong light-absorbing object, the reflection signal is weak and the distance data error is large; objects with irregular surfaces scatter light, interfere with point cloud data, affect the judgment of the object's position and shape, and bring risks to autonomous driving decisions. In view of this, we propose an autonomous driving system for cars that analyzes surrounding road conditions. Summary of the invention

[0004] The purpose of the present invention is to solve the problem that cameras and lidars are prone to collection deviations due to the characteristics of objects. When the camera encounters highly reflective metal or glass, light spots and flares appear in the image, affecting recognition. Objects with a single texture or a color similar to the background are easily confused and difficult to extract information from. When the lidar faces strongly light-absorbing objects, the reflected signal is weak, resulting in large distance errors. Objects with irregular surfaces scatter light, interfere with point cloud data, affect the judgment of objects, and bring risks to autonomous driving decisions.

[0005] To achieve the above-mentioned purpose, the present invention provides an automobile automatic driving system for surrounding road condition analysis, comprising an environment perception data acquisition module, a data processing and edge extraction module, an object recognition module and a safety distance warning module, wherein:

[0006] The environment perception data acquisition module uses camera technology to obtain visual images around the vehicle, which include digital signals, and uses laser radar to detect the environment around the car in real time, and converts the information obtained from the detection into points in three-dimensional space, and combines the points into point cloud data;

[0007] The data processing and edge extraction module senses the point cloud data and digital signals, projects the point cloud data onto a two-dimensional plane, and then uses the Canny algorithm to extract edge information of objects in the point cloud data and digital signals;

[0008] The object recognition module adopts a template matching method to determine the category of objects in the surrounding environment through object edge information. After determining the category of objects in the surrounding environment, it senses the distance between the corresponding cars of different objects in the cloud point data, associates the object distance with the object in the digital image, generates a new digital image, and adjusts the edge information determined in the data processing and edge extraction module again in this process.

[0009] As a further improvement of the present technical solution, the camera in the environmental perception data acquisition module utilizes the principle of optical imaging. The camera lens focuses the light from objects around the vehicle onto the image sensor. The image sensor converts the light signal into an electrical signal, which is then converted into a digital signal through analog-to-digital conversion to finally form a digital image.

[0010] As a further improvement of the technical solution, the laser radar in the environment perception data acquisition module includes a transmitting module, a receiving module and a control circuit. The control circuit is used to accurately control the emission frequency and angle range of the laser pulse. The specific working principle is as follows:

[0011] The transmitting module emits laser pulses into the space around the car according to the frequency and angle range in the control circuit. The laser beam is emitted when it encounters an object in the car's surrounding environment. The receiving module receives the reflected laser signal and calculates the distance between the object in the surrounding environment and the lidar based on the time difference between laser emission and reception.

[0012] As a further improvement of the technical solution, the environmental perception data acquisition module specifically calculates the distance between the object in the surrounding environment and the laser radar as follows: ;in, is the known speed of light, It is the time from when the laser beam of the transmitting module arrives at the receiving module to when the reflected laser signal is received.

[0013] As a further improvement of the technical solution, the environment perception data acquisition module will obtain a set of distance data as the laser radar transmitting module and the receiving module continuously transmit and receive laser pulses. , and then through the control circuit in the horizontal angle and vertical angle The geometric relationship between trigonometric functions and mapping them to a three-dimensional space coordinate system In the process of combining different points into point cloud data with the same coordinates, the point cloud data generation process is as follows: Perception distance , horizontal angle and vertical angle , the corresponding three-dimensional coordinate point is obtained through the following coordinate transformation formula :

[0014] ; This formula calculates the The coordinates of the axis direction, is the distance from the object to the lidar, Used to adjust the vertical projection. Used to adjust the horizontal Projection on axis;

[0015] ;calculate The coordinates of the axis direction, To determine the horizontal direction The projection on the axis, Also used to adjust the vertical projection;

[0016] ;calculate The coordinates of the axis direction, Directly reflects the distance component in the vertical direction.

