CN-119986696-B - Car environment perception obstacle avoidance system based on visual image technology

CN119986696BCN 119986696 BCN119986696 BCN 119986696BCN-119986696-B

Abstract

The invention provides an automobile environment sensing obstacle avoidance system based on a visual image technology, which comprises the steps of starting laser radar scanning when an automobile runs, constructing a running map of the automobile by combining with a SLAM technology, determining a plurality of environment objects within a specified range of the automobile, acquiring running data of the automobile, marking the running data in the running map to synchronously position the automobile, determining collision probability between the automobile and each environment object, screening a plurality of running obstacles in the running map, respectively establishing an obstacle avoidance plan according to obstacle attribute corresponding to each running obstacle, assisting in obstacle avoidance, recording an obstacle avoidance process of a driver, determining driving preference of the driver through training, determining preference degrees of the driver for different obstacle avoidance modes, generating an obstacle avoidance preference suggestion, displaying, and identifying and tracking other automobiles, pedestrians and obstacles in real time, thereby helping the driver avoid potential collision.

Inventors

CHEN ANZHU
JIANG SHUYING
BIAN YUXUAN
XU HAOMING
Yue kexin
TONG YI

Assignees

盐城工业职业技术学院

Dates

Publication Date: 20260508
Application Date: 20250115

Claims (8)

