Abstract: Image dehazing technology for high-dust and foggy mine environments plays a pivotal role in ensuring safe production and automated management in coal mines. The operational environment in mines is unique; in deep shafts and areas with high temperature and humidity, mineral dust, particulate matter, and water vapor tend to accumulate continuously, creating a dense and non-uniformly distributed dusty and hazy atmosphere. This environment induces strong light scattering, leading to significant degradation in surveillance imagery, such as severely reduced visibility, low contrast, and blurred details. These issues directly impede vision-based applications like underground safety hazard identification, dynamic personnel tracking, and equipment status monitoring, thereby constraining the progress of informatization and intelligent mining. Therefore, research into image dehazing techniques for high-dust mine environments is of indispensable practical significance for safeguarding excavation safety and enhancing automated management efficiency. To provide an in-depth exploration of the development trajectory, current state, and future trends of image dehazing algorithms in such environments, this paper categorizes these algorithms into four main types based on their underlying principles: image enhancement, image restoration, fusion-based methods, and deep learning-based methods. It elaborates on the development history of each category, presents application examples in the context of high-dust mine imagery, and analyzes the advantages and disadvantages of classic algorithms. Through experiments, the paper comparatively analyzes and summarizes the performance of various dehazing algorithms on both natural and high-dust mine images. The results indicate that while CNN and Transformer models show leading performance on natural images, CNN models perform better in mine environments due to their stronger adaptability. The paper also summarizes the current challenges in mine image dehazing and discusses the future prospects of this technology. Key challenges include: the lack of dedicated datasets, interference from dynamic dust, fog, and noise, the difficulty of handling multi-dimensional degradation, the need for lightweight algorithms, and the absence of evaluation standards for safety-critical features. Future research should focus on developing more robust and adaptive dehazing algorithms. By considering the specific characteristics of the mining environment, this includes leveraging advanced technologies like sensor fusion and deep learning to improve dehazing effects, constructing realistic mine datasets, exploring unsupervised learning, integrating multi-source data, optimizing algorithms for real-time performance and lightweight deployment, and establishing safety-oriented evaluation systems to advance the intelligent upgrade of visual monitoring systems in mines.