Remote-sensing image semi-automatic water body extraction method

Abstract

The invention discloses a remote sensing image semi-automatic water body extraction method. The method comprises the following steps: modeling features of data; carrying out image segmentation by adopting a graph segmentation technology; judging the connectivity; carrying out vectorization; performing vector graphic operation; and carrying out semi-automatic water body extraction to achieve the effect of instant operation and instant preview. Through the method, the extraction operation of an operator is greatly reduced and simplified; the extraction result is ensured to be more accurate and fit the edge of the ground object; the purpose of optimizing the extraction result is achieved; positive and negative extraction in the water body region extraction process is achieved; the extractionefficiency of an operator is improved, the interaction is simple, the operation is easy, and the extraction is preview.

Description

technical field [0001] The invention relates to the technical field of remote sensing images, in particular to a semi-automatic water extraction method for remote sensing images. Background technique [0002] Water body information plays an important role in water conservancy planning, water resource investigation, environmental monitoring, and disaster assessment. The use of remote sensing images for information acquisition has the advantages of short cycle, strong adaptability, and saving manpower and material resources; with the popularization of remote sensing technology with high spectral resolution, high spatial resolution and high temporal resolution, for Rapid, accurate and timely extraction of water body information provides the basis and can gradually make up for the lack of field measurements. [0003] At present, the methods of water body extraction from remote sensing images are distinguished according to the degree of human participation, which can be divided ...

AI Technical Summary

Problems solved by technology

[0008] The water body index method calculates the water body index of the image, and then selects the appropriate threshold for manual interactive interpretation to segment the water body information; this method is difficult to solve the problem of a large number of small faces and small holes in the extraction results after threshold segmentation The threshold is also difficult to accurately divide the water body on the entire image;

[0009] The pixel-based region growth method, through manual interaction to select the appropriate seed point and set the appropriate growth threshold, using the region growth algorithm for water body extraction; The algorithm only considers the spectral difference information between pixels. For areas with complex boundaries and uneven distribution, multiple repeated extractions are required to obtain a suitable result.

[0010] The image segmentation method uses image segmentation to segment the extracted area once to obtain a series of segmented patch units, which can better solve the problem of image noise points, and then combine manual judgment to extract appropriate patches as water body information; The method depends heavily on the quality of the segmentation algorithm and the efficiency. At the same time, at the complex boundary, a certain amount of manual editing work is still required.

Images