EDGE DETECTION TECHNIQUES APPLIED TO DEFENCE IMAGERIES

Abstract

Edge detection finds its utilization in numerous applications such as feature extraction, image segmentation, object tracking, diagnosis in medical imaging, defence applications etc. and forms a critical component in many computer vision systems. Number of techniques are being used for filtering out the redundant or less relevant information from the images while preserving the basic structural properties. In this Dissertation, study and implementation of conventional and recent edge detection techniques has been undertaken for defence imageries. Further, a new algorithm has been proposed and applied to defence images. Various techniques like Sobel, Prewitt, Kirsch, Canny, Gabor Wavelet, Structured Forests, SUSAN, Type-2 Fuzzy logic techniques, and the proposed algorithm have been implemented using MATLAB, ver-R2013a. The results of above methods obtained using defence images are presented and interpreted in Chapter four. By looking at the output images, the proposed algorithm is found to be capable of detecting the strong as well as weak edges. Qualitative comparison of the obtained results with respect to all other techniques, clearly establish that the edges obtained by proposed algorithm are better in terms of missing edges, false edges, bifurcation and clarity.