Automobile environment perception obstacle avoidance system based on visual image technology

Abstract

The invention provides an automobile environment perception obstacle avoidance system based on a visual image technology. The system comprises the following steps: starting laser radar scanning when an automobile runs, constructing a running map of the automobile by combining an SLAM (Simultaneous Localization and Mapping) technology, and determining a plurality of environment articles within a specified range of the automobile, the method comprises the following steps: acquiring driving data of an automobile, marking the driving data in a driving map to synchronously position the automobile, determining a collision probability between the automobile and each environmental article, screening a plurality of driving obstacles in the driving map, establishing an obstacle avoidance plan according to an obstacle attribute corresponding to each driving obstacle, and performing auxiliary obstacle avoidance. The obstacle avoidance process of the driver is recorded, the driving preference of the driver is determined through training, the preference degree of the driver for different obstacle avoidance modes is determined, the obstacle avoidance preferable suggestion is generated and displayed, other automobiles, pedestrians and obstacles can be recognized and tracked in real time, and therefore the driver is helped to avoid potential collision.

Description

Technical Field

[0001] The present invention relates to the technical field of automobile assisted obstacle avoidance, and in particular to an automobile environment perception obstacle avoidance system based on visual imaging technology. Background Art

[0002] As one of the most common means of transportation in our daily lives, cars play an increasingly important role in our lives. With the technological progress in the fields of artificial intelligence, sensors, control, drive and materials, the production technology of cars has become more and more mature, making cars have changed from luxury goods to consumer goods in daily life. Due to the increase in the number of cars, many people will inevitably encounter car body collisions during driving due to insufficient driving experience. According to relevant information, when a car driving at high speed encounters an obstacle, the driver's reaction time is very short, and it is easy to cause a collision due to untimely avoidance, causing a traffic accident. The smart car equipped with auxiliary obstacle avoidance has the decision-making and execution ability to avoid dangerous factors, and can make timely obstacle avoidance reactions when danger comes to avoid car collisions. It is necessary to design an environmental perception obstacle avoidance system to solve these safety problems.

[0003] The current car obstacle avoidance system is only designed for the kinematic model of the car when it is driving at low speed. When the car is driving at high speed, the cyclic dynamic characteristics inside the car need to be considered. In order to accurately calculate the driving state of the car, the car obstacle avoidance system needs to be calculated through a complex physical dynamic model. The increased reaction time of the car obstacle avoidance increases the risk of car collision, and the safety performance is poor, resulting in more limited use.

[0004] Therefore, the present invention provides an automobile environment perception and obstacle avoidance system based on visual imaging technology. Summary of the invention

[0005] The present invention discloses an automobile environment perception obstacle avoidance system based on visual imaging technology, which can identify and track other automobiles, pedestrians and obstacles in real time, thereby helping the driver avoid potential collisions.

[0006] The present invention provides a vehicle environment perception and obstacle avoidance system based on visual imaging technology, comprising:

[0007] An environmental perception module is used to start laser radar scanning when the car is driving, build a driving map of the car in combination with SLAM technology, and determine a number of environmental objects within a specified range of the car;

[0008] A perception analysis module, used to obtain the driving data of the car and mark it in the driving map to synchronously locate the car, and determine the first collision probability between the car and each of the environmental objects;

[0009] an auxiliary obstacle avoidance module, configured to screen a plurality of driving obstacles in the driving map based on the first collision probability, establish an obstacle avoidance plan according to the obstacle attribute corresponding to each of the driving obstacles, and perform auxiliary obstacle avoidance;

[0010] The obstacle avoidance training module is used to record the driver's obstacle avoidance process, train the obstacle avoidance process to determine the driver's driving preference, determine the driver's preference for different obstacle avoidance methods, generate and display obstacle avoidance optimization suggestions.

[0011] In one practicable manner,

[0012] Also includes:

[0013] An in-vehicle obstacle avoidance module, used to obtain real-time information within a specified range of the vehicle when the vehicle stops;

[0014] Determine a plurality of dynamic objects within a specified range of the vehicle using the real-time information, and respectively establish activity characteristics of the objects within the stopped range of the vehicle;

[0015] Determine the second collision probability between each dynamic object and each door of the vehicle based on the object activity characteristics;

[0016] When the second collision probability is greater than the safety probability, the corresponding dangerous door is located and locked.

[0017] In one practicable manner,

[0018] Also includes:

[0019] A non-straight obstacle avoidance module, used to determine the forward direction of the vehicle according to the state of the turn signal of the vehicle;

[0020] Acquiring road condition information related to the forward direction;

[0021] The automobile assisted forward movement suggestion is generated and displayed according to the road condition information.

