A fast ranging method for pedestrians on the road ahead based on a vehicle-mounted binocular camera

Abstract

The invention is a method for quickly measuring pedestrians on the road ahead based on a vehicle-mounted binocular camera, and discloses a binocular pedestrian detection and ranging method suitable for an automatic driving system. Methods Through the pedestrian detection method based on learning, the pedestrian detection model is trained, and the pedestrian detection model is used to detect pedestrians in the binocular scene. Calculate the matching points of the points selected in the left image detection results in the right image through the color features and scale invariant features, and calculate the disparity value between the matching pairs. The accurate parallax value of the detection result is obtained by finding the median value of the parallax, and the distance between the pedestrian and the camera is calculated according to the geometric relationship between the camera coordinate system and the image coordinate system. Method Aiming at the problems of slow calculation speed and inaccurate distance measurement in the existing methods based on binocular pedestrian detection and distance measurement, the road columnar model is used to speed up the detection speed, and multi-feature point matching is used to improve the accuracy of pedestrian distance measurement in road scenes. accuracy.

Description

technical field [0001] The invention belongs to the technical fields of computer vision, image processing and the like, and in particular relates to a fast distance measuring method for pedestrians on the road ahead based on a vehicle-mounted binocular camera, which can be used for a vehicle automatic driving assistance system in complex scenes. Background technique [0002] Self-driving cars can not only greatly improve the efficiency and safety of the transportation system, but also save us a lot of precious time, and will be the mainstream of future car development. In the automatic driving system based on visual navigation, target detection and ranging are the key technologies, which directly determine the performance of the automatic driving system. [0003] In vision-based navigation, binocular vision technology is the mainstream application technology in autonomous driving assistance systems. Binocular ranging uses the principle of binocular stereo vision, that is, t...

AI Technical Summary

Problems solved by technology

[0006] The technical problem to be solved by the present invention is to overcome the shortcomings of insufficient detection speed and insufficient detection accuracy existing in the existing binocular-based pedestrian detection and ranging, and make full use of the characteristics of binocular vision to provide a fast and effective road scene downlink Human detection and ranging method

Images