FAST AND ROBUST MULTIMODAL REMOTE SENSING IMAGE MATCHING METHOD AND SYSTEM

摘要

Disclosed are a fast and robust multimodal remote sensing image matching method and system, capable of integrating different local feature descriptors for automatic multimodal remote sensing image matching. First, a local feature descriptor such as Histogram of Oriented Gradient (HOG), local self-similarity (LSS) or Speeded-Up Robust Features (SURF) is extracted for each pixel of an image to form a pixel-by-pixel feature expression diagram. Then a three-dimensional Fourier transform is used to establish a fast matching similarity measure in a frequency domain based on the feature expression diagram. Finally, a template matching policy is used for homonymy point identification. In addition, for the matching method and system, the present invention further proposes a new pixel-by-pixel feature expression technology, namely channel feature of orientated gradient (CFOG). The technology is superior to a pixel-by-pixel feature expression method based on descriptors such as HOG, LSS and SURF in terms of matching performance and computational efficiency. According to the present invention, non-linear radiation difference among multimodal images such as visible light, infrared, laser radar, synthetic aperture radar and map can be effectively overcome; homonymy points are quickly and accurately identified among images; and automatic image matching is realized.