[0022] In one practicable manner,

[0023] The environment perception module comprises:

[0024] A synchronous scanning unit, used for performing synchronous scanning when the car is started, obtaining environmental information within a specified range of the car, and performing de-distortion processing on the environmental information to obtain a synchronous sensing scene within the specified range of the car;

[0025] A map construction unit, used to use SLAM technology to perform contour recognition on the synchronous sensing scene, obtain contours of several objects within a specified range, and build a driving map of the specified range in combination with historical sensing scenes;

[0026] The object recognition unit is used to respectively search for the appearance attributes corresponding to each of the object contours, perform in-depth recognition on the object contours using the appearance attributes, and determine a number of environmental objects within the specified range.

[0027] In one practicable manner,

[0028] The map construction unit comprises:

[0029] A contour recognition subunit is used to locate a number of key points in the synchronous sensing scene using SLAM technology, and to perform a radial diffusion search with each of the key points as the center, to obtain the intersection information between the synchronous sensing scene and each search line, and to establish a number of sparse contours of the synchronous sensing scene;

[0030] A contour improvement subunit is used to perform boundary expansion on each of the sparse contours to obtain a corresponding complete contour, obtain a smoothness corresponding to each of the complete contours, determine a filtering level of each of the complete contours based on the smoothness, and perform classification filtering on the complete contours to obtain contours of several objects in the synchronous sensing scene;

[0031] A feature exploration subunit is used to respectively identify each of the object contours in the historical sensing scene, and obtain the original features, repeated features and unique features corresponding to each of the object contours;

[0032] A map construction subunit is used to determine the fixed position of the corresponding object outline within the specified range according to the original features, determine the distribution position of the corresponding object outline within the specified range according to the repeated features, determine the active position of the corresponding object outline within the specified range according to the unique features, and generate a driving map of the specified range.

[0033] In one practicable manner,

[0034] The perception analysis module comprises:

[0035] A collection and processing unit, used for collecting the driving data of the vehicle, establishing the driving track of the vehicle according to the driving data, marking the driving track in the driving map to obtain the coverage overlap information between the vehicle and the ground surface;

[0036] A depth positioning unit, used to obtain a plurality of overlapping positions of the coverage overlapping information in the driving map, respectively obtain the road features corresponding to each of the overlapping positions, synchronously position the vehicle using the road features, and determine the real-time position of the vehicle;

[0037] The collision analysis unit is used to respectively obtain the activity speed characteristics and activity direction characteristics corresponding to each of the environmental objects, and determine the first collision probability between the environmental objects and each of the environmental objects based on the forward direction of the car.

[0038] In one practicable manner,

[0039] The collision analysis unit comprises:

[0040] A feature analysis subunit is used to locate each of the environmental objects in the driving map, collect the position change information of the environmental objects at different times, and construct the activity speed feature and activity direction feature corresponding to each of the environmental objects;

[0041] A spatial analysis subunit is used to establish a vehicle space vector according to the forward direction and forward speed of the vehicle, select target environmental objects corresponding to the target activity direction characteristics intersecting with the vehicle space vector, and establish an object space vector corresponding to each of the environmental objects using the target activity speed characteristics and the target activity direction characteristics corresponding to the target environmental objects;

[0042] The collision analysis subunit is used to respectively obtain the overlap duration between each of the object space vectors and the car space vector, construct a first collision probability between the car and the target environmental object according to the overlap duration, and regard the first collision probability of the environmental object that does not match the collision probability as 0.

[0043] In one practicable manner,

[0044] The auxiliary obstacle avoidance module comprises:

[0045] an obstacle avoidance preprocessing unit, configured to respectively locate each driving obstacle in the driving map, and sort the driving obstacles in descending order according to the first collision probability, so as to generate an obstacle avoidance queue for the automobile;

[0046] an obstacle avoidance pre-analysis unit, configured to respectively obtain obstacle attributes corresponding to each obstacle, determine a remaining obstacle avoidance distance between the vehicle and each obstacle according to the obstacle attributes, and determine a success probability of the vehicle avoiding each obstacle in a current lane in combination with the obstacle avoidance queue;

[0047] The auxiliary obstacle avoidance unit is used to find a target driving obstacle with a success probability lower than a standard probability, change lanes to avoid the target driving obstacle, and establish an obstacle avoidance plan for the car in combination with the avoidance measures corresponding to the current lane of the car and display it.

[0048] In one practicable manner,

[0049] The obstacle avoidance training module comprises:

[0050] An obstacle avoidance recording unit, used to record the driver's obstacle avoidance process and determine the driver's avoidance method for each of the driving obstacles;

[0051] A similar training unit is used to classify the avoidance methods according to obstacle attributes, generate a corresponding avoidance measure set, and select a number of habitual actions of the driver with a repetition frequency higher than a preset frequency from the avoidance measure set;

[0052] a preference analysis unit, for simulating each of the habitual actions respectively, obtaining a driving result corresponding to each of the habitual actions to generate a plurality of preferred obstacle avoidance modes, and determining a preference degree of each of the preferred obstacle avoidance modes according to a corresponding repetition frequency;

[0053] The obstacle avoidance optimization unit is used to match the target result of the obstacle avoidance with the driving result when performing obstacle avoidance, determine the corresponding preferred obstacle avoidance method according to the matching result, generate and display the obstacle avoidance optimization suggestion according to the preferred obstacle avoidance method.

