Lane identification method and system for autonomous vehicle

Through a variety of sensors and advanced algorithms, the lane model is constructed, which solves the problem of lane identification error in autonomous vehicles under harsh conditions, achieves higher driving safety and navigation accuracy, and improves driving efficiency and the efficiency of intelligent transportation systems. Lay the foundation for the development of autonomous driving technology.

CN119037440AInactive Publication Date: 2024-11-29JIANGSU QUESTYLE WISDOM NETWORK TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411453544.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-17
Publication Date
2024-11-29
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The accuracy of lane identification in autonomous vehicles in severe weather and complex road conditions is limited by sensor technology, resulting in error and real-time problems, affecting driving safety and efficiency.

Method used

Using a variety of sensors such as high-definition cameras, lidar and millimeter-wave radar, combined with data preprocessing, computer vision and deep learning algorithms, lane models are built and continuous tracking is used with Kalman filters to ensure the accuracy and stability of lane recognition.

Benefits of technology

It improves driving safety and navigation accuracy, reduces the burden on drivers, improves driving efficiency and the efficiency of intelligent transportation systems, and lays the foundation for the development of autonomous driving technology.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention relates to the technical field of automatic driving, in particular to a lane identification method and system for an automatic driving vehicle, and solves the problems that the accuracy of lane identification in the prior art is limited by a sensor technology, and particularly under severe weather or complex road conditions, the sensing ability of a sensor is greatly reduced, and the lane identification accuracy is influenced. And errors in lane identification are caused. According to the lane recognition method and system of the automatic driving vehicle, all-directional road and environment information is captured at different positions of the vehicle, a camera is used for capturing visible light images, especially lane lines, traffic signs and road surface conditions, and a laser radar provides accurate three-dimensional space information and recognizes road edges and obstacles. According to the method, the driving safety is improved, and the automatic driving vehicle can more accurately judge the road boundary and the obstacle through accurate lane recognition, so that the situation that the vehicle deviates from the lane or collides with the obstacle is avoided, and the driving safety is greatly improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a lane recognition method and system for an autonomous driving vehicle. Background Art

[0002] Autonomous vehicles, also known as driverless cars, are intelligent vehicles that integrate advanced sensors, algorithms and control systems to enable cars to autonomously perceive the surrounding environment, make decisions and perform driving tasks without direct control by human drivers. Autonomous vehicles have broad application prospects. In mining areas, sanitation, public transportation, taxis, logistics and distribution, autonomous vehicles are expected to replace traditional manned vehicles and improve work efficiency and safety. For example, in mining areas, autonomous vehicles can achieve unmanned operation, reduce personnel costs and safety risks; in the field of sanitation, autonomous sanitation vehicles can efficiently complete tasks such as road cleaning and watering; in the field of public transportation and taxis, autonomous vehicles can provide more convenient and comfortable travel services. Although autonomous vehicles have great potential, they still face many challenges in practical applications. First, technical problems such as the limitations of perception capabilities in bad weather and complex road conditions, the balance between safety and comfort in decision-making planning, and the stability and robustness of the control system need to be further resolved. Second, the promotion of autonomous vehicles also requires support from laws, ethics and infrastructure. At present, the legal provisions of various countries on autonomous driving are still unclear, and ethical issues such as decision-making choices in emergencies also need to be discussed. In addition, the construction of infrastructure such as high-precision maps and communication networks and the standardization of autonomous driving vehicles are also important factors restricting its development.

