An intelligent understanding method of land use change based on multi-temporal remote sensing images

Abstract

The invention provides an intelligent understanding method of land use change, based on multi-temporal remote sensing images. The intelligent understanding method of land use change, based on multi-temporal remote sensing images includes the steps: detecting objects which are changed in the two-temporal remote sensing image target area and the content of the changes described by natural language,wherein detection of the objects which are changed in the two-temporal remote sensing image target area includes acquisition of the data of two-temporal, preprocessing of the remote sensing images, blocked cutting of the two-temporal data, construction of a mask layer, model training, model testing, detection of the content which is changed, and highlighting of the changed blocks; and the contentof the changes described by natural language includes determination of statement templates for describing the content of changes, and generation of description statement. The intelligent understandingmethod of land use change, based on multi-temporal remote sensing images respectively segments the two-temporal remote sensing images, uses the natural object class information to determine the change of the same position natural object class, is more accurate in change detection, is deterministic in the described content, describes the content of changes by the natural language, visually displays the content of changes of the two-temporal images, and is convenient for deep understanding of remote sensing images.

Description

technical field [0001] The invention relates to the technical field of computer vision and image semantic understanding, in particular to an intelligent understanding method of land use change based on multi-temporal remote sensing images. Background technique [0002] Remote sensing images are an important data source for the management and utilization of land resources. Quickly and automatically locate data with information differences from massive data, and dig out hidden information in remote sensing images, which can save a lot of manpower and material resources, and is of great significance to the effective use of land resources and scientific decision-making of relevant departments . Traditional remote sensing image change detection is usually based on spectral and other features to detect and judge whether the ground objects at the same location have changed. It is difficult to effectively use high-level information such as geometry, shape, and semantics in high-res...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide an intelligent understanding method of land use change based on multi-temporal remote sensing images to solve the problem that traditional remote sensing image change detection is difficult to effectively use high-level information such as geometry, shape, and semantics in high-resolution remote sensing images, and the detection accuracy is not high. High, what kind of changes in ground features can not be intuitively expressed in the form of knowledge technical issues

Images