Method for processing remote sensing images by fusing elevation information

A method for processing remote sensing images by fusing elevation information, comprising the steps:

• acquiring an image data set, and preprocessing the image data set to obtain an image set;
• performing SLIC segmentation processing on the image set, introducing elevation information through weighting and setting a threshold in the SLIC segmentation processing process to obtain pre-segmentation results;
• a method for the SLIC segmentation processing comprising:
• establishing spectral feature vectors and three-dimensional space feature vector among pixel points in a preset range;
• counting spectral Euclidean distances and three-dimensional space Euclidean distances between each pixel and a cluster center, and allocating the pixels to the segmentation regions with similar features;
• performing multiple iterations to optimize the cluster center;
• traversing the whole domain to remove regions with excessively small sizes;
• performing region merging on the basis of the pre-segmentation results to complete image processing;
• a method for obtaining the pre-segmentation results comprising:
• introducing elevation information into distance calculation and setting a comprehensive threshold of spectral weighting combined with the elevation information during the SLIC segmentation process;
• adding elevation information on the basis of a two-dimensional plane and spectral Euclidean distances to establish six-dimensional feature vectors;
• and
• calculating the spectral Euclidean distances, the two-dimensional plane, and the elevation information for pixel i and pixel j within an initial superpixel range, expressed as respectively:
• d
• lab
• =
• (
• l
• i
• -
• l
• j
• )
• 2
• +
• (
• a
• i
• -
• a
• j
• )
• 2
• +
• (
• b
• i
• -
• b
• j
• )
• 2
• d
• xy
• =
• (
• x
• i
• -
• x
• j
• )
• 2
• +
• (
• y
• i
• -
• y
• j
• )
• 2
• d
• z
• =
• z
• i
• -
• z
• j
• in the formulas, l, a and b represent spectral band values of pixels, x and y represent row and column numbers of pixels, and z represents an elevation value corresponding to the pixel.