[0003] The accuracy of lane recognition in existing technologies is limited by sensor technology, especially in bad weather or complex road conditions, the sensor's perception ability will be greatly reduced, resulting in errors in lane recognition. LiDAR is easily disturbed in bad weather such as rain, snow and fog, while the camera's performance is limited when the lighting conditions are poor. These limitations directly affect the accuracy and stability of lane recognition. Autonomous vehicles face challenges when dealing with non-standard lanes or complex road conditions. Current lane recognition technology is mainly based on machine learning and image processing algorithms. These algorithms have good recognition effects for standard lanes, but the recognition effect will be greatly reduced when facing non-standard lanes, damaged roads, temporary construction and other situations. This may cause autonomous vehicles to misjudge or fail to accurately identify lanes under complex road conditions. The real-time performance of lane recognition is also a problem. In order to ensure driving safety, autonomous vehicles need to accurately identify lanes in real time so that they can adjust their driving trajectories in a timely manner. However, current lane recognition technology may have a certain delay when processing a large amount of image data, which may affect the real-time decision-making and response speed of the vehicle. The reliability of lane recognition is also affected by hardware equipment and software algorithms. Factors such as performance differences, installation locations and calibration accuracy of sensors and cameras may affect the results of lane recognition. At the same time, the complexity and robustness of the software algorithm are also key factors in determining the reliability of lane recognition. Summary of the invention

[0004] The purpose of the present invention is to provide a lane recognition method and system for an autonomous driving vehicle, which solves the problem in the prior art that the accuracy of lane recognition is limited by sensor technology, especially in severe weather or complex road conditions, the sensor's perception ability will be greatly reduced, resulting in errors in lane recognition.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A lane recognition method and system for an autonomous driving vehicle, including data collection, data preprocessing, lane line detection, lane model construction, lane tracking, decision-making and control. A variety of sensors, including high-definition cameras, laser radars, millimeter-wave radars, etc., are installed at different positions of the vehicle to capture all-round road and environmental information. The camera is used to capture visible light images, especially lane lines, traffic signs and road conditions. The laser radar provides accurate three-dimensional spatial information, which helps to identify road edges and obstacles. All sensor data are transmitted to a central processing unit in real time through the vehicle's internal network. The collected raw image data is preprocessed, including denoising, contrast enhancement, image correction and other operations to improve image quality and facilitate subsequent lane line detection. Computer vision and image processing technologies, such as edge detection, Hough transform or more advanced deep learning algorithms, are applied to identify and extract lane line features in the image.

[0007] Preferably, the image captured by the camera is subjected to distortion correction to eliminate image distortion caused by the lens, the image contrast and brightness are adjusted to ensure that lane lines can be clearly identified under different lighting conditions, and the lidar data is filtered to remove unnecessary noise and stray points to improve data quality.

[0008] Preferably, image processing techniques, such as Canny edge detection or Sobel operator, are applied to highlight edge information in the image to facilitate lane line identification. Hough transform or line detection algorithm is used to locate and extract straight line features in the image, which are likely to represent lane lines. In some advanced systems, deep learning models, such as convolutional neural networks, may be used to more accurately identify and classify lane lines.

[0009] Preferably, based on the detected lane line feature points, a mathematical model of the lane line is constructed through a fitting algorithm, usually a linear or quadratic curve model, and the lane width, curvature and direction are estimated according to the model parameters. The lane is located in the global coordinate system in combination with the vehicle's own positioning information.

[0010] Preferably, an algorithm such as a Kalman filter or a particle filter is used to continuously track the lane line in combination with historical data and current sensor readings. During vehicle driving, the lane model is continuously updated to adapt to road changes, such as curves and forks. By comparing the changes in lane line position between consecutive frames, the relative displacement and heading angle changes of the vehicle are estimated.

[0011] Preferably, based on the results of lane recognition and the current state of the vehicle, the decision layer of the autonomous driving system will plan the next driving trajectory, and the control algorithm will generate corresponding control instructions, steering angle, throttle opening and braking force according to the planned trajectory. These instructions are executed through the vehicle control system, enabling the vehicle to autonomously maintain stable driving in the identified lane.

[0012] The present invention has at least the following beneficial effects:

[0013] 1. Improve driving safety. Through precise lane recognition, autonomous vehicles can more accurately judge road boundaries and obstacles, thereby avoiding lane deviation or collision with obstacles, greatly improving driving safety.

