The integration of spectral and textural information using neural networksfor land cover mapping in the Mediterranean

The aim of this study was to develop an efficient and accurate procedure fo r classifying Mediterranean land cover with remotely sensed data. Combinati ons of artificial neural networks (ANN) and texture analysis on a per-field basis were used to classify a Landsat Thematic Mapper image of the Cukurov a Deltas, Turkey, into eight land cover classes. This study integrated spec tral information with measures of texture, in the form of the variance and the variogram. The accuracy of the ANN was greater than that of maximum lik elihood (ML) when using spectral data alone and when using spectral and tex tural data. The use of texture measures through the per-pixel and per-field majority rule approaches were found to reduce classification accuracy beca use the held boundaries were enlarged and so overwhelmed the measures of te xture. In contrast, the per-held approach (where the field was specified pr ior to analysis) combined with texture information increased significantly classification accuracy. However, the accuracy decreased as the variogram l ag increased. The accuracy with which land cover could be classified in thi s region was maximised at 89% by using a per-held, ANN approach in which se mivariance at a lag of 1 pixel was incorporated as textural information. Th is is 15% greater than the accuracy achieved using a standard per-pixel ML classification. The primary limitation of the use of the per-held approach was noted to be the need for prior knowledge of field boundaries which may be resolved using existing data or through some form of edge-detection rout ine. (C) 2000 Elsevier Science Ltd. All rights reserved.