Abstract
The unmanned surface vehicle has the characteristics of high maneuverability and flexibility. Object detection and tracking skills are required to improve the ability of unmanned surface vehicle to avoid collisions and detect targets on the surface of the water. Mean-shift algorithm is a classic target tracking algorithm, but it may fail when pixel interference and occlusion occur. This article proposes a tracking algorithm for unmanned surface vehicle based on an improved mean-shift optimization algorithm. The method uses the self-organizing feature map spatial topology to reduce the interference of the background pixels on the target object and predicts the center position of the object when the target is heavily occluded according to the extended Kalman filter. First, a self-organizing feature map model is built to classify pixels in a rectangular frame and the background pixels are extracted. Then, the method optimizes the extended Kalman filter solution process to complete the prediction and correction of the target center position and introduces a similarity function to determine the target occlusion. Finally, numerical analyses based on a ship model sailing experiment are performed with the help of OpenCV library. The experimental results validated that the proposed method significantly reduces the cumulative error in the tracking process and effectively predicts the position of the target between continuous frames when temporary occlusion occurs. The research can be used for target detection and autonomous navigation of unmanned surface vehicle.
Highlights
Environmental awareness is essential for unmanned surface vehicle (USV) to acquire the autonomous ability in changing environments.[1]
As for the problems that background pixel interference and target loss while occlusion occurred, this article proposes an improved mean-shift optimization algorithm based on Self-organizing feature map (SOFM) and extended Kalman filter (EKF), which can improve the tracking accuracy while ensuring the antiocclusion ability
The SOFM is introduced into the traditional mean-shift target tracking algorithm to detect target pixels using the time continuity of SOFM and the constraints of the spatial topology
Summary
Zuquan Xiang[1,2], Tao Tao[1,2], Lifei Song1,2 , Zaopeng Dong1,2 , Yunsheng Mao1,2 , Shixin Chu[1,2] and Hanfang Wang[3]
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.