Segmentation of infrared images
MetadataShow full item record
In this thesis, we study some segmentation methods for infrared images. First we summarize the characteristics of infrared images and survey different approaches of segmentation. Our objective is to develop an image segmentation method for surveillance images, which helps isolate all major objects in the image. We study fuzzy logic based segmentation for this purpose. Our study shows that use of a suitable thresholding method is important prior to use of the fuzzy logic based method for refinement of boundaries of objects of interest. The existing fuzzy logic based method requires prior knowledge of the size of the target objects. We remove this requirement by using adaptive global thresholding heuristic based on image histogram. Next we extend this segmentation method for object matching between frames of surveillance image sequence and for object tracking.