Autors: Christoff, N. V.
Title: Feature extraction and classification using minimal curvature of 3D mesh for automatic crater detection
Keywords: Mars Orbiter Laser Altimeter, 3D mesh, Automatic craters detection, Machine learning

Abstract: In this paper the significance of tree classes of feature selection algorithms is examined. The features are extracted from 3D mesh data, generated from the Mars Orbiter Laser Altimeter (MOLA) for a classification task to automatically detect craters, while at the same time testing the performance of five classifiers. The key idea of this study is to examine the discriminative power of the original values, hereafter called “pure” values, of a minimal curvature by only converting them in the range of grey scale. The experimental results with five different classifiers show that better accuracy results are obtained over the features selected from the grey scale image. The employed technique from computer vision usually used for face detection, is applied in the task of crater detection.



    52nd International Scientific Conference on Information, Communication and Energy Systems and Technologies – Serbia, issue 1, pp. 44-47, 2017, Serbia, Publishing House, Technical University of Sofia, ISSN 2603-3259

    Вид: пленарен доклад в международен форум, публикация в реферирано издание, индексирана в Google Scholar