Abstract : In order to promote the efficiency and accuracy of image registration, this paper proposes an improved registration algorithm for fruit tree images acquired by a dual-sensor vision system. In the algorithm, feature points are extracted by FAST detectors from Gaussian scale space, and main orientation is determined and SURF descriptor is created with statistical method of neighbor intensity distribution, matching pairs are determined with a relative ratio method and a iterative purifying method in succession. Experimental results show that the proposed algorithm outperforms previous algorithms comprehensively.
https://hal.inria.fr/hal-02180000
Contributor : Hal Ifip <>
Submitted on : Friday, July 12, 2019 - 2:12:47 PM Last modification on : Friday, July 12, 2019 - 2:13:51 PM
File
Restricted access
To satisfy the distribution rights of the publisher, the document is embargoed
until : 2021-07-12
Juan Feng, Lihua Zeng, Jianping Li. Fruit Tree Image Registration Based on Improved FAST Algorithm. 10th International Conference on Computer and Computing Technologies in Agriculture (CCTA), Oct 2016, Dongying, China. pp.212-217, ⟨10.1007/978-3-030-06155-5_21⟩. ⟨hal-02180000⟩