The invention discloses an autumn harvest main
crop information extraction method and
system based on
remote sensing data. According to the invention,
wave band merging is carried out on multi-temporal
remote sensing images; training samples are selected by combining phenological characteristics, spectral characteristics, texture characteristics and the like of all crops, a B-P neural
network method is adopted to supervise and classify
remote sensing data, classification results are classified and then processed, and
information extraction of main crops such as summer corn, cotton, peanuts, soybeans and middle-aged rice of all counties and the whole province in a research area is achieved. Through multi-temporal remote sensing image
wave band combination, under the condition that the number of training samples is reduced, the
crop classification precision is improved, the labor cost in the classification process is reduced, and the working efficiency is improved.