An automatic registration method based on runway detection

Abstract

Runway is seen distinctly that is a crucial condition in the process of approaching and landing. One of the enhanced vision methods is image fusion method between the infrared and visible images in EVS (Enhanced Vision System). The image registration plays a very important role in image fusion. So, an automatic image registration method is proposed based on the accurate runway detection. Firstly, runway is detected using DWT (discrete wavelets transform) from the infrared and visible images respectively. Then, a fitting triangle is constructed according to the edges of runway. The corresponding feature points extracted from the middle points of edges and the centroid of triangle are used to compute the transform parameters. The results of registration are more accurate and efficient than those of registration based on mutual information. This method is robust and has less computation which can be applied to real-time system.