Vehicle anti-collision early warning method based on machine vision

Abstract

The invention provides a vehicle anti-collision early warning method based on machine vision. The vehicle anti-collision early warning method comprises the following steps that S1, image data of a front road are collected through a camera in an image collection module; s2, preprocessing the acquired image data through an image processing module; s3, carrying out obstacle recognition on the preprocessed image data based on an obstacle recognition module, and determining the position and speed of an obstacle; s4, calculating a collision risk index through a risk assessment module; and S5, when the collision risk index exceeds a preset threshold value, an early warning module sends out an anti-collision early warning signal. According to the vehicle anti-collision early warning method based on machine vision, real-time detection of obstacles in front of the vehicle and dynamic evaluation of collision risks are achieved, risk collision indexes are evaluated through the risk evaluation module, and the safety distance threshold value is dynamically adjusted according to the current speed of the vehicle and road conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle anti-collision, and in particular to a vehicle anti-collision warning method based on machine vision. Background Art

[0002] With the continuous increase in the number of cars, the incidence of road traffic accidents has also increased accordingly, posing a huge threat to people's lives and property safety. In order to improve driving safety and reduce the occurrence of traffic accidents, the vehicle collision warning system (CAS) came into being. This type of system monitors the vehicle's surroundings in real time, predicts potential collision risks, and promptly issues warnings to the driver, and even automatically takes braking measures in some cases to avoid or mitigate collisions.

[0003] Traditional anti-collision warning systems mainly rely on a single type of sensor, such as radar or camera, for obstacle detection. These systems have limitations in terms of obstacle recognition accuracy, response speed, and completeness of information transmission. For example, a single sensor system may have difficulty accurately identifying obstacles in complex environments, especially in bad weather or low light conditions. In addition, traditional systems are usually based on fixed rules or algorithms and lack flexibility and adaptability.

[0004] As an important branch of artificial intelligence, machine vision technology has made significant progress in recent years. It converts the target into an image signal through an image acquisition device and uses an image processing system for analysis. In the field of vehicle collision warning, machine vision technology can be used to monitor the road conditions in front of the vehicle in real time and identify potential obstacles. However, traditional machine vision systems have challenges in feature extraction and target recognition, especially in complex environments and dynamic scenes. Summary of the invention

[0005] The purpose of the present invention is to provide a vehicle anti-collision warning method based on machine vision, which realizes real-time detection of obstacles in front of the vehicle and dynamic assessment of collision risk, evaluates the risk collision index through a risk assessment module, and dynamically adjusts the safety distance threshold according to the current speed of the vehicle and road conditions.

[0006] To achieve the above object, the present invention provides a vehicle anti-collision warning method based on machine vision, comprising the following steps:

[0007] Step S1, collecting road image data through a camera in an image acquisition module;

[0008] Step S2: preprocessing the collected image data through an image processing module;

[0009] Step S3: performing obstacle recognition on the preprocessed image data based on the obstacle recognition module, and determining the position and speed of the obstacle;

[0010] Step S4: According to the position and speed of the obstacle and the current driving state of the vehicle, the collision risk index R is calculated by the risk assessment module:

[0011]

[0012] Among them, V rel is the relative speed between the vehicle and the obstacle; d is the distance between the vehicle and the obstacle; d0 is the safety distance threshold; λ is the adjustment parameter;

[0013] Step S5: Compare the collision risk index with a preset threshold.

[0014] Preferably, in step S1, the cameras in the image acquisition module are divided into a front camera, a side camera and a rear camera. The cameras use high-resolution cameras. The front camera is installed above the vehicle windshield, the side cameras are installed at the doors on both sides of the vehicle, and the rear camera is installed at the rear of the vehicle. The high-speed data interface in the image acquisition module transmits the image data to the image processing module in real time.

[0015] Preferably, in step S2, the preprocessing in the image processing module includes graying, edge detection and feature extraction.

[0016] Preferably, in step S3, the obstacle recognition module uses a YOLOv8 learning model to perform obstacle recognition, obstacle position determination and obstacle speed determination on the preprocessed image data.

[0017] Preferably, in step S4, the safety distance threshold d0 is dynamically adjusted according to the current speed of the vehicle and the road conditions, and the specific formula is:

[0018] d0=k×v 2 ×μ;

[0019] Among them, k is the proportional coefficient; v is the vehicle speed; μ is the road friction coefficient.

[0020] Preferably, in step S5, the warning module compares the collision risk index R in step S4 with a preset threshold value to determine whether it is necessary to issue a warning signal.

[0021] Preferably, the warning signal includes visual warning, auditory warning and tactile warning. The visual warning displays the location and distance information of the obstacle through the warning light on the vehicle dashboard or the display screen; the auditory warning reminds the driver through a sound alarm; and the tactile warning reminds the driver through seat vibration.

