High-resolution image building area extraction method based on multi-scale texture spatial clustering

Abstract

The present invention discloses a high-resolution image building area extraction method based on multi-scale texture spatial clustering. The method comprises: performing multi-scale wavelet decomposition on an input image to obtain a low-frequency approximate coefficient image and a multi-directional high-frequency detail coefficient image; in each decomposition layer, performing feature integration on the multi-directional high-frequency detail coefficient image to obtain a texture intensity image; performing spatial autocorrelation statistical analysis to obtain a spatial distribution pattern image of each layer texture intensity; performing cross-scale feature fusion; and performing spatial clustering detection, and obtaining a building area through threshold segmentation. According tothe method provided by the present invention, that the texture features of the building area in the high-resolution image have the characteristics of multi-scale and spatial aggregation distribution are fully considered; and not only automatic extraction of the high-resolution image building area can be effectively implemented, but also complete contour boundaries can be maintained for the extracted building area, and further vectorization processing and application can be facilitated.

Description

technical field [0001] The invention belongs to the field of remote sensing image processing, and relates to a high-resolution image building area extraction method based on multi-scale texture space clustering. Background technique [0002] The built-up area is a typical artificial geographic element on the earth's surface, especially in the urban environment. It refers to the area covered by buildings. Timely and accurate acquisition of built-up area information is of great importance for land use planning, urban dynamic monitoring, and basic geographic database update. Value. [0003] In recent years, with the development of earth observation technology, automatic identification of built-up areas using high-resolution satellite images has received more and more attention. Due to the improvement of resolution, high-resolution images have the characteristics of large spectral variability, rich texture details, and complex spatial structures, which bring great challenges to...

AI Technical Summary

Problems solved by technology

However, there are still local spectrally homogeneous areas inside the built-up area (such as the open space between buildings, building roofs), and they will have lower frequency characteristics in the frequency domain, so it is not enough to be effective based on the magnitude of the frequency value Realize the discrimination between built-up area and non-built-up area

Images