Real time pattern selective image fusion

We propose to develop an image fusion system to fuse images from an intensified TV and a thermal imaging system into a single image by using a pattern selective approach. Pattern selective approach is different from pixel averaging approach in its ability to retain the contrast of the important elements in the scene as obtained by the imaging systems. The sytem will be able to look into the shadows, and see subtle thermal variations in the scene, see low light emissions and reflections, while retaining individual detail from each sensor. The system will be built and delivered during Phase II of the program. During the Phase I of the program, the recording of scenes from the cameras will be obtained from the customer after which they will be digitized and aligned to each other. Differernt fusion algorithms will be tried to find the best algorithm for the image combination. Based on the results, a high level design will be done for the system. BAsed on teh design the system will be built during the Phase II.