Object random walk-based visual saliency detection method and system for remote sensing image

Abstract

The invention discloses an object random walk-based visual saliency detection method and an object random walk-based visual saliency detection system for a remote sensing image. The method comprises the following steps: performing multi-scale segmentation, and combining adjacent regions with similar color characteristics under each scale respectively; for a division result under each scale, extracting visual characteristics of each divided region to construct of an object set under the current scale respectively; for the object set under each scale, calculating a corresponding edge weight by virtue of inter-object characteristic differences, and calculating the transfer probability of a focus of attraction between objects to obtain a transfer probability matrix of the focus of attraction, calculating the stable distribution of the focus of attraction among all the objects according to the transfer probability matrix of the focus of attraction respectively, further calculating visual saliency by virtue of the probability of each object in the stable distribution, and performing normalization to obtain a normalized visual saliency map under the current scale; fusing the visual saliency maps under each scale to obtain a final visual saliency map of the remote sensing image.

Description

technical field [0001] The invention relates to the technical field of image processing, and more specifically, to a method and system for detecting visual saliency of remote sensing images based on object random walk. Background technique [0002] As a major earth observation technology, remote sensing obtains high-resolution optical images that are the most intuitive and true portrayal of the spatial distribution of various objects on the earth's surface. Due to the huge surface area of ​​the earth and the various types and complex changes of ground cover, the high-resolution remote sensing images obtained present a large amount of data, a variety of content, and a complex structure. These characteristics make it time-consuming and difficult to obtain accurate feature descriptions of main objects or regions of interest in images when using computers to automatically process remote sensing images. At present, this phenomenon has become a bottleneck problem restricting the ...

AI Technical Summary

Problems solved by technology

However, the traditional visual attention model based on random walk still has certain limitations when extracting the salient area of ​​the image, mainly in two aspects: First, the traditional model uses pixels as the basic unit to calculate the salient value corresponding to each pixel , in the process of building the Markov chain, too many nodes are set, the transition probability matrix is ​​too large, and the calculation complexity is very high; secondly, the saliency graph calculated by the traditional model has been processed by Gaussian smoothing, and the edges of the saliency region are very Fuzzy, which is very unfavorable for extracting prominent ground objects in high-resolution remote sensing images

Images