[0054] In one practicable manner,

[0055] Also includes:

[0056] A road condition analysis module, used to determine the smoothness characteristics of the current road where the vehicle is located according to the driving map;

[0057] When the smoothing feature is abnormal, it is determined that the current road has a collapse phenomenon, and a warning message is generated.

[0058] The achievable beneficial effects of the above technical solution are as follows: in order to achieve synchronous obstacle avoidance and efficient obstacle avoidance, radar scanning is performed synchronously when the car is driving, and then SLAM technology is used to build a driving map. By identifying environmental objects in the driving map and combining the car's driving data to analyze the collision probability between the car and different environmental objects, auxiliary obstacle avoidance is performed for driving obstacles with a high collision probability. At the same time, the driver's obstacle avoidance process is recorded, and the obstacle avoidance method that the driver likes to use is determined through long-term training. When the same obstacle is encountered next time, the driver's preferred obstacle avoidance method is preferentially used for obstacle avoidance. In this way, not only can synchronous obstacle identification and obstacle avoidance be achieved, but also the driver's proficiency in obstacle avoidance methods can be enhanced through continuous training, so that the obstacle avoidance process is continuously optimized, the obstacle avoidance time is reduced, and the obstacle avoidance efficiency is improved.

[0059] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0060] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0062] Figure 1 A schematic diagram of the composition of a vehicle environment perception and obstacle avoidance system based on visual imaging technology in an embodiment of the present invention;

[0063] Figure 2 The figure is a schematic diagram of the composition of an environment perception module of a vehicle environment perception and obstacle avoidance system based on visual imaging technology in an embodiment of the present invention. DETAILED DESCRIPTION

[0064] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0065] Example 1

[0066] This embodiment provides a vehicle environment perception and obstacle avoidance system based on visual imaging technology. Figure 1 As shown, including:

[0067] An environmental perception module is used to start laser radar scanning when the car is driving, build a driving map of the car in combination with SLAM technology, and determine a number of environmental objects within a specified range of the car;

[0068] A perception analysis module, used to obtain the driving data of the car and mark it in the driving map to synchronously locate the car, and determine the first collision probability between the car and each of the environmental objects;

[0069] an auxiliary obstacle avoidance module, configured to screen a plurality of driving obstacles in the driving map based on the first collision probability, establish an obstacle avoidance plan according to the obstacle attribute corresponding to each of the driving obstacles, and perform auxiliary obstacle avoidance;

[0070] The obstacle avoidance training module is used to record the driver's obstacle avoidance process, train the obstacle avoidance process to determine the driver's driving preference, determine the driver's preference for different obstacle avoidance methods, generate and display obstacle avoidance optimization suggestions.

[0071] In this example, the laser radar scanning is composed of a depth and wide-angle camera, a detection radar, a door lock system, an alarm system, and a central processing unit. A depth and wide-angle camera and a detection radar are respectively installed on the left and right rearview mirrors of the car. The glass of the rearview mirror of the car is a one-way perspective glass. The lens of the depth and wide-angle camera is hidden behind the glass of the rearview mirror. The detection radar is installed at the end of the rearview mirror away from the car body, and can make pitching motion and horizontal swing relative to the rearview mirror. The depth and wide-angle camera and the detection radar can exchange data with the central processing unit; a door lock system is installed on each door of the car, and the door lock system is an external door lock or a door lock that comes with the car, and the door lock system is controlled by the central processing unit; the alarm system is an external alarm system or a built-in alarm system of the car, and the alarm system is controlled by the central processing unit; the manual intervention device is used to intervene in the opening and closing of the entire vision-based intelligent driving assistance system or only to intervene in the opening and closing of the door lock system;

[0072] In this example, environmental items refer to items that exist in the environment where the car is located;

[0073] In this example, SLAM technology refers to a technology that simultaneously performs self-localization and environment mapping in an unknown environment;

[0074] In this example, the driving map refers to a map used to present the route of the car during driving;

[0075] In this example, the first collision probability represents the collision probability between the car and the environmental object;

[0076] In this example, when the first collision probability is greater than 75%, the corresponding environmental object is regarded as a driving obstacle;

[0077] In this example, the obstacle attributes represent the appearance, specifications, and materials of the obstacle;

[0078] In this example, the role of assisted obstacle avoidance is to: provide the driver with an obstacle avoidance plan, improve the driver's obstacle avoidance efficiency, and reduce thinking time;

