A method for detecting a salient object in an image based on dark channel prior and regional covariance

Abstract

The invention relates to a method for detecting a salient object in an image based on dark channel prior and regional covariance. Local and global estimation is mainly adopted to define the significance of each super-pixel. The accuracy of salient object detection is directly influenced by feature extraction, the feature extraction is a key step of converting visual stimulation into visual information for processing, and the saliency of each superpixel is calculated through two descriptors, namely a dark channel feature and a covariance feature. In order to optimize the saliency map, a map model is utilized to enhance the visual effect. The overall performance of the proposed saliency model is verified on a foggy image dataset in order to prove the robustness of the model and detect the advantage of a salient target in a low-contrast image. According to the invention, challenging foggy weather images can be well processed. This means that the model can more uniformly display the foreground object and can more fully suppress the foggy day background, which proves the superiority of the proposed model.

Description

technical field [0001] The invention relates to a method for detecting salient objects in foggy images through dark channel prior and region covariance description, belongs to the field of image processing, and can provide theoretical and technical foundations for hot issues such as security monitoring, rain, fog, and haze environment target positioning. Background technique [0002] According to the statistical report of the Ministry of Public Security, more than 10% of road traffic accidents are directly related to severe weather, such as smog, and the visibility of these scenes is significantly reduced. Optically, the reason is due to the fact that floating particles in the air absorb and scatter a lot of light. To solve this problem, a large number of dehazing algorithms have been developed in recent decades. By utilizing image defogging techniques, the color and visibility of foggy images can be restored to a certain extent. While dehazing can substantially increase t...

AI Technical Summary

Problems solved by technology

While dehazing can substantially increase the recognition of salient regions, the background information is also enhanced

However, the state-of-the-art saliency models still cannot correctly detect real salient objects in foggy images.

[0003] Usually, fog images are characterized by low contrast and low resolution, which makes it difficult for general systems to extract visual features

Saliency detection in foggy images faces several problems: 1) Traditional feature extraction methods cannot guarantee the accuracy of saliency results because most features become invalid under low visibility conditions

2) It is difficult to distinguish the difference between foreground and background in foggy images, which leads to the lack of edge and contour information in the extraction process

Images