Multi-source Image Fusion Based On Fuzzy Reasoning

Multi-sensor message integration means integrating the data from different sensor sources to understand the object better.Image Fusion technology is to fuse the message collected from multi-channel about the same object, by distilling the message of each channel and integrating to one image witch provides to be observed or more processed. Its purpose is to integrate the message from the multi-wave-band of single sensor or the message from multi-sensor, suppress the redundancy or conflicting possibly existing in the multi-sensor to increase the transparency of the message and improve the precision of decrypting, reliability and use rate of the image, in order to form the distinct, complete, accurate message description.At the beginning of this paper, author introduced and compared some existing method of image fusion, then imported the theory of fuzzy inference. At the basis of this theory we constructed a fuzzy inference system, presented a fusion algorithm, and testified it by imitation experiments.At next step, this paper applied the algorithm into multi-sensor image road identifying, to get rid of the road shadow of visible light images. And author created a special fuzzy inference system, and explained all the part of this system in detail. Especially in the input membership function, and the manner of machine learning. Finally we testified the feasibility of this system in experiments.