1. An automobile environment perception obstacle avoidance system based on visual image technology, which is characterized by comprising: the environment sensing module is used for starting laser radar scanning when the automobile runs, constructing a running map of the automobile by combining with SLAM technology, and determining a plurality of environment objects within a specified range of the automobile; The perception analysis module is used for acquiring the driving data of the automobile and marking the driving data in the driving map to synchronously position the automobile, and determining a first collision probability between the automobile and each environmental object; The auxiliary obstacle avoidance module is used for screening a plurality of driving obstacles in the driving map based on the first collision probability, establishing an obstacle avoidance plan according to the obstacle attribute corresponding to each driving obstacle and carrying out auxiliary obstacle avoidance; The obstacle avoidance training module is used for recording an obstacle avoidance process of a driver, training the obstacle avoidance process to determine driving preference of the driver, determining preference degrees of the driver on different obstacle avoidance modes, generating an obstacle avoidance preference suggestion and displaying the obstacle avoidance preference suggestion; Wherein the context awareness module comprises: The synchronous scanning unit is used for carrying out synchronous scanning when the automobile is started to obtain environment information in a specified range of the automobile, and carrying out de-distortion treatment on the environment information to obtain a synchronous induction scene in the specified range of the automobile; The map construction unit is used for carrying out contour recognition on the synchronous induction scene by utilizing the SLAM technology to obtain a plurality of object contours in a specified range, and constructing a running map in the specified range by combining the history induction scene; The article identification unit is used for searching appearance attributes corresponding to the article outlines respectively, carrying out depth identification on the article outlines by utilizing the appearance attributes, and determining a plurality of environmental articles in the specified range; Wherein, the map construction unit includes: The contour recognition subunit is used for positioning a plurality of key points in the synchronous induction scene by utilizing the SLAM technology, respectively carrying out radial diffusion search by taking each key point as a center to obtain cross information between the synchronous induction scene and each search line, and establishing a plurality of sparse contours of the synchronous induction scene; The profile perfecting subunit is used for carrying out boundary expansion on each sparse profile to obtain corresponding complete profiles, obtaining smoothness corresponding to each complete profile, determining a filtering level of each complete profile based on the smoothness, and carrying out classified filtering on the complete profiles to obtain a plurality of object profiles in the synchronous induction scene; the characteristic deep exploration subunit is used for respectively identifying each article contour in the history induction scene to obtain original characteristics, repeated characteristics and unique characteristics corresponding to each article contour; And the map construction subunit is used for determining the fixed position of the corresponding object outline in the specified range according to the original characteristics, determining the distribution position of the corresponding object outline in the specified range according to the repeated characteristics, determining the active position of the corresponding object outline in the specified range according to the unique characteristics and generating the running map in the specified range.
2. The visual image technology-based automotive environment-aware obstacle avoidance system of claim 1, further comprising: the in-car obstacle avoidance module is used for acquiring live information in a specified range of the car when the car stops; determining a plurality of dynamic objects within a specified range of the automobile by utilizing the live information, and respectively establishing object activity characteristics within a stopping range of the automobile; Determining a second collision probability between each dynamic object and each door of the automobile based on the object activity characteristics; And when the second collision probability is larger than the safety probability, positioning the corresponding dangerous vehicle door and locking the dangerous vehicle door.
3. The visual image technology-based automotive environment-aware obstacle avoidance system of claim 2, further comprising: the non-straight obstacle avoidance module is used for determining the advancing direction of the automobile according to the state of the steering lamp of the automobile; acquiring road condition information related to the advancing direction; and generating and displaying the automobile auxiliary forward advice according to the road condition information.
4. The vehicle environment sensing obstacle avoidance system of claim 1 wherein said sensing analysis module comprises: the acquisition processing unit is used for acquiring the running data of the automobile, establishing a running track of the automobile according to the running data, and marking the running track in the running map to obtain the overlapping information of the automobile and the ground surface; the depth positioning unit is used for acquiring a plurality of overlapping positions of the overlapping information in the running map, respectively acquiring road characteristics corresponding to each overlapping position, synchronously positioning the automobile by utilizing the road characteristics, and determining the real-time position of the automobile; And the collision analysis unit is used for respectively acquiring the corresponding activity speed characteristic and activity direction characteristic of each environmental object, and determining a first collision probability between the environmental object and each environmental object based on the advancing direction of the automobile.
5. The visual image technology-based automotive environment-aware obstacle avoidance system of claim 4, wherein the collision analysis unit comprises: the characteristic analysis subunit is used for respectively positioning each environmental object in the running map, collecting the corresponding position change information of the environmental object at different moments, and constructing the corresponding activity speed characteristic and activity direction characteristic of each environmental object; The space analysis subunit is used for establishing an automobile space vector according to the advancing direction and the advancing speed of the automobile, screening target environment objects corresponding to target activity direction characteristics crossed with the automobile space vector, and establishing object space vectors corresponding to each environment object by utilizing the target activity speed characteristics and the target activity direction characteristics corresponding to the target environment objects; And the collision analysis subunit is used for respectively acquiring the coincidence time between each article space vector and the automobile space vector, constructing a first collision probability between the automobile and the target environment article according to the coincidence time, and considering the first collision probability of the environment article which is not matched with the collision probability as 0.
6. The vehicle environment-aware obstacle avoidance system of claim 1, wherein the auxiliary obstacle avoidance module comprises: the obstacle avoidance preprocessing unit is used for respectively positioning each driving obstacle in the driving map, sequencing the driving obstacles according to the sequence from the high collision probability to the low collision probability, and generating an obstacle avoidance queue of the automobile; The obstacle avoidance pre-analysis unit is used for respectively acquiring the obstacle attribute corresponding to each obstacle, determining the residual obstacle avoidance distance between the automobile and each driving obstacle according to the obstacle attribute, and determining the success probability of the automobile to avoid each driving obstacle in the current lane by combining the obstacle avoidance queues; the auxiliary obstacle avoidance unit is used for searching a target driving obstacle with success probability lower than the standard probability, carrying out lane change avoidance on the target driving obstacle, and establishing and displaying an obstacle avoidance plan of the automobile by combining avoidance measures corresponding to the automobile in a current lane.
7. The vehicle environment-aware obstacle avoidance system based on visual imaging technology of claim 1, wherein the obstacle avoidance training module comprises: The obstacle avoidance recording unit is used for recording the obstacle avoidance process of a driver and determining the avoidance mode of the driver on each driving obstacle; The similar training unit is used for classifying the avoidance modes according to the obstacle attribute to generate a corresponding avoidance measure set, and screening a plurality of habit actions of the driver, the repetition frequency of which is higher than a preset frequency, from the avoidance measure set; the preference analysis unit is used for respectively simulating each habit action, obtaining a driving result corresponding to each habit action, generating a plurality of preference obstacle avoidance modes, and determining the preference degree of each preference obstacle avoidance mode according to the corresponding repetition frequency; The obstacle avoidance optimization unit is used for matching the target result of the obstacle avoidance with the driving result when the obstacle avoidance is performed, determining a corresponding preferable obstacle avoidance mode according to the matching result, generating an obstacle avoidance optimization suggestion according to the preferable obstacle avoidance mode and displaying the obstacle avoidance optimization suggestion.
8. The visual image technology-based automotive environment-aware obstacle avoidance system of claim 1, further comprising: the road condition analysis module is used for determining the smooth characteristic of the current road where the automobile is located according to the running map; and when the smooth characteristic is abnormal, determining that the collapse phenomenon exists on the current road, and generating early warning information.

