Machine vision-based forward-rotating-avoidance type three-dimensional parking obstacle avoidance method and system

Abstract

The application discloses a front-moving rotation avoidance type three-dimensional parking obstacle avoidance method and system based on machine vision, wherein an image containing a garage bottom plate and a target detection object is collected, the garage bottom plate and the target detection object in the collected image are recognized through a YOLO target detection algorithm, and position information of the garage bottom plate and the target detection object is acquired; an adaptive histogram equalization algorithm is used to process the image of the garage bottom plate, edge detection is performed, and the contour of the garage bottom plate is drawn, four points in the contour are acquired through an iterative method, a perspective transformation matrix is calculated according to the four points, and a one-dimensional first-order equation of six line segments for distinguishing the equipment area and the operation area of the obstacle avoidance area is obtained through inverse deduction of the formula of the perspective transformation matrix, the central point of the target detection object is taken as a main judgment point, the y value of the point is calculated, and then the y value is compared with the actual y value of the central point, so that the relationship between the target detection object and the obstacle avoidance area is judged.

Images