Door Detection Based on Geometrical Features and Harris Corner

Door is a very significant element as it enables person to enter house or room. Though identifying doorway an easy task for regular person, robots visually impaired people challenging task. To overcome this challenge, we have proposed door detection method. Our method based on Prewitt edge and Harris corner detector. Here, are using number of predefined rules detect the doorframe correctly. establish robustness our method, formed substantial dataset scene images that captured in various unfamiliar environments. experimental results validate robust against changes viewpoint, shapes, occlusions, illumination, colors, sizes, orientations, textures door. The show reaches 87.45% accuracy well achieves lower false positive rate computational time.