A Survey on Conventional Edge Detection Techniques
Sharaf Ansari, Srikanth Prabhu, N. Gopalakrishna Kini, Govardhan Hegde, Yusuf Haider · 2014
An edge defines a boundary between two homogeneous regions. By highlighting the boundaries it can be used to identify different objects present in an image, therefore it is of integral importance in image processing as it filters out insignificant information from the image. It finds use in important applications such as computer vision and motion detection and 3d reconstruction. In this paper, we review five conventional edge detection methods: Sobel, Canny, Prewitt, Roberts, and Laplacian of Gaussian. We find that Canny edge detection though computationally expensive gives superior results