Detection of Anthropogenic Zones by Means of Spatial Anomaly Detection in Large-Scale Satellite Images
S. M. Borzov1, O. I. Potaturkin2
1 Institute of Automation and Electrometry, Siberian Branch, Russian Academy of Sciences
2 Institute of Automation and Electrometry, Siberian Branch, Russian Academy of Sciences Novosibirsk State University
borzov@iae.nsk.su
Keywords: remote sensing of the Earth, spatial features, image structure, spatial anomalies, object detection
Pages: 104-111
Abstract
A comparative study of the informativeness of statistical and structural features in anthropogenic zone detection in high-resolution satellite images by using decision rules formed on the basis of the Bayes and Neumann-Pearson criteria is performed. It is found that the greatest efficiency of man-made object detection on the natural background is reached by anomaly detection in the local structure of images with the use of the Harris response function.
|
