Method and device for inverting visible image of remote sensing image in bare place

Abstract

The invention discloses a method and a device for inverting a visible image of a remote sensing image in a bare place. The method comprises the following steps: preprocessing original remote sensing data to obtain a test image of a target geographic area and an RGB image of the target geographic area; performing pixel-level semantic segmentation on the test image of the target geographic area by using the first classification network model; performing region contour selection on the test image after semantic segmentation by using a second classification network model; and aligning and superposing the selected region contour to a corresponding position on the RGB image of the target geographic region. The device comprises a preprocessing unit, a semantic segmentation unit, a contour selection unit and a contour output unit. The method and device fully automatically draw the bare land inversion graph, save a large amount of manpower, have generalization, and can be made into an interface to be popularized to various cities. Full-vector diagram drawing is carried out, the definition of contours and numbers cannot be influenced by zooming in and zooming out of the drawn diagram, and the diagram is clearer than an Arcgis manually drawn diagram.

Description

technical field [0001] The present invention relates to the field of combined application of remote sensing images and deep learning technologies. More specifically, the present invention relates to a method and device for retrieving visual images from remote sensing images in bare land. Background technique [0002] Remote sensing is the imaging of the earth in a specific electromagnetic spectrum through the sensor on the satellite, and it is a technology developed on the basis of aerial photography technology. Through remote sensing detection, it is possible to conduct multi-level and multi-perspective observations on a large area of ​​the earth in a short period of time, which is an important means of obtaining environmental information and earth resources. [0003] Bare land refers to land whose surface is soil and basically no vegetation coverage or land whose surface is rocks and gravels and whose coverage area is greater than or equal to 70%. Bare land includes unuse...

AI Technical Summary

Problems solved by technology

Manually interpret and correct the bare land areas generated by deep learning, and use Arcgis and other labeling software to draw the generated bare land. The labor cost is too large, time-consuming and labor-intensive, it is difficult to promote, and it is not suitable for popularization

[0006] In the prior art, in the method for displaying the results of bare land inversion of remote sensing images, it is mainly to artificially use the Arcgis software tool to manually classify the obtained bare land categories, and to correct and draw through manual experience. The disadvantages of this method are: 1) Arcgis software can only process rgb band images. For multi-band data, other software such as ENVI needs to be used cross-use, and it has high hardware requirements, and it is easy to fill up the computer memory, causing the computer to freeze, which takes a long time and is inefficient; 2) only processing It takes about 4 days to invert a visual map of bare land in a prefecture-level city. Since bare land inversion maps are often used in bare land management, drawing all target cities requires a lot of human labor

Images