Ground plane identification method and device

Abstract

The invention discloses a ground plane identification method and device. The method comprises the following steps: preprocessing left and right eye original images of a current period to obtain a first dense disparity map; obtaining a current period pose parameter; affine transformation is carried out on a to-be-rotated image and the first dense disparity map according to the current pose parameter to obtain a rotated image and a second dense disparity map, wherein the to-be-rotated image is a preset left-eye original image or a preset right-eye original image; and identifying the rotated image and the second dense disparity map image to obtain a ground plane area. The method is high in applicability, can accurately recognize the ground plane under the condition that the vehicle body vibrates, simplifies a detection algorithm based on the disparity map on the basis, further improves the detection speed, and can be widely applied to a ground unmanned platform or an automatic driving auxiliary system and the like.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a ground plane recognition method and device. Background technique [0002] Ground level detection is a key technology in the field of intelligent driving. The ground level detection methods currently proposed mainly include the following: [0003] 1. Detection methods based on active sensors such as ultrasonic sensors, laser radars, and millimeter-wave radars. Utilizes active sensors to detect the area ahead to determine a passable ground level. Ultrasonic sensors are cheap but have a short detection distance, so they are not suitable for occasions with fast movement speeds; lidar and millimeter-wave radars have long detection distances and high accuracy, but are expensive and bulky, and are not suitable for low-cost and lightweight applications. body of the car. [0004] 2. Detection method based on monocular vision. Mainly use 2D information such as chromaticity a...

AI Technical Summary

Problems solved by technology

Such methods are susceptible to lighting, shadows, etc., and cannot achieve better performance in more complex environments

[0005] However, in actual use, due to vibrations (wheels entering pits, rolling over stones, etc.) when the car body is moving forward, the baseline of the binocular vision sensor is not parallel to the ground (that is, there is a roll angle between the binocular vision sensor and the ground) In the case of the situation, it is prone to the problem that the ground plane cannot be calculated and detected

Images