[0017] As a further improvement of the technical solution, the calculation formula of the data processing and edge extraction module for projecting the point cloud data onto a two-dimensional plane is as follows:

[0018] Perceiving point cloud data , project it onto a two-dimensional plane, assuming that the projection plane is Plane, projected point , then , .

[0019] As a further improvement of the technical solution, the working principle and calculation formula of the Canny algorithm in the data processing and edge extraction module for detecting the edge information of objects in point cloud data and digital images are consistent, and the specific working steps are as follows:

[0020] Step 1:

[0021] Gaussian function is used to smooth point cloud data and digital images to reduce the impact of noise. The calculation formula is: ,in is the standard deviation of the Gaussian distribution, The coordinates corresponding to different point cloud data and digital images;

[0022] Step 2: Use the Sobel operator to calculate the gradient strength and direction of different point cloud data and digital images. The Sobel operator has two templates in the horizontal and vertical directions. , vertical template , after convolution, the horizontal gradient is obtained and vertical gradient , gradient strength , gradient direction ;

[0023] Step 3: Set the threshold range , the gradient value is greater than Point cloud data and digital images are marked as strong edge points, and The marks between are weak edge points, strong edge points are definitely edge points, and weak edge points are considered edge points only when they are connected to strong edge points. The point cloud data and digital images are traversed, and the strong edge points and the weak edge points connected to them are connected to form the final edge information.

[0024] As a further improvement of the present technical solution, the object recognition module adopts web crawler technology to establish a template library, which contains image information of multiple objects. The working principle of web crawler technology is as follows: the web crawler first sends an HTTP request to the target website according to a given URL list, obtains the HTML code of the web page, and parses the HTML code to extract the image information in the web page.

[0025] As a further improvement of the technical solution, the object recognition module senses the contour of the object in the digital image and defines it as the contour A to be recognized, and uses the Euclidean distance calculation method to match and calculate the Euclidean distance between the contour A to be recognized and each edge information in the template library;

[0026] Specifically adjusting the edge information in the data processing and edge extraction module: setting a similarity threshold, if the Euclidean distance corresponding to the contour A to be identified and the edge information B in the template library is greater than the upper limit of the similarity threshold, the object category corresponding to the edge information B is mapped to the contour A to be identified;

[0027] Otherwise, the reflection intensity of the contour A to be identified and the reflection intensity corresponding to the edge information B in the module library are called up, and the Euclidean distance between the reflection intensity of the contour A to be identified and the edge information B is calculated again, and the reflection threshold is set;

[0028] If the Euclidean distance between the contour A to be identified and the edge information B in the template library is within the similarity threshold, and the reflection intensity Euclidean distance is greater than the reflection threshold, the object category corresponding to the edge information B is called out and mapped to the contour A to be identified.

[0029] As a further improvement of the present technical solution, the safety distance warning module senses the safety distance corresponding to different vehicle speeds in the vehicle control system, the safety distance threshold, and the distance between different objects and the vehicle in the new digital image. If the safety distance threshold is less than the distance between different objects and the vehicle, a reminder signal is output, and the output signal includes an object image and object information.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] In the automobile automatic driving system for surrounding road condition analysis, point cloud data and digital signals are sensed by a data processing and edge extraction module, and the point cloud data are projected onto a two-dimensional plane. Then, the Canny algorithm is used to extract the edge information of objects in the point cloud data and the digital signal. After extraction, the object recognition module uses a template matching method to match and calculate the object category corresponding to the object edge information in the point cloud data and the digital signal, sense the distance between different objects corresponding to the cars in the cloud point data, associate the object distance with the object in the digital image, and generate a new digital image. The object recognition module sets a similarity threshold and a reflection threshold, and the edge information of the object in the data processing and edge extraction module is determined again through the similarity threshold and the reflection threshold, so that the safety distance warning module can more accurately know the distance between different objects and the car in the new digital image. Thanks to the precise processing of the object recognition module, the distance between different objects and the car in the new digital image can be more accurately known, so that the system can timely and accurately judge the safety distance between the vehicle and the surrounding objects. When the safety distance threshold is exceeded, a warning signal is quickly issued to effectively ensure driving safety and reduce the risk of accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is the overall module principle diagram of the present invention;

[0033] Figure 2 It is a flow chart of the working principle of the present invention.

