Remote sensing identification method and system suitable for crop types in landscape fragmentation area

Abstract

The invention discloses a remote sensing identification method and system suitable for crop types in a landscape fragmentation area, and the method comprises the steps: obtaining a remote sensing image of a target area according to the phenological characteristics of crops, and carrying out the data preprocessing; calculating the crushing degree of the target area based on the cultivated land data and partitioning; combining cultivated land data and crop distribution data obtained through SVM algorithm identification to construct a sample set; randomly selecting a training sample set and a test set from each fragmentation subarea, and inputting the training sample set and the test set as samples to a multi-feature deep learning model MFsNet comprising a detail feature extraction module, a semantic feature extraction module, a shallow feature jump operation module and a feature fusion module to complete the construction of a crop identification model; and inputting the to-be-identified target area image into the model to obtain a crop type identification result. According to the method, the training samples are selected from the partitions with different fragmentation degrees, so that the representativeness of the training samples is ensured, and the model training result can realize relatively high crop type identification precision in the landscape fragmentation area.

Description

technical field [0001] The invention belongs to the technical field of remote sensing image processing and crop type identification, in particular to a method and system suitable for remote sensing identification of crop types in landscape fragmentation areas. Background technique [0002] Based on the advantages of short detection period, large coverage, and strong current potential, remote sensing technology can support accurate and rapid identification of crop types. The process of distinguishing different crops and other types of ground objects. The extraction of remote sensing recognition and classification methods and features will directly affect the accuracy of crop recognition, requiring a large amount of manual intervention, such as manual feature extraction, which is difficult to achieve automatic classification; it is difficult to extract deep-level features, and the extracted features are all low-level A single or a small number of shallow features, resulting i...

AI Technical Summary

Problems solved by technology

[0007] In view of the above situation, the present invention proposes a method and system suitable for remote sensing identification of crop types in landscape fragmented areas, which can solve the shortcomings of traditional classification methods and the limitations of CNN models with simple structure and single module in land fragmented areas

Images