Way extracting method integrating object-oriented segmentation and grayscale morphology

Abstract

The invention discloses a way extracting method integrating object-oriented segmentation and the grayscale morphology. The method comprises the following steps that image fusion is performed on original data; object-oriented segmentation is performed on a fusion result; overall classification is performed according to an object obtained through segmentation, and classification comprises all levels of broad ways, urban construction regions and the like; segmentation is performed on an original panchromatic image by adopting a horizontal operator, and narrow-way extracting and noise reducing are performed according to the correlation technique like the grayscale morphology; all levels of broad ways are corrected by adopting the backbone based on the grayscale morphology to obtain a broad way network; non-narrow-way textures in the narrow way extracted result are removed by using an urban construction region mask to obtain a narrow way network; the way networks are combined to obtain a final classification way network extracted result. The method integrates the way extracting thoughts of object orienting, segment matching and ridge extracting, and is a multi-level way extracting method which can be suitable for broad and narrow ways and have higher accuracy and practicality by weakening or removing noise possibly existing in four kinds of way extracting.

Description

technical field [0001] The present invention relates to the field of remote sensing image information extraction, more specifically, relates to an object-oriented segmentation combined with gray-scale morphology road extraction method, which is a semi-automatic road extraction method for multi-level road extraction. Background technique [0002] As an important man-made object, road is the main body of modern transportation system and has important political, economic and military significance. Therefore, road has also become an important information carrier and processing object in maps and geographic information systems. In the mid-1970s, due to the need for digitized geographic traffic information, and the disadvantages of traditional manual road digitization, such as high cost and low efficiency, the image automatic and semi-automatic road extraction technology appeared and gradually developed. Nowadays, the emergence of multi-spectral and high-resolution satellites, ima...

AI Technical Summary

Problems solved by technology

As far as the various automatic and semi-automatic road extraction methods based on high-resolution satellite images in the field of remote sensing are concerned, there are three common problems: first, it is difficult to take into account the width and narrowness, and due to the single algorithm idea, it is often suitable for extracting wide road patches. The algorithm cannot extract narrow roads, and the algorithm suitable for extracting narrow roads at the pixel and sub-pixel level will have extremely strong distortion on wide roads; the second is that the extraction result is single, and the general extraction result is a simple mask. It can distinguish between roads and non-roads, but the roads are not classified internally, and the use value is limited; third, the research area of ​​the algorithm is too small, or the roads in the larger research area are too simple, and the resulting algorithm is generally not universal. It is difficult to avoid the serious impact of noise when expanding the scope of research or increasing the complexity of the road network

Images