A method for extract water body from urban river network area

Abstract

The invention discloses a method for extracting water body in urban river network area. There are a large number of building shadows in urban areas, some on water and some on land. Because of the similarity of shadow spectrum and water body, many shadow regions are often mistakenly extracted in water body extraction. The invention selects the water body located in the shadow area as the dark pixel, because the water body in the shadow area is shielded by the shadow, the water body in the shadow area is not irradiated by the direct sunlight, and the radiation of the water body out of the shadowarea can be considered as 0, and the aerosol parameter when the image is imaged can be calculated by the water body, thereby the atmospheric correction is carried out on all the water body pixels onthe image; Using the normalized water index after atmospheric correction and single-band threshold method, the shadow of many buildings in urban areas can be removed effectively and the water in urbanareas can be extracted quickly and accurately. The invention solves the problem that it is difficult to remove a large number of urban shadows in the extracted water body of the urban river network area.

Description

technical field [0001] The invention belongs to the technical field of remote sensing information extraction, and relates to a method for extracting water bodies in urban river network areas based on high-resolution remote sensing images. Background technique [0002] The urban river network area undertakes the functions of accumulating rainwater, regulating flood discharge, and beautifying waterfront landscapes. It is the foundation and support for social and economic development and ecological environment construction. Accurately extracting the location, area and corresponding change information of surface water bodies is a prerequisite for carrying out urban water resources investigation, water conservancy planning, environmental monitoring and protection, etc. [0003] Because satellite remote sensing data monitoring has the characteristics of wide range and strong real-time dynamics, remote sensing technology can greatly reduce the time cost and labor cost of obtaining ...

AI Technical Summary

Problems solved by technology

[0005] When the above method is applied to large-scale open water extraction, generally good results can be obtained. However, when the application scene becomes an urban river network area, the accuracy of water body extraction generally decreases.

This is because the shape of urban waters is uneven and the width of rivers is small, which requires high spatial resolution of remote sensing images, which highlights the following three problems: 1) High-resolution remote sensing images can clearly observe the urban high-rise The shadow of the building, and the shadow is easily confused with the spectral characteristics of the water body, so that a large number of shadow pixels are misjudged as water body pixels; 2) There are many optical types of water bodies in urban river network areas, such as clean water, eutrophic water, turbid water The spectral forms of water bodies, black and odorous water bodies, etc. are different, and it is easy to miss the extraction of water body pixels; 3) "restore" the actual spectral forms of ground objects through preprocessing processes such as image radiation calibration and atmospheric correction, which is the key to all remote sensing water body identification. However, at present, the design intention of wide-spectrum and high-resolution satellite sensors is mostly for land observation, which will inevitably reduce the accuracy of preprocessing results of water body pixels

Images