DEVELOPMENT OF IMPROVED EDGE DISCOVERY ALGORITHM FOR NOISY COLOURED IMAGES USING PARTICLE SWARM OPTIMIZATION

ABSTRACT

This research work presents an improved edge discovery algorithm using particle swarm optimization based on vector order statistics. In order to address the shortcomings of the existing traditional edge discovery algorithms that could not process coloured images directly unless been converted to gray scale, Vector Order Statistic technique was employed. The profile edge intensity was applied to generate a collection scheme. This collection scheme was obtained using set of pixels with respect to the step and roof edge profiles. The collection scheme was then applied as a mask in both vertical and horizontal directions to the images.An improved figure of 0.8480 and 0.8458 were obtained in clean and noisy environment respectively using the PFOM. The proposed algorithm achieved a PSNR of 57.7320dB in environment containing ≤ 33% of noise level.