Authors: Gu Gong; Hua Zhu
Addresses: School of Mechatronic Engineering, China University of Mining and Technology, Xuzhou 221116, China; School of Computer Science and Technology, Jiangsu Normal University, Xuzhou 221116, China ' School of Mechatronic Engineering, China University of Mining and Technology, Xuzhou 221116, China
Abstract: The redundant wavelet transform has the translation invariant property, and the motion estimation can achieve better results in the redundant wavelet domain. This paper proposes the redundant discrete wavelet transformation (RDWT) for motion estimation and compensation in video compression and compensation of moving target image, and a fast adaptive block matching motion estimation algorithm based on redundant wavelet transform. This paper also introduces the methods of image domain search, starting point prediction and motion estimation into the redundant wavelet transform, and the study method based on optical flow which can enhance the transmission robustness with low redundancy. The experimental results show that the algorithm can effectively reduce the time required for motion estimation and improve the efficiency of motion estimation, and the subjective quality of the motion estimation is very good. The research results can provide technical support for robot vision system.
Keywords: redundant wavelet; pattern recognition; image features; time-frequency spectrum.
International Journal of Wireless and Mobile Computing, 2017 Vol.13 No.4, pp.323 - 329
Received: 17 Apr 2017
Accepted: 18 Aug 2017
Published online: 11 Jan 2018 *