Obstacle recognition method in intelligent aided driving

The invention discloses an obstacle recognition method in intelligent aided driving, and the method comprises the following steps of S1, collecting an obstacle image, namely collecting the obstacle image in a scene through a binocular camera; S2, analyzing the obstacle distance, namely according to a binocular distance measurement principle, performing analysis in combination with the acquired obstacle image to obtain the distance information of an obstacle; S3, constructing an image processing system, combining a deep belief network with a support vector machine algorithm, and constructing and training the image processing system; and S4, recognizing the obstacle, namely transmitting the obstacle image to an image processing system by using a machine vision technology to complete the recognition of the obstacle in the intelligent aided driving. According to the method, the deep learning, an SVM algorithm model and a binocular distance measurement principle are comprehensively utilized, so that the distance i