图像在拍摄、传输和存储过程中常会出现退化情况，影响视觉感知和信息理解。图像复原增强旨在将降质图像恢复为干净图像，以提升视觉感知体验，并提高如语义分割和目标检测等计算机视觉任务的精度，在自动驾驶和智能医疗等数据敏感的应用场景有重要作用。近年来，视觉及多模态大模型在多个领域取得了重要进展，并在图像复原增强任务中展现出巨大潜力。对此，本文系统地总结并分析了近年来国内外图像复原增强领域应用视觉（大）模型和多模态大模型的重要研究进展。 1）总结介绍基于Vision Transformer（ViT）的图像复原增强方法，探讨ViT在处理图像退化和增强方面所具有的长距离依赖潜力； 2）详细阐述基于扩散模型的图像复原增强方法，讨论其在处理复杂图像退化和恢复细节方面的优势； 3）分析X-anything模型在图像复原增强任务上的潜力，尤其是Segment Anything Model（SAM）等视觉大模型在退化样本上提供的鲁棒零样本预测先验信息能力； 4）介绍多模态大模型如CLIP和GPT-4V在图像复原增强任务中的应用，展示了这些预训练模型在图像复原过程中所提供的语义信息指导能力； 5）分析当前图像复原增强技术面临的挑战，如数据获取困难、计算资源需求高和模型稳定性不足等，同时展望了图像复原增强技术的发展方向，为未来的研究和应用提供新的思路和参考。

Images, as essential carriers of visual information, play an integral role in various facets of human life, from daily interactions to complex technological applications. However, throughout the processes of acquisition, transmission, and storage, images are often exposed to numerous environmental and technical factors that lead to quality degradation. This degradation not only results in diminished visual perception and information loss but also has broader implications, adversely affecting computer vision tasks. When such quality degradation occurs, it can reduce the accuracy of critical computer vision applications, including semantic segmentation and object detection, which rely heavily on high-quality input images. In application scenarios where high precision and reliability are paramount, such as autonomous driving, intelligent healthcare, and other safety-critical environments, image degradation can significantly undermine the user experience and compromise the reliability of data-driven systems. To address these challenges, image restoration and enhancement technologies are designed with the goal of recovering degraded images to their original clarity and fidelity. These technologies aim to restore distortion-free images, thereby improving subjective visual quality and enhancing the performance of downstream tasks that depend on these images. Traditional image restoration techniques have shown some effectiveness in repairing images with mild degradation, but they often encounter difficulties when addressing complex or severe degradations, especially when multiple degradation factors are involved. This limitation has driven researchers to explore advanced methods capable of handling diverse and intricate degradation scenarios. In recent years, advancements in hardware computational power, coupled with rapid developments in deep learning, have led to significant breakthroughs in vision and multimodal large models. These models, powered by sophisticated architectures and extensive training, have demonstrated extraordinary potential across multiple fields. Leveraging these advances, image restoration and enhancement technologies have achieved notable progress, offering promising solutions to previously challenging problems. This paper provides a systematic review of the current research landscape in image restoration and enhancement, conducting an in-depth analysis of several core technologies driving advancements in this area. The primary contributions of this paper are structured around the following six focal areas: 1) Compilation and Analysis of Datasets for Image Restoration and Enhancement Tasks: The effectiveness of image restoration methods is greatly influenced by the quality and scale of datasets used for training and evaluation. This paper offers a comprehensive compilation of datasets commonly applied in image restoration tasks, such as denoising, deraining, and dehazing. We provide insights into the characteristics of these datasets, including their scale, quality, and the techniques employed to generate low-quality images, enabling a thorough understanding of dataset influences on restoration performance. 2) Exploration of Vision Transformer (ViT) in Image Restoration and Enhancement: Vision Transformers (ViT) have introduced the powerful Transformer architecture to the field of image processing. By enabling the processing of long-range dependencies, ViT has demonstrated considerable promise in image restoration and enhancement tasks. This paper systematically reviews the application of ViT in recent restoration tasks, discussing the advantages and limitations of ViT-based methods and evaluating its potential to manage complex image degradation patterns. 3) Summary of Diffusion Model-Based Image Restoration and Enhancement Methods: Diffusion models have emerged as effective solutions for handling complex image degradation and restoring fine details in challenging cases. This paper summarizes the recent advancements in diffusion model-based image restoration, focusing on the unique strengths of the iterative denoising process. Compared to traditional methods, diffusion models show strong capabilities in detail recovery for severely degraded images, though they also present risks related to generating content that may appear less realistic. 4) Analysis of the Potential of X-anything Models in Image Restoration and Enhancement Tasks: Represented by models such as the Segment Anything Model (SAM), X-anything models leverage extensive pre-training and prior information to achieve robust zero-shot predictions, even when applied to degraded images with limited labeling. This paper explores the application potential of SAM and similar models in image restoration, highlighting their ability to provide stable restoration capabilities through zero-shot learning, which could be highly advantageous in scenarios with unlabeled or weakly labeled data. 5) Application of Multimodal Large Models in Image Restoration and Enhancement: With the rise of multimodal large models like CLIP and GPT-4V, researchers have begun to leverage the powerful information fusion capabilities of these models for image restoration and enhancement. This paper demonstrates the advantages of multimodal models in complex restoration tasks by analyzing how they utilize pre-trained semantic information to guide restoration processes. The assistance of these semantic features allows multimodal models to achieve superior performance in challenging scenarios, where traditional methods may fall short. 6) Challenges and Prospects of Image Restoration and Enhancement Technologies: Despite the significant progress made in recent years, image restoration and enhancement technologies still face substantial challenges in practical applications. Key obstacles include difficulties in acquiring high-quality and diverse training data, high computational resource demands, and the need for enhanced model stability under various conditions. This paper discusses these challenges in depth and explores prospective research directions, such as improving model adaptability to resource constraints, developing more efficient data acquisition methods, and enhancing model robustness. These directions aim to provide valuable insights for both researchers and practical applications, fostering further development in the field. In conclusion, this paper aims to provide readers with a comprehensive overview of the research advancements in image restoration and enhancement over recent years, both domestically and internationally. By systematically summarizing the current progress and analyzing key technological innovations, this paper seeks to inspire new ideas and open up innovative directions for future research and applications in this rapidly evolving field.