Crop classification method based on deep learning and system thereof

Abstract

The invention provides a crop classification method based on deep learning and a system thereof. The method comprises the steps of dividing a to-be-classified operation area into a plurality of sub operation areas, and acquiring a multi-temporal multi-characteristic data set of each sub operation area; according to the multi-temporal multi-characteristic data set and crop sample data in the growthperiod of the to-be-classified crops in the to-be-classified operation area, acquiring a multi-temporal multi-characteristic data sequence of each pixel in a random sub operation area; according to the multi-temporal multi-characteristic data sequence of each pixel, acquiring the growth characteristic graph of each pixel; identifying the growth characteristic graph of each pixel through a trainedneural network model, and acquiring the classification result of the to-be-classified crops. According to the method and the system of the invention, a crop classification problem is converted to a problem of identifying a time sequence growth characteristic graph; through a deep learning method, the time sequence which is irregular in a main grain crop main output area scale can be normally usedin a normal-state data environment, thereby improving classification precision.

Description

technical field [0001] The present invention relates to the field of agricultural remote sensing, and more specifically, to a method and system for classifying crops based on deep learning. Background technique [0002] Crop classification based on remote sensing data is an important basic scientific issue in agricultural remote sensing. There are many methods for crop classification based on remote sensing data, and the design principles of these methods come from the following three aspects: 1) the difference in spectral reflectance of different crops; 2) the difference in image texture of different crops; Differences in growth characteristics of crops over time series. [0003] With the deepening of the research on the reflection mechanism of vegetation to solar radiation, researchers have created many indices with special sensitivity to different vegetation and ground features according to the differences in reflection and absorption of the spectrum by different vegetat...

AI Technical Summary

Problems solved by technology

The acquisition of texture information requires meter-level or even sub-meter-level resolution remote sensing data, and this ultra-high-resolution remote sensing data cannot cover regional or national scales due to its limited width.

Images