[0079] In this example, the specified range means a range that is centered on the car and expands the volume of the car by 2 times;

[0080] In this example, the specific workflow of this system is as follows:

[0081] (1) Collect real vehicle data, and simulate the environmental perception and chassis wire control system installation and adjustment process through text, voice, and perspective positioning prompts. In view of the shortcomings of traditional ultrasonic obstacle avoidance systems that are easily affected by driving environments such as road conditions and temperature, have low detection accuracy and high cost, laser sensors are used to achieve high-precision detection between vehicles and obstacles at medium and long distances;

[0082] (2) Analyze real vehicle data, simulate abnormal CAN line waveforms, unresponsive host computer line control, and other phenomena, and provide real-time feedback on operation results. Design a matrix filter to process the detection data of the laser sensor to improve the reliability of the detection data;

[0083] (3) Combined with the detection data of laser and wheel speed sensors, the vehicle's motion state is systematically analyzed and controlled through a neural network algorithm to achieve emergency braking of the vehicle;

[0084] (4) According to the system function requirements, complete the design of hardware parts such as laser emission system and SLAM audio-visual system;

[0085] (5) Complete the software design of the intelligent connected car environment perception system and conduct system simulation virtual debugging;

[0086] (6) Realize the software and hardware coordination of the intelligent driving assistance obstacle avoidance system to improve system performance.

[0087] The working principle and beneficial effects of the above technical solution are as follows: In order to achieve synchronous obstacle avoidance and efficient obstacle avoidance, radar scanning is performed synchronously when the car is driving, and then SLAM technology is used to build a driving map. By identifying environmental objects in the driving map and combining the car's driving data to analyze the collision probability between the car and different environmental objects, auxiliary obstacle avoidance is performed for driving obstacles with a high collision probability. At the same time, the driver's obstacle avoidance process is recorded, and the obstacle avoidance method that the driver likes to use is determined through long-term training. When the same obstacle is encountered next time, the driver's preferred obstacle avoidance method is preferentially used for obstacle avoidance. In this way, not only can synchronous obstacle identification and obstacle avoidance be achieved, but also the driver's proficiency in obstacle avoidance methods can be enhanced through continuous training, so that the obstacle avoidance process is continuously optimized, the obstacle avoidance time is reduced, and the obstacle avoidance efficiency is improved.

[0088] Example 2

[0089] On the basis of Embodiment 1, the vehicle environment perception and obstacle avoidance system based on visual imaging technology further includes:

[0090] An in-vehicle obstacle avoidance module, used to obtain real-time information within a specified range of the vehicle when the vehicle stops;

[0091] Determine a plurality of dynamic objects within a specified range of the vehicle using the real-time information, and respectively establish activity characteristics of the objects within the stopped range of the vehicle;

[0092] Determine the second collision probability between each dynamic object and each door of the vehicle based on the object activity characteristics;

[0093] When the second collision probability is greater than the safety probability, the corresponding dangerous door is located and locked.

[0094] In this example, the second collision probability represents the collision probability between the dynamic object and the car door;

[0095] In this example, the safety probability is 45%.

[0096] The working principle and beneficial effects of the above technical solution are as follows: when the car stops and the passenger wants to open the door, the system detects in advance whether there is a vehicle behind the car. If so, it detects its distance, movement trend and other information. The door lock system can prevent the passenger from opening the door rashly. In addition, deep processing can also detect in advance whether there is any danger and issue an alarm when the driver is about to change lanes, turn, pull over, etc., with a high degree of intelligence.

[0097] Example 3

[0098] Based on Example 2, the vehicle environment perception and obstacle avoidance system based on visual imaging technology further includes:

[0099] A non-straight obstacle avoidance module, used to determine the forward direction of the vehicle according to the state of the turn signal of the vehicle;

[0100] Acquiring road condition information related to the forward direction;

[0101] The automobile assisted forward movement suggestion is generated and displayed according to the road condition information.

[0102] The working principle and beneficial effects of the above technical solution are as follows: when the car stops and the passenger wants to open the door, the system detects in advance whether there is a vehicle behind the car. If so, it detects its distance, movement trend and other information. The door lock system can prevent the passenger from opening the door rashly. In addition, deep processing can also detect in advance whether there is any danger and issue an alarm when the driver is about to change lanes, turn, pull over, etc., with a high degree of intelligence.

[0103] Example 4

[0104] Based on Example 1, the vehicle environment perception and obstacle avoidance system based on visual imaging technology is as follows: Figure 2 As shown, the environment perception module includes:

[0105] A synchronous scanning unit, used for performing synchronous scanning when the car is started, obtaining environmental information within a specified range of the car, and performing de-distortion processing on the environmental information to obtain a synchronous sensing scene within the specified range of the car;

[0106] A map construction unit, used to use SLAM technology to perform contour recognition on the synchronous sensing scene, obtain contours of several objects within a specified range, and build a driving map of the specified range in combination with historical sensing scenes;

[0107] The object recognition unit is used to respectively search for the appearance attributes corresponding to each of the object contours, perform in-depth recognition on the object contours using the appearance attributes, and determine a number of environmental objects within the specified range.

