Analysis of Probabilistic multi-scale fractional order fusion-based de-hazing algorithm

TL;DR

This paper introduces a probabilistic multi-scale fractional order fusion de-hazing algorithm that enhances image contrast and brightness while avoiding sky over-enhancement, outperforming existing methods.

Contribution

It proposes a novel fusion-based de-hazing method that improves local contrast, edge sharpness, and brightness without over-enhancing sky regions.

Findings

Better de-hazing performance in most cases

Enhanced local contrast and edge sharpness

Effective brightness improvement without sky over-enhancement

Abstract

In this report, a de-hazing algorithm based on probability and multi-scale fractional order-based fusion is proposed. The proposed scheme improves on a previously implemented multiscale fraction order-based fusion by augmenting its local contrast and edge sharpening features. It also brightens de-hazed images, while avoiding sky region over-enhancement. The results of the proposed algorithm are analyzed and compared with existing methods from the literature and indicate better performance in most cases.