Fencing systems: Several fencing systems are frequently used in order to designate the locations of high risk in mine infected areas. We have developed an automated approach for the identification of fencing indicators that exhibit a regular pattern in terms of co-linearity and periodicity. A representative example is the case of periodically placed poles, with their shadow oriented in a certain direction. The method initially involves an oriented denoising algorithm, followed by a shadow detection method, based on the HLS color space representation of the image. The Hough transform is used for the detection of the principal orientations (co-linearity criterion), while linear patterns that exhibit equal spacing regularity are then selected using the FFT transform (periodicity criterion) [3].
Land cover: One of the requirements of an air- or space-borne minefield area reduction system is the characterization of the land cover of the surveyed area. Its major aim is the discrimination between areas of human activity (e.g. residential areas, agricultural fields) and natural undisturbed environments (areas with high and dense vegetation). In the case of high resolution airborne images, the defined land cover classes have a non homogeneous appearance with respect to spectral and textural responses, something that hampers the classification efficiency of conventional pixel-based classifiers (like Maximum Likelihood, fuzzy c-means clustering, neural networks). Taking into account this constraint, we have investigated a variety of classification approaches based on the framework of Markov Random Field theory, in order to
In the last decade, several humanitarian demining actions have acknowledged the role of remote sensing as a useful tool, able to enhance the productivity, cost-effectiveness and safety of ground-based minefield detection methods [1] [2] [3] [4]. Air- and s
capture contextual information, in terms of spatial consistency. The investigated methods involve non-causal energy-based models defined on the image lattice, hierarchical, multiresolution models defined on pyramidal representation of the image, as well as hierarchical Markovian models defined on the hierarchy of multiscale region adjacency graphs [7].
References
[1] B. Maathuis, "Remote Sensing based detection of landmine suspect areas and minefields", PhD Thesis, Department of Geosciences, University of Hamburg, 2001.
[2] S. Batman, J. Goutsias, ”Unsupervised Iterative Detection of Land Mines in Highly Cluttered Environments”, IEEE Trans. on Image Processing, vol. 12 (5), pp. 509-523, 2003.
[3] V. Pizurica, A. Katartzis, J. Cornelis, H. Sahli, "What can be expected from computerised image analysis techniques for airborne minefield detection?", 2nd International Symposium 'Operationalization of Remote Sensing', Enschede, Nederland, August 16-20,pp. 399-406, 1999.
[4] L. van Kempen, A. Katartzis, V. Pizurica, J. Cornelis, H. Sahli, "Digital signal/image processing for mine detection. Part1: Airborne approach", MINE'99, Euroconference on Sensor systems and signal processing techniques applied to the detection of mines and unexploded ordnance, Firenze, Italy, October 1-3, pp. 48-53, 1999.
[5] A. Katartzis, H. Sahli, V. Pizurica, J. Cornelis, "A Model-Based Approach to the Automatic Extraction of Linear Features from Airborne Images", IEEE Trans. on Geoscience and Remote Sensing, vol. 39 (9), pp. 2073-2079, 2001.
[6] A. Katartzis, H. Sahli, E. Nyssen , J. Cornelis, "Detection of Buildings from a Single Airborne Image using a Markov Random Field Model", IEEE International Geoscience and Remote Sensing Symposium (IGARSS 2001), Sydney, Australia, vol. 6, pp. 2832-2834, 2001.
[7] A. Katartzis, I. Vanhamel, H. Sahli, "A Hierarchical Markovian Model for Multiscale Region-based Classification of Multispectral Images", IEEE Workshop on Advances in Techniques for Analysis of Remotely Sensed Data (WARDS 2003), NASA Goddard Visitor Center, Greenbelt Maryland, USA, 2003.