[0014] 2. Achieve precise navigation: Lane recognition technology provides accurate navigation information for autonomous vehicles, enabling them to travel along predetermined routes and reduce the risk of getting lost or deviating from their destination.

[0015] 3. Improve driving efficiency. Autonomous driving vehicles can independently select the optimal driving path through lane recognition, avoiding congestion and unnecessary detours, thereby improving driving efficiency and saving time and energy.

[0016] 4. Reduce the burden on drivers. Lane recognition technology can significantly reduce the operating burden on drivers, making driving easier and more convenient. During long-distance driving, this can effectively relieve driver fatigue.

[0017] 5. Promote the development of intelligent transportation systems. Accurate lane recognition is an important part of intelligent transportation systems. By communicating with other vehicles and transportation infrastructure, autonomous vehicles can use road resources more efficiently and reduce traffic accidents and congestion.

[0018] 6. Lay the foundation for the comprehensive development of autonomous driving technology in the future. Lane recognition technology is one of the key links in the development of autonomous driving technology. With the continuous advancement and improvement of technology, lane recognition will lay the foundation for higher levels of autonomous driving and promote the rapid development of the entire autonomous driving industry. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0020] Embodiment 1

[0021] It includes data collection, data preprocessing, lane line detection, lane model construction, lane tracking, decision-making and control. A variety of sensors are installed, including high-definition cameras, lidar, millimeter-wave radar, etc., at different positions of the vehicle to capture all-round road and environmental information. Cameras are used to capture visible light images, especially lane lines, traffic signs and road conditions. Lidar provides accurate three-dimensional spatial information, which helps to identify road edges and obstacles. All sensor data are transmitted to the central processing unit in real time through the vehicle's internal network. The collected raw image data is preprocessed, including denoising, contrast enhancement, image correction and other operations to improve image quality and facilitate subsequent lane line detection. Computer vision and image processing techniques, such as edge detection, Hough transform or more advanced deep learning algorithms, are used to identify and extract lane line features in images.

[0022] Driving safety is achieved. Through precise lane recognition, autonomous vehicles can judge road boundaries and obstacles more accurately, thereby avoiding lane deviation or collision with obstacles, greatly improving driving safety and achieving precise navigation. Lane recognition technology provides autonomous vehicles with accurate navigation information, allowing vehicles to travel along predetermined routes and reduce the risk of getting lost or deviating from the destination.

[0023] Embodiment 2

[0024] The images captured by the camera are distorted to eliminate image distortion caused by the lens, and the image contrast and brightness are adjusted to ensure that lane lines can be clearly identified under different lighting conditions. The lidar data is filtered to remove unnecessary noise and stray points to improve data quality.

[0025] Embodiment 3

[0026] Apply image processing techniques, such as Canny edge detection or Sobel operator, to highlight edge information in the image to facilitate lane line identification. Use Hough transform or line detection algorithm to locate and extract straight line features in the image. These features are likely to represent lane lines. In some advanced systems, deep learning models, such as convolutional neural networks, may be used to more accurately identify and classify lane lines.

[0027] Embodiment 4

[0028] Based on the detected lane feature points, a mathematical model of the lane is constructed through a fitting algorithm, usually a linear or quadratic curve model. The width, curvature and direction of the lane are estimated according to the model parameters, and the lane is located in the global coordinate system in combination with the vehicle's own positioning information.

[0029] Embodiment 5

[0030] Using algorithms such as Kalman filters or particle filters, combined with historical data and current sensor readings, the lane lines are continuously tracked. During the vehicle's driving process, the lane model is continuously updated to adapt to road changes, such as curves and forks. By comparing the changes in lane line position between consecutive frames, the relative displacement and heading angle changes of the vehicle are estimated.

[0031] Embodiment 6

[0032] Based on the results of lane recognition and the current state of the vehicle, the decision-making layer of the autonomous driving system will plan the next driving trajectory. The control algorithm will generate corresponding control instructions based on the planned trajectory, including steering angle, throttle opening and braking force. These instructions are executed through the vehicle control system, enabling the vehicle to autonomously maintain stable driving in the identified lane.

