西部尘霾天气下图像退化模型及快速去尘霾算法研究

本项目以西部尘霾天气条件下图像实时恢复为研究目的。在深入分析西部尘霾天气对图像成像影响的基础上，针对目前去尘霾算法普遍存在的边缘恢复效果不佳、处理速度比较慢等问题，对去除图像尘霾算法进行深入研究。通过精确构建图像退化模型，合理确定适合西部尘霾天气下图像梯度稀疏性的先验约束，快速估计每块图像的深度信息，对尘霾图像实现清晰复原，从而开发出快速去除尘霾图像算法。本项目研究拟取得的成果，不仅对于深化图像恢复技术研究具有一定的学术价值，而且对于促进计算机视觉和图像处理系统在相关领域的应用具有积极作用。

The research objective of this project is real time image restoration on the condition of dust haze weather in Western China. Based on detailed analysis to the influence of western dust haze weather on imaging, aiming at the common problems of present algorithms such as edge restoration in poor effect and slower processing speed, image dust haze removal algorithm will be in depth study. Through accurate construction of image degradation model and reasonable determination to the prior constraint of image gradient sparsity for western dust haze weather, depth information of each block of an image will be fastly estimated and clear restoration of dust haze image will be achieved. Thus a rapid removal algorithm of dust haze image will be developed. Achievements that this project intend to obtain, not only for the deepening of image restoration technology research has certain academic value, but also has a positive effect for the promotion of application of computer vision and image processing system in the relevant fields.