摘要
The crop area estimaton is one of the main fields in application of remotesensing. The paper focuses on the operational method for rice planting areaestimation, in which TM datu is used to ertract base rice area in a given year of1992. The NOAA AVHRR data is used to prwhct the changing tendency of the nceplanting area. The base area data needs to be updated for every rice growth penodupon the availability of TM data. Three methods can be used to extract the base riceplanting area. They are (1) visual interpretation with interaedve adjustmant on thescreen, (2) iflteraCtive automatic classification with manual elinunating of the non-rice pixels on the screen, and (3) automatic dassification with GIS spatial analysis.These methods can be combined to increase reliability and accuracy. The currentpaper is only concemed with the description of the second method. MultitemporalNOAA AVHRR SAVI data are combined as multiband image and are classifiedusing supetwsed makimum likelihood classifier on ERDAS to prediCt the changingtendency of rice planting area. The method has been successfully used in extraCtingearly nce area in Hubei Province in 1994 and acceptable result was obtained.
The crop area estimaton is one of the main fields in application of remotesensing. The paper focuses on the operational method for rice planting areaestimation, in which TM datu is used to ertract base rice area in a given year of1992. The NOAA AVHRR data is used to prwhct the changing tendency of the nceplanting area. The base area data needs to be updated for every rice growth penodupon the availability of TM data. Three methods can be used to extract the base riceplanting area. They are (1) visual interpretation with interaedve adjustmant on thescreen, (2) iflteraCtive automatic classification with manual elinunating of the non-rice pixels on the screen, and (3) automatic dassification with GIS spatial analysis.These methods can be combined to increase reliability and accuracy. The currentpaper is only concemed with the description of the second method. MultitemporalNOAA AVHRR SAVI data are combined as multiband image and are classifiedusing supetwsed makimum likelihood classifier on ERDAS to prediCt the changingtendency of rice planting area. The method has been successfully used in extraCtingearly nce area in Hubei Province in 1994 and acceptable result was obtained.