[0034] The meaning of each number in the figure is:

[0035] 100. Environmental perception data acquisition module; 200. Data processing and edge extraction module; 300. Object recognition module; 400. Safety distance warning module. DETAILED DESCRIPTION

[0036] The following will be combined with the accompanying drawings in the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0037] refer to Figure 1-Figure 2 As shown, a car automatic driving system for surrounding road condition analysis includes an environment perception data acquisition module 100, a data processing and edge extraction module 200, an object recognition module 300 and a safety distance warning module 400, wherein:

[0038] The existing car autonomous driving system is an intelligent system that integrates a variety of advanced technologies and is designed to enable the car to automatically complete driving tasks without the direct intervention of human drivers. During the autonomous driving process, the camera is responsible for collecting information about objects in the road conditions around the vehicle, and the lidar is used to accurately measure the distance between the object and the vehicle. The control system will use this data to determine the driving speed according to the distance between the vehicle and different objects.

[0039] However, when cameras or lidars are used, collection deviations may occur due to the characteristics of the objects; when the camera faces highly reflective metal or glass, light spots and flares are prone to appear in the image, interfering with the capture of shape and other features and causing recognition errors; objects with a single texture or a color similar to the background are easily confused with the background, making it difficult to accurately extract information; when the lidar encounters a strong light-absorbing object, the reflection signal is weak and the distance data error is large; objects with irregular surfaces scatter light, interfering with point cloud data, affecting the judgment of the object's position and shape, and bringing risks to autonomous driving decisions.

[0040] The environment perception data acquisition module 100 uses camera technology to obtain visual images around the vehicle. The camera uses the principle of optical imaging. The camera lens focuses the light of objects around the vehicle on the image sensor. The image sensor converts the light signal into an electrical signal, and then converts it into a digital signal through analog-to-digital conversion, and finally forms a digital image. Then, according to the RGB value of each pixel in the digital image, the luminous intensity of the corresponding sub-pixel is adjusted, various colors are mixed, and the original visual image is restored;

[0041] Optical imaging principle: The camera lens is a complex optical lens group, equivalent to a convex lens. The convex lens has a converging effect on light. Light parallel to the main optical axis converges to a point after passing through the convex lens to form a focus. According to the imaging law of the convex lens, when the object is located outside the double focal length of the convex lens, the light emitted from the same point on the object will converge to a point between one and two focal lengths on the other side of the convex lens after being refracted by the convex lens. Thus, the light will be gathered on the camera image sensor to form an inverted, reduced real image.

[0042] The image sensor is composed of a large number of pixels, each of which contains a photodiode. When light shines on the photodiode, the light signal is converted into an electrical signal according to the photoelectric effect, generating an electric charge proportional to the intensity of the incident light.

[0043] The structure and principle of the photodiode are as follows:

[0044] Photodiodes are made of semiconductor materials. The conductivity of semiconductors is between that of conductors and insulators. The electronic state inside them has a special energy band structure, with valence bands and conduction bands. Electrons in the valence band will not jump to the conduction band at room temperature, so the conductivity of semiconductors is weak.

[0045] The core structure of the photodiode is the PN junction. By doping different types of impurities in the semiconductor, P-type semiconductors and N-type semiconductors are formed, and their interface constitutes the PN junction. In the PN junction, due to the high concentration of holes in the P-type semiconductor and the high concentration of electrons in the N-type semiconductor, diffusion movement will occur, resulting in a built-in electric field near the PN junction, with its direction pointing from the N region to the P region. The built-in electric field will prevent the further diffusion movement;

[0046] The process of photoelectric effect is as follows:

[0047] Photon energy absorption: When light shines on a photodiode, the energy of the photon will be absorbed by the semiconductor material. The energy of the photon is proportional to the frequency of the light. Only when the energy of the photon is greater than the bandgap width of the semiconductor material can the electrons in the valence band absorb the photon energy and transition to the conduction band.

