作者: Sunil Arya , Daniel DeMenthon , Peter Meer , Larry S Davis
DOI: 10.1016/0167-8655(91)90016-F
关键词:
摘要: Abstract This paper addresses the problem of texture discrimination in range images. The data is transformed to a common coordinate system, and then resampled generate regular grid points order eliminate effect different sampling rates. Eight statistical textural features based on co-occurrence matrices are computed data, used discriminate between two classes natural surfaces consisting pebbles sizes lying plane. Seven eight found be useful discriminating these classes. experiments also confirm importance resampling before computing features.