Multi-spectral Imaging To Detect Image Registration Techniques

Image registration is an important topic in the field of image processing and is widely applied in remote sensing, computer vision, motion estimation, the medical image analysis as well as other fields. It is the first step of fusion process, particularly in data layer image fusion. There are relative translation, rotation, zoom and even different ratio distortion between the images gathered from different sensors, different imaging modalities and different time. In order to actualize a comprehensive analysis and application, the image registration problems have to be resolved at the first. With the multi-spectral imaging detection technology development, there will be higher requirements for accurate real-time images automatic registration under the low signal to noise ratio environment.Based on the analysis of the multi-spectral imaging night vision technology, a edge detection and phase-related automated registration method for multi-spectral images was designed after a research on methods of multi-spectral images automated registration. Compared with the inhered arithmetic based on the pixel's spatial relativity, the new method was appraised in computing efficiency and accuracy. It is proved that the new registration method is faster and more reliable than the arithmetic based on the pixel's spatial relativity and has great significance for real-time multi-spectral image registration .