[0048] Electron-hole pair generation: After the electron absorbs the photon energy and jumps to the conduction band, it will leave a hole in the valence band, thus forming an electron-hole pair. The greater the intensity of the incident light, the more photons irradiate the photodiode per unit time, and the more electron-hole pairs are generated, thus establishing a proportional relationship between the light intensity and the amount of charge generated;

[0049] Carrier separation and drift: Under the action of the built-in electric field of the PN junction, the generated electrons and holes will separate. The electrons are pushed to the N region by the built-in electric field, and the holes are pushed to the P region. If the photodiode is in a reverse bias state (that is, the P region is connected to the negative pole of the power supply, and the N region is connected to the positive pole of the power supply), the external electric field will enhance the built-in electric field, so that more electron-hole pairs can be effectively separated and collected, thereby improving the sensitivity and response speed of the photodiode;

[0050] The circuit in the image sensor collects the charge generated by each pixel, samples and quantizes the analog voltage signal through an analog-to-digital converter, and converts it into a digital code. The digital signal represents the brightness and color information of each pixel in the image. The digital signal is then encoded and stored in a certain format to form a digital image.

[0051] The analog-to-digital converter is an electronic device that converts analog signals into digital signals. Its working principle is mainly based on four processes: sampling, holding, quantization and encoding. The following is a detailed introduction:

[0052] Sampling principle: the input analog signal is sampled at certain time intervals to obtain discrete analog signal samples; holding principle: after sampling, the sampled analog signal value needs to be kept for a period of time so that subsequent quantization and encoding operations can be performed under stable signal values; quantization principle: the analog signal after sampling and holding is mapped to a finite number of discrete digital levels; encoding principle: the quantized discrete level value is represented by a binary code to obtain a digital signal.

[0053] The environment perception data acquisition module 100 also uses a laser radar to detect the environment around the car in real time, and converts the detected information into points in three-dimensional space, and combines the points into point cloud data;

[0054] The laser radar in the environment perception data acquisition module 100 includes a transmitting module, a receiving module and a control circuit. The control circuit is used to accurately control the emission frequency and angle range of the laser pulse. The specific working principle is as follows:

[0055] The transmitting module emits a bunch of laser pulses to the space around the car according to the frequency and angle range in the control circuit. The laser beam is emitted when it encounters an object in the environment around the car. The receiving module receives the reflected laser signal and calculates the distance between the object and the laser radar in the surrounding environment based on the time difference between laser emission and reception. The specific calculation formula is: ;in, is the known speed of light, The time from when the laser beam of the transmitting module arrives when the receiving module receives the reflected laser signal;

[0056] As the laser radar transmitting module and receiving module continuously transmit and receive laser pulses, a set of distance , and then through the control circuit in the horizontal angle and vertical angle The geometric relationship between trigonometric functions and mapping them to a three-dimensional space coordinate system In the process of combining different points into point cloud data with the same coordinates, the point cloud data generation process is as follows: Perception distance , horizontal angle and vertical angle , the corresponding three-dimensional coordinate point is obtained through the following coordinate transformation formula :

[0057] ; This formula calculates the The coordinates of the axis direction, is the distance from the object to the lidar, Used to adjust the vertical projection. Used to adjust the horizontal Projection on axis;

[0058] ;calculate The coordinates of the axis direction, To determine the horizontal direction The projection on the axis, Also used to adjust the vertical projection;

[0059] ;calculate The coordinates of the axis direction, Directly reflects the distance component in the vertical direction.

