Comparative Analysis of Edge Detection Techniques for extracting Refined Boundaries

Edge detection is an important task in Computer Vision for extracting meaningful information from digital images. The main goal of our proposed technique is to obtain thin edges, so that the result is more suitable for further application such as boundary detection, image segmentation, motion detection/estimation, texture analysis, object identification and so on. We tested four edge detectors that use different methods for detecting edges and compared their results under a variety of situations to determine which detector was preferable under different sets of conditions. This data could then be used to create a multi-edge-detector system, which analyzes the scene and runs the edge detector best suited for the current set of data. For one of the edge detectors we considered two different ways of implementation, the one using intensity only and the other using color information. We also considered one additional edge detector which takes a different philosophy to edge detection. Rather than trying to find the ideal edge detector to apply to traditional photographs, it would be more efficient to merely change the method of photography to one which is more conducive to edge detection. It makes use of a camera that takes multiple images in rapid succession under different lighting conditions.