A Wide Range Multi-obstacle Detection Method Based on VIDAR and Active Binocular Vision

Binocular vision-based obstacle detection system has a wide range of applications, such as intelligent vehicles, mobile robots and other fields. However, the system has a large blind area because of the fixed camera pose of the traditional binocular vision system. This limits the system to accuratel...

Full description

Saved in:
Bibliographic Details
Published in:IAENG international journal of applied mathematics 2023-03, Vol.53 (1), p.1-12
Main Authors: Zhu, Ruoyu, Xu, Yi, Wang, Liming, Sun, Teng, Yu, Jinxin, Ding, Shaohong, Wang, Yuqiong, Guo, Dong, Wang, Pengwei, Liu, Bingzheng
Format: Article
Language:English
Subjects:
Accuracy
Algorithms
Binocular vision
Cameras
Field of view
Intelligent vehicles
Obstacle avoidance
Research methodology
Target detection
Vehicles
Vision systems
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Binocular vision-based obstacle detection system has a wide range of applications, such as intelligent vehicles, mobile robots and other fields. However, the system has a large blind area because of the fixed camera pose of the traditional binocular vision system. This limits the system to accurately obtain the depth information of obstacles only in the field of view where two cameras overlap, and it is not possible to obtain accurate information otherwise. Aiming at this problem, we propose a wide range multi-target active binocular vision obstacle detection method based on VIDAR (Vision-IMU based detection and range method) for intelligent vehicles. First, two separate cameras are used to detect the three-dimensional obstacles by VIDAR. Second, the optical axis angle of the camera is adjusted in real-time according to the position of the obstacle. This causes the target to be detected such that it always lies in the overlapping field of view of the two cameras, and the obstacle can be actively tracked. Third, the obstacle search strategy is proposed to maximize the number of obstacles detected by the active binocular system. Last, experimental results show that compared with the traditional obstacle detection method based on binocular vision, the proposed method can not only detect unknown types of obstacles, but also have a wider detection range and a guaranteed ranging accuracy.
ISSN:1992-9978
1992-9986