Target Segmentation and Image Enhancement through Multisensor Data Fusion

摘要

This thesis effort develops a data fusion process for laser radar and forward looking infrared sensors. The process contains an algorithm which segments man-made objects away from a variety of backgrounds. Importantly, the segmentation algorithm does not use pre-stored templates. Other algorithms within the process also enhance and extract target features from both laser radar and infrared images. The overall process provides target details which were not easily found in either one of the sensors by themselves. Furthermore, key algorithms within the overall process have possible optical implementations. (Author)