Automatic parking sensing method integrating look-around image and laser radar

Abstract

The invention relates to an automatic parking sensing method fusing a look-around image and a laser radar, and the method comprises the following steps: 1, obtaining an original image through four fisheye cameras disposed on a vehicle body, and carrying out the distortion correction of the original image; step 2, by selecting four mark points with known ground coordinates, respectively performing projection transformation on the original image after distortion correction to obtain four aerial view images projected to the ground; 3, splicing the four aerial view images based on the spatial position consistency of the mark points to obtain an all-round view image; step 4, inputting the look-around image into a parking space detection model for parking space detection to obtain pixel coordinates of the parking space; 5, the global coordinates of the parking spaces are obtained according to the pixel coordinates of the parking spaces, and automatic parking path planning is carried out. Compared with the prior art, the method has the advantages that the tracking algorithm of the parking spaces is simplified, the intelligence and robustness of path planning are improved, and the safety in the automatic parking process is guaranteed.

Description

technical field [0001] The invention relates to the technical field of automatic parking, in particular to an automatic parking perception method for fusing surround-view images and laser radar. Background technique [0002] The perception information sources of existing automatic parking schemes are mostly based on the surround view image, the fusion of the surround view image and millimeter-wave radar, or the fusion of the surround view image and the ultrasonic sensor. The surround view image is used to detect parking spaces. The detection methods include traditional vision methods and deep learning methods. Judging whether there is a car in the target parking space through millimeter-wave radar and ultrasonic sensors, and obtaining the distance of obstacles, and the existing automatic parking solutions mostly use the kinematic information of the own vehicle for positioning, and then use positioning to track the coordinates of the parking space. [0003] Parking solutions ...

AI Technical Summary

Problems solved by technology

[0003] Parking solutions that only rely on surround-view images cannot perceive obstacles. The sensing distance of ultrasonic sensors is relatively short. The range of sensing angles of millimeter-wave radars is limited, and there are blind spots. The above sensors are generally used to judge whether there is a car in the parking space. For obstacles It can only be stopped; in addition, the positioning method that only relies on the vehicle's motion state will cause cumulative errors

Therefore, for the automatic parking function, the existing methods do not have a comprehensive perception of the environment. On the one hand, it limits the intelligence of automatic parking, and on the other hand, it is difficult to ensure the safety of the automatic parking process.

Images