Unsupervised iterative detection of land mines in highly cluttered environments

Abstract

An unsupervised iterative scheme is proposed for land mine detection in heavily cluttered scenes. This scheme is based on iterating hybrid multispectral filters that consist of a decorrelating linear transform coupled with a nonlinear morphological detector. Detections extracted from the first pass are used to improve results in subsequent iterations. The procedure stops after a predetermined number of iterations. The proposed scheme addresses several weaknesses associated with previous adaptations of morphological approaches to land mine detection. Improvement in detection performance, robustness with respect to clutter inhomogeneities, a completely unsupervised operation, and computational efficiency are the main highlights of the method. Experimental results reveal excellent performance.

Small bottom object density analysis from side scan sonar data by a mathematical morphology detector

“…Mathematical morphology (MM) applied to image processing aims at analysing the discrete geometry of objects within the image under study through set theory and topology [1,4]. It provides a wide range of image to image operations for processing both binary and grayscale images.…”

Stand-off thermal IR minefield survey: system concept and experimental results

“…The implement method follows roughly the work presented in [11], where the authors apply MM operators on spectral data taken at a specific time, whereas in the presented method multi-temporal and multi-spectral data is available. The flow diagram for the detector is depicted in Figure 4.…”

Intravital leukocyte detection using the gradient inverse coefficient of variation

“…If denotes an image then the mean of the image gradient over the entire contour computed in the outward normal direction is given by (7) where is the unit outward normal to the contour at and is the length of the contour given by . The incorporation of directional information yields superior results when the contour intersects adjacent object boundaries.…”

“…Based on local edge strength, radial distribution, and symmetry, the image pixels are ranked to provide a sorted list of the most likely positions of all dominant ellipses. Detection methods, based on mathematical morphology, have been introduced by Batman et al [7], and by Mukhopadhyay et al [8]. In the related work of Egmont-Petersen et al leukocytes in contact with the vessel wall are detected using a neural network that is trained with synthetic images generated by a stochastic model [9].…”