[0108] In this example, the de-distortion processing refers to the process of correcting the image distortion caused by factors such as lens distortion and shooting angle in the environmental information;

[0109] In this example, the synchronous sensing scene means constructing a virtual scene that is consistent with the actual driving state of the car;

[0110] In this example, the historical sensing scene refers to a number of synchronous sensing scenes that the car has passed through during this driving process. When a new synchronous sensing scene appears, the previous synchronous sensing scene is regarded as a historical sensing scene;

[0111] In this example, the appearance attribute represents the appearance presented by the object outline;

[0112] In this example, deep recognition refers to the process of inputting appearance attributes and object contours into big data for recognition and determining the name of the object contour.

[0113] The working principle and beneficial effects of the above technical solution are as follows: when the car starts to drive, a synchronous scan is performed, and then the deformation in the environmental information is eliminated through de-distortion processing to ensure that the generated synchronous sensing scene is consistent with the actual scene of the environment where the car is located, and the accuracy of obstacle identification and obstacle avoidance is guaranteed. Then, SLAM technology is used for contour recognition, and the driving map is built in combination with historical sensing scenes. Finally, the contours of the objects are deeply identified and analyzed to determine the name of each object. Finally, the environmental objects within the specified range are obtained. In this way, the driving process of the car can be fully presented, which is convenient for fault identification.

[0114] Example 5

[0115] On the basis of Embodiment 4, the vehicle environment perception and obstacle avoidance system based on visual imaging technology, the map construction unit includes:

[0116] A contour recognition subunit is used to locate a number of key points in the synchronous sensing scene using SLAM technology, and to perform a radial diffusion search with each of the key points as the center, to obtain the intersection information between the synchronous sensing scene and each search line, and to establish a number of sparse contours of the synchronous sensing scene;

[0117] A contour improvement subunit is used to perform boundary expansion on each of the sparse contours to obtain a corresponding complete contour, obtain a smoothness corresponding to each of the complete contours, determine a filtering level of each of the complete contours based on the smoothness, and perform classification filtering on the complete contours to obtain contours of several objects in the synchronous sensing scene;

[0118] A feature exploration subunit is used to respectively identify each of the object contours in the historical sensing scene, and obtain the original features, repeated features and unique features corresponding to each of the object contours;

[0119] A map construction subunit is used to determine the fixed position of the corresponding object outline within the specified range according to the original features, determine the distribution position of the corresponding object outline within the specified range according to the repeated features, determine the active position of the corresponding object outline within the specified range according to the unique features, and generate a driving map of the specified range.

[0120] In this example, the key points represent a number of points randomly selected in the synchronous sensing scene and having the characteristics of being arranged at equal distances;

[0121] In this example, the radial diffusion search means a search process that takes a key point as the center and the total number of key points as the radial line, radiating outward from the key point. During the search, when the radial line intersects with the contour in the synchronous sensing scene, it is recorded as an intersection point, and the intersection information is updated;

[0122] In this example, the sparse contour represents the contour generated by the radial diffusion search;

[0123] In this instance, boundary dilation refers to the process of expanding the boundary of the sparse image by comparing each pixel in the sparse contour with other pixels in its neighborhood;

[0124] In this example, the higher the smoothness, the lower the filtering level;

[0125] In this example, the original feature represents the object contour that has existed in the historical synchronous sensing scene, the repeated feature represents the object contour that appears repeatedly in the historical synchronous sensing scene, and the unique feature represents the object contour that appears for the first time in the synchronous sensing scene.

[0126] The working principle and beneficial effects of the above technical solution: In order to further improve the driving map and reduce the error of object recognition, SLAM technology is first used to locate several key points in the synchronous sensing scene, and then a radial diffusion search is performed with these points as the starting point, so as to use the cross-information obtained from the search to establish a sparse contour, and the sparse contour is further expanded and filtered to obtain the object contour in the synchronous sensing scene, and then the characteristics of each object contour are identified, and then it is positioned accordingly, and finally a driving map is generated. By processing the object contour multiple times, the accuracy of the generated driving map can be guaranteed, and synchronous positioning can be better achieved.

