Brief Review of Remote Sensing Literature Pertaining to Classification andGeostatistics

摘要

This first report summarizes the most relevant information for the project fromamong the wide range of literature on remote sensing. The facilities offered by IMAGINE have been examined. IMAGINE is a digital image processing package that can store, manipulate and analyse georeferenced data. It is the most widely used software for remote sensing applications. The other literature cited relates specifically to the aim of the project, i.e. to optimize field sampling from remote images using geostatistics. It includes applications of kriging and spatial filtering to classify remotely sensed date, and to enhance the image, the neural net approach, fuzzy classification, sampling images effectively and economically, the use of training data, confidence levels for classification, and the redundancy of information.