High-resolution remote sensing image building extraction method and device based on U-shaped attention control network

Abstract

The invention discloses a high-resolution remote sensing image building extraction method based on a U-shaped attention control network. The method comprises the following steps: a remote sensing image is input to an encoder; the encoder extracts features of different grid dimension hierarchies, generates an encoder feature map and outputs the encoder feature map to the converter; an attention valve is connected in series in the jump connection between the convolution blocks of the encoder and the decoder; the converter extracts the abstract feature map and outputs the abstract feature map to the decoder; the decoder performs step-by-step up-sampling processing on the abstract feature map input by the converter, generates a decoder feature map and outputs the decoder feature map to the segmentation unit; and the segmentation unit adjusts the channel number of the input decoder feature map, and obtains a building segmentation result in the remote sensing image. In addition, the invention also discloses a high-resolution remote sensing image building extraction device based on the U-shaped attention control network. By adopting the U-shaped attention control network disclosed by the invention, the building extraction quality of the high-resolution remote sensing image is effectively improved in the aspects of prediction performance and result precision.

Description

technical field [0001] The invention relates to the technical field of remote sensing image extraction, in particular to a method and device for extracting buildings from high-resolution remote sensing images based on a U-shaped attention control network. Background technique [0002] Buildings, as real estate resources in urban and rural areas, are of great significance in regional construction planning, regional population estimation, economic development evaluation, topographic map production and update, etc. With the development of satellite sensor technology, the imaging quality of remote sensing images has been continuously improved, and researchers can obtain batches of high-resolution remote sensing images more quickly. However, as the volume of available remote sensing images continues to increase, how to automatically, accurately and effectively extract building information from images has gradually become a difficult problem. [0003] The amount of data in early ...

AI Technical Summary

Problems solved by technology

However, the CNN model alone cannot directly generate accurate building outlines, and the post-processing steps after image segmentation cannot be ignored. Therefore, in recent years, scholars have begun to try to introduce fully convolutional neural network (Fully Convolutional Networks, referred to as FCN) model. In the field of remote sensing image building extraction, the "end-to-end" extraction of building outlines is realized

Images