High and cold mountainous area accumulated snow extraction method using high-resolution remote sensing image

Abstract

The invention relates to the technical field of remote sensing, and discloses a high and cold mountain area accumulated snow extraction method using a high-resolution remote sensing image. The methodcomprises the following steps of: S1) segmenting the high-resolution remote sensing image by using an optimal-scale fractal network evolution segmentation algorithm to obtain a high-resolution image object; S2) constructing a high-dimensional image object feature map by using the high-resolution image object; S3) constructing a deep belief network model; S4) constructing a full connection condition random field model; and S5) outputting an accumulated snow extraction result. The method is higher in applicability to different sensor data, and is more suitable for snow extraction of high-resolution remote sensing images. A deep belief network is combined with a full-connection condition random field model by using an object-oriented idea, spectrum, shape, texture and spatial relationship information of a high-resolution image is fully utilized, a salt and pepper phenomenon in an accumulated snow extraction result is eliminated, and the accumulated snow extraction precision is improved.

Description

technical field [0001] The invention relates to the technical field of remote sensing, in particular to a method for extracting snow in alpine mountainous areas by using high-resolution remote sensing images. Background technique [0002] Restricted by factors such as topography, geology and environmental protection, it is difficult to quantitatively obtain snow cover data in alpine mountainous areas by traditional methods of geography and meteorology. With the rapid development of remote sensing technology, new technology means are provided for snow cover monitoring. [0003] Currently MODIS data and its snow cover products are widely used in snow cover monitoring. "Study on the Method of Identifying Snow Cover in Qilian Mountains Using MODIS Data" (Wang Xing, Zhang Qiang, Guo Niobium, etc., "Arid Meteorology", Volume 25, No. 2, June 2007), reveals the use of MODIS images of The snow cover area was extracted using the Normalized Difference Snow Index (NDSI) index. "Study...

AI Technical Summary

Problems solved by technology

Furthermore, considering that using traditional pixel-based methods to extract snow areas on high-resolution images is prone to salt and pepper phenomena, and high-resolution images contain rich information such as texture, geometry, and spatial relationships, it is urgent to provide a high-resolution image of alpine mountain area. Snow extraction method to effectively eliminate the salt and pepper phenomenon in the snow extraction results, so as to obtain accurate and continuous snow areas

Images