[0022] Therefore, the present invention adopts the above-mentioned vehicle anti-collision warning method based on machine vision, realizes real-time detection of obstacles in front of the vehicle and dynamic assessment of collision risk, evaluates the risk collision index through the risk assessment module, and dynamically adjusts the safety distance threshold according to the current speed of the vehicle and road conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 The present invention is a flowchart of the overall system of a method for vehicle anti-collision warning based on machine vision. DETAILED DESCRIPTION

[0024] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0025] Unless otherwise defined, technical or scientific terms used in the present invention shall have the common meanings understood by one having ordinary skills in the field to which the present invention belongs.

[0026] Embodiment 1

[0027] like Figure 1 As shown, the present invention provides a vehicle anti-collision warning method based on machine vision, comprising the following steps:

[0028] Step S1, collecting road image data through the camera in the image acquisition module; in step S1, the camera in the image acquisition module is divided into a front camera, a side camera and a rear camera, and the camera adopts a high-resolution camera. The high-resolution camera (such as 1920×1080 or higher) can provide clearer and more detailed image data, which helps the system to more accurately identify the shape, size and position of obstacles. The front camera is installed above the windshield of the vehicle to monitor the road and obstacles ahead. Including pedestrians, vehicles, traffic signs, etc. The front camera is the core part of the anti-collision warning system, which can capture the dynamic situation ahead in real time and provide the system with detailed information on the environment ahead. The side cameras are installed at the door positions on both sides of the vehicle to monitor the side traffic situation. When the vehicle changes lanes or turns, it can provide a side view to help the system detect side obstacles, such as other vehicles, pedestrians or obstacles. The rear camera is installed at the rear of the vehicle to monitor the rear traffic situation. When reversing or reversing into the garage, it can provide a rear view to help the system detect rear obstacles, such as other vehicles, pedestrians or obstacles. The high-speed data interface in the image acquisition module transmits the image data to the image processing module in real time. The high-speed data interface uses USB3.0. The USB3.0 interface has the characteristics of high bandwidth and low latency, which can ensure the fast and stable transmission of image data. During the transmission process, the image data is usually compressed to reduce bandwidth usage while maintaining image quality. The high bandwidth characteristics of the USB3.0 interface can support the fast transmission of high-resolution images and ensure the real-time performance of the system.

[0029] The image acquisition module collects image data of the vehicle's surrounding environment in real time through multiple high-resolution cameras, and transmits the data to the image processing module through USB3.0. The reasonable layout and high-resolution characteristics of the cameras, combined with the efficient transmission capability of the high-speed data interface, provide high-quality input for subsequent image processing and obstacle recognition, ensuring the accuracy and reliability of the entire anti-collision warning system.

[0030] Step S2: preprocessing the collected image data through an image processing module;

[0031] In step S2, preprocessing in the image processing module includes graying, edge detection and feature extraction.

[0032] Data volume reduction: Color images contain three channels, RGB, and each channel contains a large amount of data. By graying, the color image is converted into a single-channel grayscale image, which significantly reduces the amount of data and reduces the computational complexity; retaining structural information: Graying retains the brightness information of the image, that is, the light and dark contrast and shape information of the image, which is crucial for subsequent edge detection and feature extraction; enhancing processing efficiency: The processing speed of grayscale images is faster, because the amount of calculation of single-channel images is much smaller than that of three-channel images, thereby improving the real-time performance of the entire system; the role of graying is to convert color images into grayscale images, reduce the amount of data, and retain important structural information of the image for subsequent processing. The purpose of edge detection is to detect the edges of objects in the image and enhance the contour information of the image, which is convenient for subsequent feature extraction and target recognition. The purpose of feature extraction is to extract features that are helpful for obstacle recognition from the preprocessed image.

[0033] Step S3, based on the obstacle recognition module, the obstacle recognition is performed on the preprocessed image data, and the position and speed of the obstacle are determined; in step S3, the obstacle recognition module uses the YOLOv8 learning model to perform obstacle recognition, obstacle position determination and obstacle speed determination on the preprocessed image data. Obstacle recognition includes model loading, image preprocessing and obstacle detection. Model loading: load the pretrained YOLOv8 model and select the appropriate weight file to adapt to the obstacle detection task; image preprocessing: further process the preprocessed image to meet the input requirements of the YOLOv8 model; obstacle detection: use the YOLOv8 model to detect obstacles on the preprocessed image, and the model will output the bounding box, category and confidence of the detected obstacle.

[0034] The bounding box coordinates output by the YOLOv8 model are used to determine the location of the obstacle; the speed of the obstacle can be calculated by tracking consecutive frames.

[0035] Model loading: Load the pre-trained YOLOv8 model to ensure that the model can adapt to the obstacle detection task; Image preprocessing: Further process the image to meet the input requirements of the YOLOv8 model; Obstacle detection: Use the YOLOv8 model to detect obstacles in the image and output the obstacle's bounding box, category, and confidence; Obstacle position determination: Determine the position of the obstacle in the image through the bounding box coordinates; Obstacle speed determination: Calculate the speed of the obstacle through continuous frame tracking to provide more comprehensive information for collision risk assessment.

