摘要
户外图像的雾化往往导致色彩降质和对比度降低,有效的去除图像中天气效果对提高视觉系统的可靠性和鲁棒性具有重要意义。根据大气散射物理模型以及有雾图像的色彩与对比度降质同场景点的景深成指数关系的先验知识,分别建立了户外图像全局去雾和局部去雾的能量最优化模型,推导了相应的求解偏微分方程。利用用户提供的简单附加信息,实现了仅从单幅图像恢复清晰图像的去雾算法。实验结果表明:去雾图像的色彩清晰度和对比度有较大改善,具有很好的应用前景。
Images of outdoor scenes captured in fog suffer from poor contrast and color. It is important to remove weather effects from the degraded image in order to make vision systems more reliable and more robust. It has the prior knowledge that degradation of color and contrast quality due to bad weather is exponential in the depths of the scene points. Based on the atmospheric scattering physics-based models, the global defogging model and the local defogging model were constructed in terms of energy optimization methodology, and the key partial differential equation was deduced accordingly. The algorithm of defogging was advanced using only a single image in terms of simple additional information provided interactively by the user. The experiment results demonstrate that the algorithms can effectively restore clear day color and contrast from weather degraded image and can be applied in practice.
出处
《系统仿真学报》
EI
CAS
CSCD
北大核心
2007年第16期3739-3744,3769,共7页
Journal of System Simulation