High-resolution remote sensing image vegetation extraction method based on sensitive feature focusing perception

Abstract

The invention provides a high-resolution remote sensing image vegetation extraction method based on sensitive feature focusing perception, which is carried out on the basis of a weighted fusion probability mapping graph and comprises the following steps of: processing a satellite image by adopting a machine learning algorithm, and manufacturing a vegetation training label image; setting the size of an input image according to the semantic segmentation network, and making a training image and a label data set through cutting and segmentation; coupling an adaptive affinity field and an attention mechanism into the semantic segmentation network, constructing an adaptive context reasoning and vegetation sensitive feature focusing sensing module, and training the semantic segmentation network based on the training sample set to obtain a trained vegetation element extraction model; predicting the test set based on the trained network model to obtain a predicted plaque image; and based on the obtained predicted plaque image, carrying out splicing by using a weighted average strategy, recovering the resolution of the area before cutting, and eliminating the splicing seam effect. According to the method, end-to-end automatic extraction of different types of vegetation regions of the high-resolution image can be realized.

Description

technical field [0001] The invention belongs to the technical field of extracting vegetation elements from high-resolution remote sensing images, and is a method for adaptively extracting vegetation elements through a semantic segmentation network. Background technique [0002] Feature extraction is the process of identifying the type, nature, spatial location, shape, size and other attributes of the feature based on the features of the feature on the remote sensing image, and extracting the mark of the target feature. Feature extraction is an important part of land use / land cover classification and an important research direction in the field of remote sensing applications. With the emergence of domestic high-resolution remote sensing satellites such as ZY-3 and Gaofen series, high-resolution remote sensing images have become one of the main data sources for land use / land cover extraction due to their rich geometric and texture features. The acquisition of land and urban g...

AI Technical Summary

Problems solved by technology

[0006] The research team of the inventor proposed the paper "Vegetation Land Use / Land Cover Extraction from High-Resolution Satellite Images Based on Adaptive ContextInference" in 2020, and proposed a high-resolution variability remote sensing image vegetation extraction method based on adaptive context reasoning. However, the idea of ​​this method is relatively one-sided, and it fails to provide a technical solution that can be implemented to achieve the effect of improving the accuracy of vegetation element extraction.

Images