Synthetic aperture radar image Voronoi polygon mosaic method considering multiple dimensions

Abstract

The invention discloses a synthetic aperture radar image Voronoi polygon mosaic method considering multiple dimensions, and the method comprises the steps: carrying out the quick mosaic of a preprocessed SAR digital orthoimage, and generating a mosaic base map image; utilizing an eight-neighborhood search algorithm to extract a boundary vector range of each preprocessed SAR digital orthoimage; on the basis of the extracted vector range, determining an overlapping region between the preprocessed SAR digital orthoimages; performing Voronoi complex polygon detection on each overlapping region, simplifying the detected complex polygon, and decomposing the complex polygon into a simplified SAR digital orthoimage; and embedding the simplified SAR digital orthoimage. According to the method, the mosaic base map image is firstly generated, and then the Voronoi seam line considering the multi-dimensional expansion overlapping surface and the linear weighted smooth mosaic processing are adopted, so that the mosaic efficiency and the mosaic effect can be greatly improved.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a method for mosaicing Voronoi polygons of synthetic aperture radar images in consideration of multi-dimensionality. Background technique [0002] Synthetic Aperture Radar (SAR) technology was born in 1951. It collects and records the electromagnetic wave radiation and scattered energy of the target and the background in the microwave band, uses signal processing technology to process and organize, compresses it into a SAR image, and extracts and recognizes the target object or phenomenon. It is widely used in surveying and mapping terrain features, detecting oceans, detecting earth resources, and reconnaissance and surveillance of military targets. [0003] SAR images are widely used in the production of digital orthophotos (DOM) and digital elevation models (DEM), becoming an indispensable basic remote sensing image product for natural resources, serving the monitoring...

AI Technical Summary

Problems solved by technology

The image seamline method of the conventional Voronoi diagram, for the image with a small degree of overlap, is prone to not being covered by any image, resulting in an invalid area in the middle of the mosaic image

Considering the image seamline generation algorithm of the overlapping surface Voronoi diagram, for images with a large degree of overlap, the seamline is not unique, resulting in cracks in the image after stitching

The twinsnakes method, due to the problem of local minimum or maximum in the snake algorithm, and the snake function is very sensitive to the variables that control the relationship between the vertices of the snake, causing the loss of the global minimum area and stopping at the local or global maximum.

[0006] Dijkstra's shortest path search method, because blind search is essentially an exhaustive search strategy, will cause the problem of path combination explosion, low search efficiency, and consume too much computing space and time

To sum up, the traditional methods are mostly based on small-scale, using data in the same direction for mosaic processing, resulting in loopholes or low efficiency. At the same time, data of different image sizes, different directions, and different times are not considered, resulting in a wide range of image applications. The application of a single image or a small-scale mosaic image does not take advantage of massive remote sensing image data, and the data utilization rate is low

Images