Improving edge detection using particle swarm optimisation

Traditional edge detection approaches often result in broken edges especially in noisy images. This study presents a particle swarm optimisation based algorithm to detect edges continuously in such images. In this paper, a new objective function and a new encoding scheme are proposed to address noise and reduce broken edges. The newly developed algorithm is compared to a modified version of the Canny algorithm as a Gaussian-based edge detector based on Pratt's figure of merit measure. Experimental results indicate that the newly developed algorithm can perform better than the Canny and Sobel algorithms in the images.