Fast Scene Matching Algorithm Based in Image Edge Detection

摘要

The problem of scene matching is very important in many practical applications such as military target recognition and scout aircraft navigation, terminal guidance of weapon system, etc. This paper firstly discusses the grayscale distortion of image for guidance of scene matching, and select Lateral Inhibition Compete (LIC) method is to detect image edge. Then, according to the main idea of Sequential Similarity Detection Algorithm (SSDA), a fast scene matching algorithm based on the edge feature of image for guidance is proposed. Finally, the comparative experimental results in relation to Absolute Difference (AD) and Normalized Product correlation (NProd) matching algorithm are given and it is indicated that the algorithm is more efficient.