Title :
A fast registration algorithm based on combination of SURF descriptor and energy minimization for detecting new-type surface mount components
Author :
Gao, Hongxia ; Chen, Xinyuan ; Chu, Fuguo ; Mai, Qian
Author_Institution :
Coll. of Autom. Sci. & Eng., South China Univ. of Technol., Guangzhou, China
Abstract :
In electronic manufacturing process, detecting characteristics of new-type surface mount components become smaller and smaller. So they are difficult to be located precisely because of image noise. To improve the detecting accuracy, this paper analyzes the difficulties of detecting SMCs and then proposes a new fast registration algorithm based on combination of energy minimization and SURF descriptor to locate new-type surface mount components´ rotating angle, which utilizes SURF descriptor to estimate rough rotation angle and then uses energy minimization to improve the accuracy of angle detected by SURF. The experiments demonstrate that the proposed algorithm can detect surface mount components´ position and angle efficiently and is more reliable than the algorithm based on edge detecting.
Keywords :
edge detection; feature extraction; image denoising; image registration; object detection; surface mount technology; SMC detection; SURF descriptor; electronic manufacturing process; energy minimization; fast registration algorithm; image noise; new-type surface mount component characteristics detection; new-type surface mount component rotating angle location; rough rotation angle estimation; Accuracy; Feature extraction; Image edge detection; Minimization; Noise; Surface treatment; Transforms; Energy Minimization; Registration; SURF; Surface Mount Components;
Conference_Titel :
Control Conference (CCC), 2012 31st Chinese
Conference_Location :
Hefei
Print_ISBN :
978-1-4673-2581-3