A hybrid method for ellipse detection in industrial images

作者：

Highlights：

• A hybrid method with high reliability to detect ellipses in low SNR industrial images is proposed.

• Using candidate elliptic arcs with high fitting precision as seed arcs to group with other arcs.

• A novel method to extract regions which may contain one ellipse is proposed.

• New selection scores for evaluating hypothetical ellipse are used.

• Experimental results show high reliability of proposed method.

摘要

•A hybrid method with high reliability to detect ellipses in low SNR industrial images is proposed.•Using candidate elliptic arcs with high fitting precision as seed arcs to group with other arcs.•A novel method to extract regions which may contain one ellipse is proposed.•New selection scores for evaluating hypothetical ellipse are used.•Experimental results show high reliability of proposed method.

论文关键词：Ellipse detection,Edge following method,Hough transform,Industrial application,Reliability

论文评审过程：Received 5 June 2016, Revised 9 February 2017, Accepted 3 March 2017, Available online 6 March 2017, Version of Record 12 March 2017.

论文官网地址：https://doi.org/10.1016/j.patcog.2017.03.007