[0127] Example 6

[0128] On the basis of Example 1, the automobile environment perception and obstacle avoidance system based on visual imaging technology, the perception analysis module includes:

[0129] A collection and processing unit, used for collecting the driving data of the vehicle, establishing the driving track of the vehicle according to the driving data, marking the driving track in the driving map to obtain the coverage overlap information between the vehicle and the ground surface;

[0130] A depth positioning unit, used to obtain a plurality of overlapping positions of the coverage overlapping information in the driving map, respectively obtain the road features corresponding to each of the overlapping positions, synchronously position the vehicle using the road features, and determine the real-time position of the vehicle;

[0131] The collision analysis unit is used to respectively obtain the activity speed characteristics and activity direction characteristics corresponding to each of the environmental objects, and determine the first collision probability between the environmental objects and each of the environmental objects based on the forward direction of the car.

[0132] In this example, the coverage coincidence information indicates the coincidence of the car and the ground surface;

[0133] In this example, the activity speed feature indicates the speed of the environmental object, and the activity direction feature indicates the direction of the environmental object. If the environmental object is fixed, the activity speed feature is 0, and the activity direction feature is vertical to the ground.

[0134] The working principle and beneficial effects of the above technical solution are as follows: the driving trajectory of the car is determined by processing the driving data of the car, and then the overlap information between the car and the ground is combined for synchronous positioning. Finally, the activity speed characteristics and activity direction characteristics of environmental objects are analyzed to determine the collision probability between the car and different objects. The driving process of the car can be monitored in real time. It is not only suitable for private cars, but also for commercial vehicles, public transportation and self-driving cars, providing drivers with a safer and smarter driving experience, which not only reduces the risk of traffic accidents, but also helps to improve road traffic efficiency.

[0135] Example 7

[0136] On the basis of Example 6, the vehicle environment perception and obstacle avoidance system based on visual imaging technology, the collision analysis unit includes:

[0137] A feature analysis subunit is used to locate each of the environmental objects in the driving map, collect the position change information of the environmental objects at different times, and construct the activity speed feature and activity direction feature corresponding to each of the environmental objects;

[0138] A spatial analysis subunit is used to establish a vehicle space vector according to the forward direction and forward speed of the vehicle, select target environmental objects corresponding to the target activity direction characteristics intersecting with the vehicle space vector, and establish an object space vector corresponding to each of the environmental objects using the target activity speed characteristics and the target activity direction characteristics corresponding to the target environmental objects;

[0139] The collision analysis subunit is used to respectively obtain the overlap duration between each of the object space vectors and the car space vector, construct a first collision probability between the car and the target environmental object according to the overlap duration, and regard the first collision probability of the environmental object that does not match the collision probability as 0.

[0140] In this example, the car space vector represents a vector representing the driving condition of the car in three-dimensional space, and the object space vector represents a vector representing the activity condition of the object in three-dimensional space.

[0141] The working principle and beneficial effects of the above technical solution are as follows: after the environmental objects included in the driving map are determined, the activity speed characteristics and activity direction characteristics of each environmental object are analyzed to establish the corresponding space vector, and the collision probability between the car and the environmental object is determined by analyzing the collision situation between the space vectors, thus completing the obstacle recognition and laying the foundation for subsequent obstacle avoidance.

[0142] Example 8

[0143] On the basis of Example 1, the vehicle environment perception and obstacle avoidance system based on visual imaging technology, the auxiliary obstacle avoidance module includes:

[0144] an obstacle avoidance preprocessing unit, configured to respectively locate each driving obstacle in the driving map, and sort the driving obstacles in descending order according to the first collision probability, so as to generate an obstacle avoidance queue for the automobile;

[0145] an obstacle avoidance pre-analysis unit, configured to respectively obtain obstacle attributes corresponding to each obstacle, determine a remaining obstacle avoidance distance between the vehicle and each obstacle according to the obstacle attributes, and determine a success probability of the vehicle avoiding each obstacle in a current lane in combination with the obstacle avoidance queue;

[0146] The auxiliary obstacle avoidance unit is used to find a target driving obstacle with a success probability lower than a standard probability, change lanes to avoid the target driving obstacle, and establish an obstacle avoidance plan for the car in combination with the avoidance measures corresponding to the current lane of the car and display it.

[0147] In this example, the standard probability is 95%.

[0148] The working principle and beneficial effects of the above technical solution are as follows: when avoiding obstacles, the current lane is given priority, which can not only achieve obstacle avoidance, but also reduce danger, and avoid the driver being distracted by performing multiple operations at the same time. In order to improve the driver's obstacle avoidance quality and avoid interfering with the driver's judgment of the actual situation, only auxiliary obstacle avoidance is used to cooperate with the driver during obstacle avoidance, thereby improving the success rate of obstacle avoidance.

