The Research And Realization Of Multisensor Image Fusion

Multisensor image fusion refers to the techniques that integrate complementary information from multi-image sensor data so that the fusion images are more suitable for the purpose of human visual perception and computer process. The main work of this thesis is concluded as follows: Firstly, several image fusion schemes are introduced including the false color fusion methods and the ones based on multiresolution analysis. Evaluation methods are also analyzed briefly.Secondly, the false color fusion method proposed by A.M.Waxman in Massachusetts Institute of Technology (MIT) is analyzed, based on which two new predigested methods are proposed. In the first predigested method, the information from the visible light image is emphasized using the module of receptive field. The parts in the infrared image are also emphasized whose temperature is higher or lower than it is around. Then they are shown in different colors. Considering the noise in the infrared image, in the second predigested method, only the part in the infrared image whose temperature is higher, and the information from the visible light images is emphasized.Thirdly, the method proposed by Toet is discussed and simulated. The improved Toet method is applied in analyzing the difference between the result image with gray fusion method and the source images.Fourthly, the fusion methods based on multiresolution analysis are studied. They are different because of different wavelet transforms and diverse fusion rules, the characteristics of which are concluded from the software simulating.Finally, a scheme for realizing the mutisensor image fusion system is proposed. It completes the missions of image fusion, target reorganization and can be surveilled from the host computer. Further more, two projects in software are put forward depending on the different amount of data. Finally, image fusion methods are emulated in evaluation module card of C6711.