Vision-based recovery of the unmanned underwater vehicle(UUV) is a hard task due to complex underwater optical conditions. Based on the fork-carrying-pole recovery system, a visual positioning method based on asymmetric guiding light array is proposed. In this method, the guiding light array is detected and located by camera, so as to complete the recovery. This paper mainly solves three problems Firstly, an asymmetric L-shaped guiding light array is designed based on the characteristics of underwater optical conditions and fork-carrying-pole recovery system; Secondly, in condition of different background light intensity, we proposed a set of detection methods for L-shaped light arrays. According to the change of background light, the threshold segmentation interval can be adaptively selected, and the segmentation threshold of light array image can be selected accurately. The pseudo light source can be eliminated by BLOB feature combined with logic regression. Thirdly, a 6-degree-of-freedom(6-DOF) pose estimation method of UUV is designed, and a location method of tracking after single Target Segmentation (TASTS) is proposed to improve the location method. This novel TASTS can solve the problem of tracking and positioning of partial target occlusion within 50%. Finally, the UUV recovery experiment is carried out. The experimental results show that this method improves the success rate of positioning, eliminates the close visual blind area of visual recovery. This method improves the accuracy, whose error is reduced from 0.5 m to 0.1 m in the final stage, and increases the rapidity rate of UUV autonomous recovery by 20%.
[1]
Bo Wang,et al.
AUV docking experiments based on vision positioning using two cameras
,
2015
.
[2]
Edward Jones,et al.
Improved image processing-based crop detection using Kalman filtering and the Hungarian algorithm
,
2018,
Comput. Electron. Agric..
[3]
Wei Zhang,et al.
Underwater recovery realization for an AUV using positioning-to-line strategy
,
2014,
2014 IEEE Conference and Expo Transportation Electrification Asia-Pacific (ITEC Asia-Pacific).
[4]
Siva Ram Krishna Vadali,et al.
Reliable pose estimation of underwater dock using single camera: a scene invariant approach
,
2015,
Machine Vision and Applications.
[5]
Wei Zhang,et al.
Autonomous Underwater Vehicle Vision Guided Docking Experiments Based on L-Shaped Light Array
,
2019,
IEEE Access.
[6]
C. Deltheil,et al.
Simulating an optical guidance system for the recovery of an unmanned underwater vehicle
,
2000,
IEEE Journal of Oceanic Engineering.
[7]
N. Otsu.
A threshold selection method from gray level histograms
,
1979
.