Fuzzy C-means and mathematical morphology for mine detection in IR image

Detection and clearance of a buried are difficult problems with lots of environmental and economical implication. In this work, the mine detection is tackled in a broader context of preprocessing and texture segmentation for the data associated with infrared sensor. Principal component analysis is used to enhance the contrast by extracting the whole dynamic information contained in a sequence of images. Texture parameters, and fuzzy C-means clustering method are proposed to segment background and mine like objects. For the residual clutter in a segmented image, a post-processing step is employed based on morphological reconstruction filter that yields accurate detection result