An Adaptive Canny Edge Detector using Histogram Concavity Analysis
Jun Zeng, Dehua Li · International Journal of Digital Content Technology and its Applications · 2011
The traditional Canny edge detector has some drawbacks. Gaussian filter can’t remove the impulsive noise. Moreover, it is difficult to automatically select the dual-threshold. Especially when the noise intensity increases, the dual-threshold selection method of traditional Canny detector is invalid. In this paper, we present an adaptive Canny edge detector using histogram concavity analysis. The improved detector uses switching median filter to remove the impulsive noise and then uses Gaussian filter to smooth other types of noise. It determines the dual-threshold through histogram concavity analysis. Experimental results show that the proposed detector is better than the LOG detector and traditional Canny detector in strong noise environment. The improved Canny detector can automatically select the dual-threshold.