Acta Optica Sinica, Volume. 35, Issue 1, 115002(2015)
Separation Algorithm of Traffic Signs Based on Curvature Scale Space Corner Detection
In view of the detection rate degradation of traffic sign due to multiple signs interconnection, a separation algorithm based on curvature scale space (CSS) corner detection is proposed. The candidate regions of multiple signs interconnection are automatically identified based on red green blue (RGB) normalization color segmentation algorithm and region feature decision criterion. The edge smoothing and contour tracking for the extracted target region are performed. For the obtained contour, the corner detection by using the CSS corner detector based on global and local curvature properties is conducted. And according to the criterion for judging the convexity- concavity of corner and separation points pair matching condition, the separation points pair between the interconnection traffic signs is extracted from the detected corners. The separation line between the separation points pair is achieved by adopting Bresenham algorithm, thus the traffic signs are separated by using the line. Experimental results verify the effectiveness of our method. The proposed method overcomes the problem of the traffic signs over- separation and improves the overall detection performance compared with the existing separation algorithm based on watershed transform and the improved adaptive separation algorithm.
Get Citation
Copy Citation Text
Li Houjie, Qiu Tianshuang, Song Haiyu, Wang Peichang, Wang Pengjie. Separation Algorithm of Traffic Signs Based on Curvature Scale Space Corner Detection[J]. Acta Optica Sinica, 2015, 35(1): 115002
Category: Machine Vision
Received: Apr. 14, 2014
Accepted: --
Published Online: Dec. 15, 2014
The Author Email: Houjie Li (lihoujie@sina.com)