[0060] The data processing and edge extraction module 200 senses the point cloud data and the digital signal, projects the point cloud data onto a two-dimensional plane, and then uses the Canny algorithm to extract the edge information of the object in the point cloud data and the digital signal;

[0061] The calculation formula for projecting the point cloud data onto a two-dimensional plane by the data processing and edge extraction module 200 is as follows:

[0062] Perceiving point cloud data , project it onto a two-dimensional plane, assuming that the projection plane is Plane, projected point , then , ;

[0063] Point cloud data is a large number of discrete points in three-dimensional space. Directly processing three-dimensional data requires large computational workload and high complexity. Projecting it onto a two-dimensional plane can achieve data dimensionality reduction, converting three-dimensional problems into two-dimensional problems for processing, reducing the amount of computation and processing difficulty, and improving the operating efficiency of subsequent algorithms.

[0064] The working principle of the Canny algorithm to detect the edge information of objects in point cloud data and digital images is consistent with the calculation formula. The specific working steps are as follows:

[0065] Step 1:

[0066] Gaussian function is used to smooth point cloud data and digital images to reduce the impact of noise. The calculation formula is: ,in is the standard deviation of the Gaussian distribution, The coordinates corresponding to different point cloud data and digital images;

[0067] Step 2: Use the Sobel operator to calculate the gradient strength and direction of different point cloud data and digital images. The Sobel operator has two templates in the horizontal and vertical directions. , vertical template , after convolution, the horizontal gradient is obtained and vertical gradient , gradient strength , gradient direction ;

[0068] Step 3: Set the threshold range , the gradient value is greater than Point cloud data and digital images are marked as strong edge points, and The marks between are weak edge points, strong edge points are definitely edge points, and weak edge points are considered edge points only when they are connected to strong edge points. The point cloud data and digital images are traversed, and the strong edge points and the weak edge points connected to them are connected to form the final edge information;

[0069] The edges in actual images are often not completely continuous and may be broken due to factors such as noise, illumination changes, or the characteristics of the object itself. Weak edge points may be parts of the real edge where the gradient value is reduced due to various reasons. By judging whether the weak edge points are connected to the strong edge points, these potential edge fragments can be connected to restore the continuity of the edge and present the edge of the object more completely.

[0070] The object recognition module 300 uses a template matching method to determine the category of objects in the surrounding environment through the edge information of the object. After determining the category of the object in the surrounding environment, it senses the distance between the corresponding cars of different objects in the cloud point data, associates the object distance with the object in the digital image, and generates a new digital image;

[0071] The object recognition module 300 uses web crawler technology to establish a template library, which contains image information of multiple objects, specifically: digital images, edge information corresponding to reflection intensity, and the edge information includes digital images corresponding to objects under different angles, lighting conditions, etc. The digital images of objects are objects such as logos and signs during the driving process of the car;

[0072] The working principle of web crawler technology is as follows: the web crawler first sends HTTP requests to the target website based on the given URL list, obtains the HTML code of the web page, and parses the HTML code to extract the image information in the web page;

[0073] The web crawler technology is used to establish a template library containing digital images and edge information of objects at different angles and lighting conditions, which provides a rich and diverse reference samples for object recognition. In the actual recognition process, it can more comprehensively cover various possible object appearance situations, thereby improving the accuracy of object recognition. The existence of the template library enables the recognition process to directly perform matching calculations, avoiding complex feature extraction and analysis from scratch each time, greatly improving the recognition efficiency. It is especially suitable for car driving scenarios that require real-time processing.

[0074] The object recognition module 300 senses the contour of the object in the digital image and defines it as the contour A to be recognized. The Euclidean distance calculation method is used to match and calculate the Euclidean distance between the contour A to be recognized and each edge information in the template library.

[0075] Set a similarity threshold. If the Euclidean distance between the contour A to be identified and the edge information B in the template library is greater than the upper limit of the similarity threshold, the object category corresponding to the edge information B is mapped to the contour A to be identified.

