LAUSR

Abstract The objective is to solve the problem of image recognition in intelligent vehicles and optimize the judgment of obstacles and the planning of subsequent routes of intelligent vehicles. Methods: The machine vision technology is used to collect images of relevant road segments, process these images with graying and binarization methods, and simulate the proposed method to observe its effect through data collection. Through the analysis of continuous direct-through driving after encountering the obstacles, it is found that the intelligent vehicles have small traveling errors once the routes are identified and planned. In addition, the error in the x-direction is no more than 0.006 m, while the error in the y-direction is no more than 0.003 m. The recognition effect of the vehicle has reached the expected result. Through the analysis of turning and rotary driving after encountering the obstacles, it is found that after the intelligent vehicles have identified the obstacles and planned the routes, a sudden change in the amplitude of the curve during the turn is caused. In addition, during the turning driving, the error in the x-direction is no more than 0.02 m, while the error in the y-direction is no more than 0.05 m. During the rotary driving, the error in the x-direction is no more than 0.03 m, while the error in the y-direction is not more than 0.04 m. The error variation range is also within the allowable error range. Through the research in this paper, it is found that the error of intelligent vehicle is within the allowable range and achieves the expected effect. Although there are some shortcomings in the experimental process, it can still provide an experimental basis for obstruction detection and route planning of intelligent vehicles in the later stage.