[0033] Improve driving efficiency. Autonomous vehicles can autonomously select the best driving path through lane recognition, avoid congestion and unnecessary detours, thereby improving driving efficiency, saving time and energy, and reducing the burden on drivers. Lane recognition technology can greatly reduce the driver's operating burden, making driving easier and more convenient. In long-distance driving, this can effectively relieve driver fatigue and promote the development of intelligent transportation systems. Accurate lane recognition is an important part of intelligent transportation systems. Through communication with other vehicles and transportation infrastructure, autonomous vehicles can use road resources more efficiently, reduce traffic accidents and congestion, and lay the foundation for the comprehensive development of autonomous driving technology in the future. Lane recognition technology is one of the key links in the development of autonomous driving technology. With the continuous advancement and improvement of technology, lane recognition will lay the foundation for higher levels of autonomous driving and promote the rapid development of the entire autonomous driving industry.

[0034] 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 to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A lane recognition method and system for an autonomous driving vehicle, including data acquisition, data preprocessing, lane line detection, lane model construction, lane tracking, decision-making and control. A variety of sensors, including high-definition cameras, laser radars, millimeter-wave radars, etc., are installed at different positions of the vehicle to capture all-round road and environmental information. The camera is used to capture visible light images, especially lane lines, traffic signs and road conditions. The laser radar provides accurate three-dimensional spatial information, which helps to identify road edges and obstacles. All sensor data are transmitted to the central processing unit in real time through the vehicle's internal network. The collected raw image data is preprocessed, including denoising, contrast enhancement, image correction and other operations to improve image quality and facilitate subsequent lane line detection. Computer vision and image processing techniques, such as edge detection, Hough transform or more advanced deep learning algorithms, are applied to identify and extract lane line features in the image.

2. A lane recognition method and system for an autonomous driving vehicle according to claim 1, characterized in that: The image captured by the camera is distorted and corrected to eliminate image distortion caused by the lens, and image contrast and brightness are adjusted to ensure that lane lines can be clearly identified under different lighting conditions. The lidar data is filtered to remove unnecessary noise and stray points to improve data quality.

3. The lane recognition method and system for an autonomous driving vehicle according to claim 1, characterized in that: The image processing technology, Canny edge detection or Sobel operator, is used to highlight the edge information in the image to facilitate the identification of lane lines. The Hough transform or line detection algorithm is used to locate and extract straight line features in the image. These features are likely to represent lane lines. In some advanced systems, deep learning models, convolutional neural networks, may be used to more accurately identify and classify lane lines.

4. The lane recognition method and system for an autonomous driving vehicle according to claim 1, characterized in that: Based on the detected lane line feature points, a mathematical model of the lane line is constructed through a fitting algorithm, which is usually a linear or quadratic curve model. The width, curvature and direction of the lane are estimated according to the model parameters, and the lane is located in the global coordinate system in combination with the vehicle's own positioning information.

5. The lane recognition method and system for an autonomous driving vehicle according to claim 1, characterized in that: The method uses algorithms such as Kalman filter or particle filter, combines historical data and current sensor readings, to continuously track lane lines. During vehicle driving, the lane model is continuously updated to adapt to road changes, such as curves and forks. By comparing the changes in lane line positions between consecutive frames, the relative displacement and heading angle changes of the vehicle are estimated.

6. The lane recognition method and system for an autonomous driving vehicle according to claim 1, characterized in that: Based on the results of lane recognition and the current state of the vehicle, the decision-making layer of the autonomous driving system will plan the next driving trajectory, and the control algorithm will generate corresponding control instructions, steering angle, throttle opening and braking force according to the planned trajectory. These instructions are executed through the vehicle control system, enabling the vehicle to autonomously maintain stable driving in the identified lane.