[0076] The contour A to be identified is points, expressed as ,in ; The edge information B in the template library is composed of points, expressed as ,in , the Euclidean distance between the contour A to be identified and the edge information B is: ;

[0077] Otherwise, the reflection intensity of the contour A to be identified and the reflection intensity corresponding to the edge information B in the module library are called up, and the Euclidean distance between the reflection intensity of the contour A to be identified and the edge information B is calculated again, and the reflection threshold is set;

[0078] If the Euclidean distance between the contour A to be identified and the edge information B in the template library is within the similarity threshold, and the reflection intensity Euclidean distance is greater than the reflection threshold, the object category corresponding to the edge information B is called out and mapped to the contour A to be identified;

[0079] Through template matching and Euclidean distance calculation, combined with edge information and reflection intensity matching, the object category corresponding to the outline to be identified can be determined more accurately, which helps the automotive system to fully and accurately perceive objects in the surrounding environment and provide a reliable basis for subsequent decision-making.

[0080] By associating object distances with objects in digital images and generating new digital images, the automotive system can build an environmental model that includes object categories and location information, presenting the surrounding environment more intuitively for further analysis and processing.

[0081] The safety distance warning module 400 senses the safety distance corresponding to different vehicle speeds in the vehicle control system, the safety distance threshold, and the distance between different objects and the vehicle in the new digital image. If the safety distance threshold is less than the distance between different objects and the vehicle, a reminder signal is output, and the output signal includes information such as the object image, object shape, and material. The reminder signal includes information such as the object image, object shape, and material, which can provide a more detailed reference for the driver or the vehicle control system to help them better assess the degree of danger and formulate a response strategy. For example, objects of different shapes and materials may have different collision consequences and avoidance methods, and this information can make decisions more accurate and effective.

[0082] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An automatic driving system for a car based on surrounding road conditions analysis, characterized in that: The system comprises an environment perception data collection module (100), a data processing and edge extraction module (200), an object recognition module (300) and a safety distance warning module (400), wherein: The environment perception data acquisition module (100) uses camera technology to obtain visual images around the vehicle, the visual images including digital signals, and uses laser radar to detect the environment around the vehicle in real time, and converts the information obtained from the detection into points in three-dimensional space, and combines the points into point cloud data; The data processing and edge extraction module (200) senses the point cloud data and the digital signal, projects the point cloud data onto a two-dimensional plane, and then uses the Canny algorithm to extract edge information of objects in the point cloud data and the digital signal; The object recognition module (300) uses a template matching method to determine the category of objects in the surrounding environment through object edge information. After determining the category of objects in the surrounding environment, the distance between the corresponding cars of different objects in the cloud point data is sensed, the object distance is associated with the object in the digital image, a new digital image is generated, and in this process, the edge information determined in the data processing and edge extraction module (200) is adjusted again.

2. The automobile automatic driving system for surrounding road condition analysis according to claim 1, characterized in that: The camera in the environment perception data acquisition module (100) uses the principle of optical imaging. The camera lens focuses the light from objects around the vehicle onto the image sensor. The image sensor converts the light signal into an electrical signal, which is then converted into a digital signal through analog-to-digital conversion, ultimately forming a digital image.

3. The automobile automatic driving system for surrounding road condition analysis according to claim 2, characterized in that: The laser radar in the environment perception data acquisition module (100) comprises a transmitting module, a receiving module and a control circuit. The control circuit is used to accurately control the emission frequency and angle range of laser pulses. The specific working principle is as follows: The transmitting module emits laser pulses into the space around the car according to the frequency and angle range in the control circuit. The laser beam is emitted when it encounters an object in the car's surrounding environment. The receiving module receives the reflected laser signal and calculates the distance between the object in the surrounding environment and the lidar based on the time difference between laser emission and reception.

4. The automobile automatic driving system for surrounding road condition analysis according to claim 3 is characterized in that: The environment perception data acquisition module (100) specifically calculates the distance between an object in the surrounding environment and the laser radar using the following calculation formula: ;in, is the known speed of light, It is the time from when the laser beam of the transmitting module arrives at the receiving module to when the reflected laser signal is received.

