Building layered detection method based on building shadow information mining

Abstract

The invention discloses a building layered detection method based on building shadow information mining. The method comprises the following steps: extracting buildings, including shadow detection according to the areas, distribution rule and morphological features of building shadows; extracting small shadows corresponding to low buildings and other shadows corresponding to medium to high buildings by using shadow area layering; performing small shadow processing corresponding to morphological algorithm low buildings to get an area of low dense building groups and independent low buildings; meanwhile, performing edge extraction on medium to high building shadows, and layering the medium to high buildings by using a linear extraction method to get medium building shadows and high building shadows; separating shadows and secret objects similar to shadows out of an image according to the grayscale statistical characteristic of the shadow image; then, judging the relationship between the shadows and the buildings; acquiring seed points of the low, medium and high buildings by using the relationship between the shadows and the buildings; and extracting the buildings from remote sensingimages by the way of regional growth.

Description

technical field [0001] The invention relates to the technical field of remote sensing image processing, in particular to a layered building detection method based on building shadow information mining. Background technique [0002] Cities are the product of the development of human civilization. Buildings in the city are the main places for human living and activities, and are closely related to human life. Therefore, building recognition is becoming more and more important in urban monitoring, urban planning, map updating and population estimation. In the past ten years, the ability to obtain remote sensing data has been greatly improved in terms of spatial and temporal resolution. High-resolution images provide us with the basis for studying the details of urban areas, especially buildings, but at the same time, they also bring us many problems such as limited spectral information, high heterogeneity, and shadows. Therefore, it is necessary to study the use of high-reso...

AI Technical Summary

Problems solved by technology

Therefore, compared with the 2D detection algorithm, the 3D detection has higher requirements on the data source, takes a long time, has high cost, and the data source is relatively limited, and has certain limitations on the detection range.

[0004] The existing two-dimensional building detection mainly has the following problems: (1) The detection result is mainly represented by a binary image, which only focuses on where the building is, and does not show the height characteristics of the building, which is not conducive to the analysis of urban scenes

(3) Some methods require relevant information such as the sun elevation angle when using shadows to assist in the detection of buildings

(4) In a large amount of existing data, there are buildings with large areas

Images