In this paper, we discuss a simple but an efficient fuzzy reasoning based edge detection as a feature extraction technique for an automated landmine detection application. Edges form the boundaries of the various regions in an image. Features such as corners, lines and curves can be extracted from the detected edges which characterize the structure of various objects present in the image. The image is scanned using a 2X2 windowing mask. The four inputs of the designed fuzzy inference system are represented by the 4 pixels in the windowing mask. The rules are applied on the mask and the output is mapped to the defined Sugeno singleton set and  the processed pixel in the input image is checked for the occurrence of an edge pixel. The proposed algorithm is tested on the acquired infrared landmine images and the performance is compared with some of the standard edge detection methods. The proposed algorithm shows improved results.

Keywords:  edge, feature, fuzzy, landmine, mask