[0149] Example 9

[0150] On the basis of Example 1, the vehicle environment perception and obstacle avoidance system based on visual imaging technology, the obstacle avoidance training module includes:

[0151] An obstacle avoidance recording unit, used to record the driver's obstacle avoidance process and determine the driver's avoidance method for each of the driving obstacles;

[0152] A similar training unit is used to classify the avoidance methods according to obstacle attributes, generate a corresponding avoidance measure set, and select a number of habitual actions of the driver with a repetition frequency higher than a preset frequency from the avoidance measure set;

[0153] a preference analysis unit, for simulating each of the habitual actions respectively, obtaining a driving result corresponding to each of the habitual actions to generate a plurality of preferred obstacle avoidance modes, and determining a preference degree of each of the preferred obstacle avoidance modes according to a corresponding repetition frequency;

[0154] The obstacle avoidance optimization unit is used to match the target result of the obstacle avoidance with the driving result when performing obstacle avoidance, determine the corresponding preferred obstacle avoidance method according to the matching result, generate and display the obstacle avoidance optimization suggestion according to the preferred obstacle avoidance method.

[0155] In this example, the preset frequency is 30%.

[0156] The working principle and beneficial effects of the above technical solution are as follows: when the driver is avoiding obstacles, his obstacle avoidance actions are recorded, and then the driver's habitual actions are screened. When the driver encounters obstacles that have occurred before, the driver's preferred obstacle avoidance method is prioritized for obstacle avoidance. This allows the driver to adapt to the vehicle, and can also improve the speed and efficiency of obstacle avoidance through multiple trainings, thereby reducing the probability of danger.

[0157] Example 10

[0158] On the basis of Embodiment 1, the vehicle environment perception and obstacle avoidance system based on visual imaging technology further includes:

[0159] A road condition analysis module, used to determine the smoothness characteristics of the current road where the vehicle is located according to the driving map;

[0160] When the smoothing feature is abnormal, it is determined that the current road has a collapse phenomenon and a warning message is generated.

[0161] The working principle and beneficial effects of the above technical solution are: analyzing the road conditions while the driver is driving, further reducing the probability of danger, and ensuring the safety of the driver and passengers.

[0162] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A vehicle environment perception and obstacle avoidance system based on visual imaging technology, characterized in that: include: An environmental perception module is used to start laser radar scanning when the car is driving, build a driving map of the car in combination with SLAM technology, and determine a number of environmental objects within a specified range of the car; A perception analysis module, used to obtain the driving data of the car and mark it in the driving map to synchronously locate the car, and determine the first collision probability between the car and each of the environmental objects; an auxiliary obstacle avoidance module, configured to screen a plurality of driving obstacles in the driving map based on the first collision probability, establish an obstacle avoidance plan according to the obstacle attribute corresponding to each of the driving obstacles, and perform auxiliary obstacle avoidance; The obstacle avoidance training module is used to record the driver's obstacle avoidance process, train the obstacle avoidance process to determine the driver's driving preference, determine the driver's preference for different obstacle avoidance methods, generate and display obstacle avoidance optimization suggestions.

2. The vehicle environment perception and obstacle avoidance system based on visual imaging technology as claimed in claim 1, characterized in that: Also includes: An in-vehicle obstacle avoidance module, used to obtain real-time information within a specified range of the vehicle when the vehicle stops; Determine a plurality of dynamic objects within a specified range of the vehicle using the real-time information, and respectively establish activity characteristics of the objects within the stopped range of the vehicle; Determine the second collision probability between each dynamic object and each door of the vehicle based on the object activity characteristics; When the second collision probability is greater than the safety probability, the corresponding dangerous door is located and locked.

3. The vehicle environment perception and obstacle avoidance system based on visual imaging technology as claimed in claim 2, characterized in that: Also includes: A non-straight obstacle avoidance module, used to determine the forward direction of the vehicle according to the state of the turn signal of the vehicle; Acquiring road condition information related to the forward direction; The automobile assisted forward movement suggestion is generated and displayed according to the road condition information.

4. The vehicle environment perception and obstacle avoidance system based on visual imaging technology as claimed in claim 1, characterized in that: The environment perception module comprises: A synchronous scanning unit, used for performing synchronous scanning when the car is started, obtaining environmental information within a specified range of the car, and performing de-distortion processing on the environmental information to obtain a synchronous sensing scene within the specified range of the car; A map construction unit, used to use SLAM technology to perform contour recognition on the synchronous sensing scene, obtain contours of several objects within a specified range, and build a driving map of the specified range in combination with historical sensing scenes; The object recognition unit is used to respectively search for the appearance attributes corresponding to each of the object contours, perform in-depth recognition on the object contours using the appearance attributes, and determine a number of environmental objects within the specified range.

