Image Matching And Fusion Algorithm Based On DSP

Sonar image processing plays an important role in underwater signal processing, has great value in economic and military. Image matching and fusion technology is an important part of image processing technology, in recent years, image matching and fusion technology's application is more and more widely. But it is lack of the application of this technology in sonar image processing.This paper mainly studied the normal distributions transform (NDT)image matching algorithm, and its application in sonar image matching. Because there is noise in sonar image,image's noise reduction is necessary. The NDT algorithm is based on probabilistic model,it is no need to feature extraction,but image segmentation is helpful. So the main research is included three parts: the pre-process of the image, image segmentation,image matching and fusion of sonar image.The main research of the pre-process of sonar image is the methods of the sonar image's noise reduction and the image segmentation. Several classical methods of sonar image's noise reduction are listed, the median method, the technology based on wavelet and so on. The methods of the image segmentation are iterative threshold segmentation method, the maximum entropy method and Otsu. Under different noise conditions, the paper studied these kinds of the noise reduction method and image segmentation method on the sonar images,gives the results and analysis of each method.The image matching. This part studied the NDT algorithm, gives its basic principle,simulation results and the results of image fusion.Finally, the image matching and fusion algorithm was achieved on DSP. The results of the NDT image maching algorithm have been given.