A methodology to isolate semi-automatically agrarian parcels from remotely sensed images and classifying their cropping systems and other land uses is described. This is achieved throughout CROPCLASS® software, which is written in IDL® and works as an “add-on” of ENVI®. Main steps are: a) parcels individualization, b) spectral band and vegetative indices calculation for each parcel; and c) cropping systems classification. We have validated this procedure using a series GeoEye-1 satellite images taken over Southern of Spain. The classification of cropping systems for each parcel was executed using CRT Decision Trees analysis. Traditionally, agricultural land use information is updated routinely in many cropland regions in USA and Europe through farmer communications or ground visit of administrative inspectors, which is tedious, time consuming, and therefore economically expensive. The patented CROPCLASS® procedure (in process) will get census agri-environmental administrative data through multitemporal remote images, therefore avoiding at large extend the traditional methodology.

Contenido: 1) Importancia de clasificar los usos de suelo en agricultura (cultivos) 2) Método de clasificación y de diagnóstico de cultivos 3) CROPCLASS-1.0 ® procedimiento & software (+CROPCLASS-2.0 en desarrollo) 4) Futuro y universalización del procedimiento 5) Proyectos, publicaciones, registro y patente

Trabajo presentado en la ESRI EUROPEAN USER CONFERENCE, celebrada en Madrid del 2 al 3 de octubre de 2013

Peer reviewed