5. The vehicle environment perception and obstacle avoidance system based on visual imaging technology as claimed in claim 4, characterized in that: The map construction unit comprises: A contour recognition subunit is used to locate a number of key points in the synchronous sensing scene using SLAM technology, and to perform a radial diffusion search with each of the key points as the center, to obtain the intersection information between the synchronous sensing scene and each search line, and to establish a number of sparse contours of the synchronous sensing scene; A contour improvement subunit is used to perform boundary expansion on each of the sparse contours to obtain a corresponding complete contour, obtain a smoothness corresponding to each of the complete contours, determine a filtering level of each of the complete contours based on the smoothness, and perform classification filtering on the complete contours to obtain contours of several objects in the synchronous sensing scene; A feature exploration subunit is used to respectively identify each of the object contours in the historical sensing scene, and obtain the original features, repeated features and unique features corresponding to each of the object contours; A map construction subunit is used to determine the fixed position of the corresponding object outline within the specified range according to the original features, determine the distribution position of the corresponding object outline within the specified range according to the repeated features, determine the active position of the corresponding object outline within the specified range according to the unique features, and generate a driving map of the specified range.

6. The vehicle environment perception and obstacle avoidance system based on visual imaging technology as claimed in claim 1, characterized in that: The perception analysis module comprises: A collection and processing unit, used for collecting the driving data of the vehicle, establishing the driving track of the vehicle according to the driving data, marking the driving track in the driving map to obtain the coverage overlap information between the vehicle and the ground surface; A depth positioning unit, used to obtain a plurality of overlapping positions of the coverage overlapping information in the driving map, respectively obtain the road features corresponding to each of the overlapping positions, synchronously position the vehicle using the road features, and determine the real-time position of the vehicle; The collision analysis unit is used to respectively obtain the activity speed characteristics and activity direction characteristics corresponding to each of the environmental objects, and determine the first collision probability between the environmental objects and each of the environmental objects based on the forward direction of the car.

7. The vehicle environment perception and obstacle avoidance system based on visual imaging technology as claimed in claim 6, characterized in that: The collision analysis unit comprises: A feature analysis subunit is used to locate each of the environmental objects in the driving map, collect the position change information of the environmental objects at different times, and construct the activity speed feature and activity direction feature corresponding to each of the environmental objects; A spatial analysis subunit is used to establish a vehicle space vector according to the forward direction and forward speed of the vehicle, select target environmental objects corresponding to the target activity direction characteristics intersecting with the vehicle space vector, and establish an object space vector corresponding to each of the environmental objects using the target activity speed characteristics and the target activity direction characteristics corresponding to the target environmental objects; The collision analysis subunit is used to respectively obtain the overlap duration between each of the object space vectors and the car space vector, construct a first collision probability between the car and the target environmental object according to the overlap duration, and regard the first collision probability of the environmental object that does not match the collision probability as 0.

8. The vehicle environment perception and obstacle avoidance system based on visual imaging technology as claimed in claim 1, characterized in that: The auxiliary obstacle avoidance module comprises: an obstacle avoidance preprocessing unit, configured to respectively locate each driving obstacle in the driving map, and sort the driving obstacles in descending order according to the first collision probability, so as to generate an obstacle avoidance queue for the automobile; an obstacle avoidance pre-analysis unit, configured to respectively obtain obstacle attributes corresponding to each obstacle, determine a remaining obstacle avoidance distance between the vehicle and each obstacle according to the obstacle attributes, and determine a success probability of the vehicle avoiding each obstacle in a current lane in combination with the obstacle avoidance queue; The auxiliary obstacle avoidance unit is used to find a target driving obstacle with a success probability lower than a standard probability, change lanes to avoid the target driving obstacle, and establish an obstacle avoidance plan for the car in combination with the avoidance measures corresponding to the current lane of the car and display it.

9. The vehicle environment perception and obstacle avoidance system based on visual imaging technology as claimed in claim 1, characterized in that: The obstacle avoidance training module comprises: An obstacle avoidance recording unit, used to record the driver's obstacle avoidance process and determine the driver's avoidance method for each of the driving obstacles; A similar training unit is used to classify the avoidance methods according to obstacle attributes, generate a corresponding avoidance measure set, and select a number of habitual actions of the driver with a repetition frequency higher than a preset frequency from the avoidance measure set; a preference analysis unit, for simulating each of the habitual actions respectively, obtaining a driving result corresponding to each of the habitual actions to generate a plurality of preferred obstacle avoidance modes, and determining a preference degree of each of the preferred obstacle avoidance modes according to a corresponding repetition frequency; The obstacle avoidance optimization unit is used to match the target result of the obstacle avoidance with the driving result when performing obstacle avoidance, determine the corresponding preferred obstacle avoidance method according to the matching result, generate and display the obstacle avoidance optimization suggestion according to the preferred obstacle avoidance method.

10. The vehicle environment perception and obstacle avoidance system based on visual imaging technology as claimed in claim 1, characterized in that: Also includes: A road condition analysis module, used to determine the smoothness characteristics of the current road where the vehicle is located according to the driving map; When the smoothing feature is abnormal, it is determined that the current road has a collapse phenomenon, and a warning message is generated.

Images