A method and system for extracting building images from remote sensing images

Abstract

The invention discloses a method for extracting building images in remote sensing images, comprising: obtaining a convolutional neural network model; the convolutional neural network model is a trained model that takes remote sensing images as input and outputs building images A neural network model; obtaining a remote sensing image of an area to be collected; inputting the remote sensing image into the convolutional neural network model, extracting an image of a building in the area to be collected, and obtaining a preliminary extraction result; using a morphological method for the preliminary extraction result The learning closed operation is optimized to obtain the final extraction result of the building image in the area to be collected. And it also discloses a specific virtual system for realizing this method. The method and system for extracting building images in remote sensing images provided by the invention have the characteristics of high image extraction accuracy and high efficiency.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a method and system for extracting building images from remote sensing images. Background technique [0002] Building extraction from high-resolution remote sensing images is a key technology in the fields of urban change monitoring, urban planning, 3D modeling, and digital city establishment. With the continuous improvement of the resolution of remote sensing satellites, higher requirements are placed on the accuracy of building extraction from remote sensing images. [0003] The extraction of buildings from remote sensing images is to perform semantic segmentation on the ground objects in the image, that is, to obtain the information of building targets and edge contours by binary classification of each pixel. Because the extraction results intuitively and effectively reflect the location and distribution of building objects in the study area, it has become an importa...

AI Technical Summary

Problems solved by technology

In the high-resolution remote sensing images of urban areas, the surrounding environment of urban buildings is complex, and the traffic network is densely intersected, which may easily cause confusion and unclear edges between buildings and road edges in the extracted results.

In addition, the resolution of the forward propagation of the convolutional neural network continues to decrease, and the extraction results often have unsatisfactory edge effects, which limits the accuracy of building extraction.

Images