A remote sensing detection and evaluation method for the area and production of large-area crop raising

Abstract

The invention provides a remote sensed estimating method of the large area planting area and the output, which establishes a spectrum which has the same growing stage and is correspondent with the field crown layer at the base of doing normal radiation and atmosphere adjustment to the remote sensed image, and computes the distance threshold value of the spectrum and compares the distance threshold value with the image initial wave spectrum and the reference wave spectrum to obtain TM plant unit, it then statistics the quantity of the TM plant unit from each image unit of middle resolution ratio imaging spectrum data MODIS with the same period and region to ascertain the planting number of the MODIS image, it then ascertains the planting area of the MODIS image according to each image unit area of the planting number and the MODIS.

Description

technical field [0001] The invention relates to a remote sensing detection and estimation method for a large-area single crop planting area and its output. Background technique [0002] At present, for the pure pixel recognition of a single crop, except for the small-scale field investigation method, most of them are supervised or unsupervised classification methods using remote sensing images. These methods either involve a large number of field work, or define mixed pixels It is very inaccurate, and the training area is also greatly affected by human factors when the supervised classification is selected, and the characteristics of the training sample area cannot be well grasped, so that it is difficult to meet the application requirements in the image conversion of different scales in remote sensing applications. Therefore, traditional classification data with low accuracy have to be used in large-scale crop area and yield estimation. Contents of the invention [0003]...

AI Technical Summary

Problems solved by technology

[0002] At present, for the pure pixel recognition of a single crop, except for the small-scale field investigation method, most of them are supervised or unsupervised classification methods using remote sensing images. These methods either involve a large number of field work, or define mixed pixels It is very inaccurate, and the training area is also greatly affected by human factors when the supervised classification is selected, and the characteristics of the training sample area cannot be well grasped, so that it is difficult to meet the application requirements in the image conversion of different scales in remote sensing applications. make it necessary to use less accurate traditional classification data in large-scale crop area and yield estimates

Images