Description

Car environment perception obstacle avoidance system based on visual image technology Technical Field The invention relates to the technical field of automobile auxiliary obstacle avoidance, in particular to an automobile environment perception obstacle avoidance system based on a visual image technology. Background Vehicles, one of the most common vehicles in our daily lives, play an increasingly important role in our lives. With the technological advances in the areas of artificial intelligence, sensors, control, drive, and materials, the technology for producing automobiles has matured, such that automobiles have changed from a former luxury to a consumer product of everyday life. Due to the improvement of the automobile storage amount, a plurality of people cannot avoid collision with the automobile body due to insufficient driving experience in the running process of the automobile. According to the information, when the automobile running at high speed encounters an obstacle, the response time of a driver is very short, and the collision is easily caused by untimely avoidance, so that traffic accidents are caused, and the intelligent automobile for assisting in avoiding the obstacle is assembled, has decision making and execution capability for avoiding dangerous factors, can timely make obstacle avoidance response when the danger comes, and avoids the collision of the automobile. It is necessary to design an environment-aware obstacle avoidance system to address these safety issues. The front automobile obstacle avoidance system is designed only aiming at a kinematic model of an automobile in low-speed running, and in a high-speed running state of the automobile, the cycle dynamics characteristics in the automobile are required to be considered, so that the automobile obstacle avoidance system is required to be calculated through a complex physical power model in order to accurately calculate the running state of the automobile, the response time of the automobile obstacle avoidance is increased, the risk of automobile collision is increased, the safety performance is poor, and the use limitation is relatively high. Therefore, the invention provides an automobile environment sensing obstacle avoidance system based on a visual image technology. Disclosure of Invention The automobile environment sensing obstacle avoidance system based on the visual image technology can identify and track other automobiles, pedestrians and obstacles in real time, so that a driver is helped to avoid potential collision. The invention provides an automobile environment sensing obstacle avoidance system based on a visual image technology, which comprises: the environment sensing module is used for starting laser radar scanning when the automobile runs, constructing a running map of the automobile by combining with SLAM technology, and determining a plurality of environment objects within a specified range of the automobile; The perception analysis module is used for acquiring the driving data of the automobile and marking the driving data in the driving map to synchronously position the automobile, and determining a first collision probability between the automobile and each environmental object; The auxiliary obstacle avoidance module is used for screening a plurality of driving obstacles in the driving map based on the first collision probability, establishing an obstacle avoidance plan according to the obstacle attribute corresponding to each driving obstacle and carrying out auxiliary obstacle avoidance; the obstacle avoidance training module is used for recording an obstacle avoidance process of a driver, training the obstacle avoidance process to determine driving preference of the driver, determining preference degree of the driver on different obstacle avoidance modes, generating an obstacle avoidance preference suggestion and displaying. In one embodiment of the present invention, in one possible implementation, Further comprises: the in-car obstacle avoidance module is used for acquiring live information in a specified range of the car when the car stops; determining a plurality of dynamic objects within a specified range of the automobile by utilizing the live information, and respectively establishing object activity characteristics within a stopping range of the automobile; Determining a second collision probability between each dynamic object and each door of the automobile based on the object activity characteristics; And when the second collision probability is larger than the safety probability, positioning the corresponding dangerous vehicle door and locking the dangerous vehicle door. In one embodiment of the present invention, in one possible implementation, Further comprises: the non-straight obstacle avoidance module is used for determining the advancing direction of the automobile according to the state of the steering lamp of the automobile; acquiring road condition information related to the advanc