摘要
城市航空影像中 ,高大建筑物形成的阴影影响了影像视觉效果 ,也给物体识别造成了困难 ,有必要对其进行补偿和消除 .本文针对Retinex图像增强算法的不足 ,提出了一种模糊Retinex算法 .模糊Retinex使中心环绕空间对比运算仅在光照强度相近的区域中进行 ,增强了Retinex在光照强度变化较大的场景中的鲁棒性 .基于影像阴影区域、非阴影区域的模糊划分 ,该算法在保持原影像自然色彩的前提下补偿和消除了阴影的不良影响 .实验表明该方法取得了较好的效果 ,丰富了Retinex的研究和应用范围 .
In urban aerial images, the poor visibility of features in shadowed regions created by buildings prevents recognition of objects. It is necessary to compensate and remove the shadowed region. In this paper, a fuzzy Retinex is proposed to overcome the limitation of the well known image enhancement technique-Retinex. The fuzzy Retinex restricts the center/surround spatially contrast operation in region with homogeneous illumination so that the robustness of Retinex is reinforced in scene with strongly variational illumination. Based on the fuzzy segmentation of an image into shadowed and unshadowed regions, shadows are compensated and removed using the fuzzy Retinex while the natural tint of original image is retained. The proposed method enriches the research and application domain of the Retinex. Experimental results are given to show its effectiveness.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2005年第3期500-503,共4页
Acta Electronica Sinica
基金
国家自然基金项目 (No .60 1 72 0 66)
国家 863计划项目 (No .2 0 0 1AA1 360 70 )
关键词
航空影像
模糊Retinex阴影消除
Aerial photography
Calculations
Fuzzy sets
Image enhancement
Image segmentation
Object recognition
