Combined spectral and spatial processing of ERTS imagery data

摘要: Abstract In addition to spectral features, texture is an important spatial feature used in identifying objects or regions of interest image. Although relatively easy for human observers recognize and describe empirical terms, it has been extremely refractory precise definition analysis by digital computers. This paper describes a procedure extracting some easily computable features the blocks image data illustrates applications combined textural (spatial) land use categories ERTS MSS (Earth Resources Technology Satellite Multi Spectral Scanner) data. The classification algorithm based on was developed tested using 614 64 × resolution cells derived from over Monterrey Bay area California coast line. applied training set 314 310 blocks. overall accuracy classifier found be 83.5% seven categories.