Analysis of Edge Detection Algorithms on DIAL (Digital Image Analysis Laboratory)

摘要

As part of an effort to deliver to the Defense Mapping Agency in FY85 a semiautomated feature extraction system, the Engineer Topographic Laboratories is conducting a feature extraction study. One of the objectives of this study is to determine whether selected edge-finding algorithms can be used to find such lineal features as roads and boundaries. The purpose of this document is to report on progress to date on an analysis of several edge-detection algorithms that exist on ETL's Digital Image Analysis Laboratory image-processing facility. In addition to evaluating the algorithms, this study investigated the possibilities of integrating the algorithms' resultant edges with other feature classification algorithms. Three fundamental edge-detection methods were compared: 1. A nonlinear edge-enhancement operator. 2. A directional edge operator. 3. An extraction of the zero-crossings from filtering an image with the Laplacian of a symmetric Gaussian function.