5. The automobile automatic driving system for surrounding road condition analysis according to claim 4 is characterized in that: The environment perception data acquisition module (100) obtains a set of distance data as the laser radar transmitting module and the receiving module continuously transmit and receive laser pulses. , and then through the control circuit in the horizontal angle and vertical angle The geometric relationship between trigonometric functions and mapping them to a three-dimensional space coordinate system In the process of combining different points into point cloud data with the same coordinates, the point cloud data generation process is as follows: Perception distance , horizontal angle and vertical angle , the corresponding three-dimensional coordinate point is obtained through the following coordinate transformation formula : ； This formula calculates the The coordinates of the axis direction, is the distance from the object to the lidar, Used to adjust the vertical projection. Used to adjust the horizontal Projection on axis; ； calculate The coordinates of the axis direction, To determine the horizontal direction The projection on the axis, Also used to adjust the vertical projection; ;calculate The coordinates of the axis direction, Directly reflects the distance component in the vertical direction.

6. The automobile automatic driving system for surrounding road condition analysis according to claim 1, characterized in that: The calculation formula for projecting the point cloud data onto a two-dimensional plane by the data processing and edge extraction module (200) is as follows: Perceiving point cloud data , project it onto a two-dimensional plane, assuming that the projection plane is Plane, projected point , then , .

7. The automobile automatic driving system for surrounding road condition analysis according to claim 6, characterized in that: The working principle and calculation formula of the Canny algorithm in the data processing and edge extraction module (200) for detecting the edge information of objects in point cloud data and digital images are consistent, and the specific working steps are as follows: Step 1: Gaussian function is used to smooth point cloud data and digital images to reduce the impact of noise. The calculation formula is: ,in is the standard deviation of the Gaussian distribution, The coordinates corresponding to different point cloud data and digital images; Step 2: Use the Sobel operator to calculate the gradient strength and direction of different point cloud data and digital images. The Sobel operator has two templates in the horizontal and vertical directions. , vertical template , after convolution, the horizontal gradient is obtained and vertical gradient , gradient strength , gradient direction ; Step 3: Set the threshold range , the gradient value is greater than Point cloud data and digital images are marked as strong edge points, and The marks between are weak edge points, strong edge points are definitely edge points, and weak edge points are considered edge points only when they are connected to strong edge points. The point cloud data and digital images are traversed, and the strong edge points and the weak edge points connected to them are connected to form the final edge information.

8. The automobile automatic driving system for surrounding road condition analysis according to claim 1, characterized in that: The object recognition module (300) uses web crawler technology to establish a template library, which contains image information of multiple objects. The working principle of the web crawler technology is as follows: the web crawler first sends an HTTP request to the target website according to a given URL list, obtains the HTML code of the web page, and parses the HTML code to extract the image information in the web page.

9. The automobile automatic driving system for surrounding road condition analysis according to claim 8, characterized in that: The object recognition module (300) senses the contour of an object in a digital image and defines it as a contour A to be recognized, and uses a Euclidean distance calculation method to match and calculate the Euclidean distance between the contour A to be recognized and each edge information in the template library; Specifically adjusting the edge information in the data processing and edge extraction module (200): setting a similarity threshold, if the Euclidean distance between the contour A to be identified and the edge information B in the template library is greater than the upper limit of the similarity threshold, then the object category corresponding to the edge information B is mapped to the contour A to be identified; Otherwise, the reflection intensity of the contour A to be identified and the reflection intensity corresponding to the edge information B in the module library are called up, and the Euclidean distance between the reflection intensity of the contour A to be identified and the edge information B is calculated again, and the reflection threshold is set; If the Euclidean distance between the contour A to be identified and the edge information B in the template library is within the similarity threshold, and the reflection intensity Euclidean distance is greater than the reflection threshold, the object category corresponding to the edge information B is called out and mapped to the contour A to be identified.

10. The automobile automatic driving system for surrounding road condition analysis according to claim 1, characterized in that: The safety distance warning module (400) senses the safety distances corresponding to different vehicle speeds in the vehicle control system, the safety distance threshold, and the distances between different objects and the vehicle in the new digital image. If the safety distance threshold is less than the distance between the different objects and the vehicle, a warning signal is output, wherein the output signal includes an object image and object information.

Images