[Combining satellite data with computerized cadastral data to control crops in Pastriz and Quinto (Aragon)]

This work deals with the combined use of satellite and computerized cadastral data to identify crops in plots, and with the incorporation of this information into a Geographic Information System (GIS). GIS may yield graphic or numeric outputs for controlling wrong crop identification and fraud dissuasion, simplifying the field inspection work for the public administration. The study area is located at the Aragonese villages of Pastriz and Quinto, where computerized plot information is available. These villages exhibit a variety of agricultural systems, with some detection problems related to the very small size of many plots. Two images taken by the Thematic Mapper (TM) sensor of the Landsat-5 satellite has been used. One is from May 28th and the other from August 16th, 1994. Unitemporal and multitemporal analysis were performed using several supervised pixel classifications and several methods, such as plot postsegmentation, for improving those classifications. Pixel classifications and their improvement by using a filter were made with the image processor ERDAS. Vectorial plot information from the cadastre and raster satellite information were integrated through the ARC/INFO program for the plot postsegmentation. Several solutions for the combined use of raster and vectorial information have been suggested. The established procedures in this work may be applied to data proceeding from other commercially available satellites, which have higher spatial resolution power than Landsat TM. This will improve the classifications' accuracy and above all, will allow the assignment of uses to smaller plots. This work shows that an automatic and synoptic crop control in individual plots is feasible on a year basis. The described method will reduce the field inspection work, and will increase the reliability of detecting errors and/or frauds in crop declarations or in similar administrative documents.