[0036] Step S4: According to the position and speed of the obstacle and the current driving state of the vehicle, the collision risk index R is calculated by the risk assessment module:

[0037]

[0038] Among them, V rel is the relative speed between the vehicle and the obstacle; d is the distance between the vehicle and the obstacle; d0 is the safety distance threshold; λ is the adjustment parameter;

[0039] In step S4, the safety distance threshold d0 is dynamically adjusted according to the current speed of the vehicle and the road conditions. The specific formula is:

[0040] d0=k×v 2 ×μ;

[0041] Where k is the proportionality coefficient; v is the vehicle speed; μ is the road friction coefficient;

[0042] Obtain the position and speed of the obstacle from the obstacle recognition module; calculate the relative speed V based on the speed of the vehicle and the obstacle rel , calculate the distance d according to the position of the vehicle and the obstacle, and calculate the safety distance threshold d0 according to the vehicle speed and road conditions. rel , d, d0 and are substituted to obtain the collision risk index R.

[0043] Step S5: Compare the collision risk index with a preset threshold.

[0044] When the collision risk index exceeds the preset threshold, the warning module sends out an anti-collision warning signal.

[0045] In step S5, the warning module compares the collision risk index R in step S4 with the preset threshold value to determine whether a warning signal needs to be issued. The warning signal includes visual warning, auditory warning and tactile warning. The visual warning displays the location and distance information of the obstacle through the warning light on the vehicle dashboard or the display screen; the auditory warning reminds the driver through a sound alarm; and the tactile warning reminds the driver through seat vibration.

[0046] Obstacle information acquisition: Obtain the position and speed information of the obstacle from the obstacle recognition module; Relative speed calculation: Calculate the relative speed between the vehicle and the obstacle; Distance calculation: Calculate the distance between the vehicle and the obstacle; Safety distance threshold calculation: Calculate the safety distance threshold based on the vehicle speed and road conditions; Collision risk index calculation: Substitute the relative speed, distance, safety distance threshold and adjustment parameters into the formula collision risk index R to obtain the collision risk index.

[0047] Therefore, the present invention adopts the above-mentioned vehicle anti-collision warning method based on machine vision, realizes real-time detection of obstacles in front of the vehicle and dynamic assessment of collision risk, evaluates the risk collision index through the risk assessment module, and dynamically adjusts the safety distance threshold according to the current speed of the vehicle and road conditions.

[0048] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A vehicle anti-collision warning method based on machine vision, characterized in that: The following steps are involved: Step S1, collecting road image data through a camera in an image acquisition module; Step S2: preprocessing the collected image data through an image processing module; Step S3: performing obstacle recognition on the preprocessed image data based on the obstacle recognition module, and determining the position and speed of the obstacle; Step S4: According to the position and speed of the obstacle and the current driving state of the vehicle, the collision risk index R is calculated by the risk assessment module: Among them, V rel is the relative speed between the vehicle and the obstacle; d is the distance between the vehicle and the obstacle; d0 is the safety distance threshold; λ is the adjustment parameter; Step S5: Compare the collision risk index with a preset threshold.

2. A vehicle anti-collision warning method based on machine vision according to claim 1, characterized in that: In step S1, the cameras in the image acquisition module are divided into a front camera, a side camera and a rear camera. The cameras use high-resolution cameras. The front camera is installed above the vehicle windshield, the side cameras are installed at the door positions on both sides of the vehicle, and the rear camera is installed at the rear of the vehicle. The high-speed data interface in the image acquisition module transmits the image data to the image processing module in real time.

3. A vehicle anti-collision warning method based on machine vision according to claim 1, characterized in that: In step S2, preprocessing in the image processing module includes graying, edge detection and feature extraction.

4. A vehicle anti-collision warning method based on machine vision according to claim 1, characterized in that: In step S3, the obstacle recognition module uses the YOLOv8 learning model to perform obstacle recognition, obstacle position determination and obstacle speed determination on the preprocessed image data.

5. A vehicle anti-collision warning method based on machine vision according to claim 1, characterized in that: In step S4, the safety distance threshold d0 is dynamically adjusted according to the current speed of the vehicle and the road conditions. The specific formula is: d0=k×v 2 ×μ; Among them, k is the proportional coefficient; v is the vehicle speed; μ is the road friction coefficient.

6. A vehicle anti-collision warning method based on machine vision according to claim 1, characterized in that: In step S5, the warning module compares the collision risk index R in step S4 with a preset threshold value to determine whether a warning signal needs to be issued.

7. A vehicle anti-collision warning method based on machine vision according to claim 6, characterized in that: Warning signals include visual warning, auditory warning and tactile warning. Visual warning displays the location and distance information of obstacles through warning lights or display screens on the vehicle dashboard. Auditory warning alerts the driver through an audible alarm; Haptic warning alerts the driver through seat vibration.

Images