Abstract: In complex disaster-prone mine environments, the utilization of underground imaging enables a clear grasp of underground conditions and enhances rescue efficiency. This study, taking two-dimensional underground image processing technology as the object, introduces and analyzes several 2D image processing techniques and their research status specifically adapted for complex mine environments, and points out the existing problems in the field of mine rescue in China. Under the complex conditions of mining environments, existing image enhancement algorithms (e.g. Retinex, CLAHE) exhibit limitations including vignetting effect, color distortion, noise sensitivity, and insufficient dynamic adaptation capabilities; existing denoising techniques (such as spatial domain filtering, wavelet transform, and deep learning-based methods) demonstrate insufficient performance in suppressing mixed noise, exhibiting limitations including detail loss and low computational efficiency; the practical application of algorithms faces bottlenecks in data and models, including the lack of real-world datasets, high computational resource demands of deep learning models, and insufficient multi-task joint optimization frameworks. The future development of image processing in mine environments should focus on the following aspects. Developing enhancement algorithms adaptive to underground scenarios by integrating deep reinforcement learning with physical models to optimize multi-scale illumination separation and edge detail enhancement; constructing a multimodal noise suppression framework that combines self-supervised learning (such as Noise2Void: Learning Denoising from Single Noisy Images) with wavelet-neural network hybrid architectures to improve denoising efficiency and accuracy; promoting collaborative optimization of data-algorithm-hardware systems through multi-task learning technology to achieve end-to-end scene adaptation and distributed real-time processing. Through systematic analysis and forward-looking discussion of existing 2D image processing technologies in underground operations, this study aims to provide theoretical foundations and technical guidance for intelligent mine monitoring systems, ultimately driving innovation in visual surveillance technology and enhancing safety standards in mining production.