Hough Circles Detection Based on Depth Information Set of Image

Fu-kang ZHU,Yi LIU,Fan-jie MENG,Ming CONG
DOI: https://doi.org/10.13462/j.cnki.mmtamt.2018.05.021
2018-01-01
Abstract:In view of using visual guiding industrial robot for shaft-hole assembly, this paper proposes a method of circular hole detection based on depth information set of image. And this method is successfully carried out on bearing hole detection of the three-axis magnetic fluid sealer. First, the aligned samples of color and depth data stream are collected via a 3D camera and all pixel′s coordinates of the bearing hole in depth image are extracted. Then corresponding data processing are implemented on color image matrices and position information of the bearing hole can be detected by Hough transform eventually. The result turns out that the Hough circles detection based on depth image can be used to improve the weak points of traditional Hough transform which can′t individually identify the specific circle and often occurs false or missing detec-tion in complex background. This approach can provide some helpful instructions for circular hole detection in complex cases.
What problem does this paper attempt to address?