基于视觉侧抑制特性的自动色彩均衡算法

针对摄影测量和可见光遥感图像存在的曝光失衡、对比度低等缺陷及标准自动色彩均衡算法未充分顾及部分区域的色彩特征问题,本文基于模拟人眼视觉特性的非循环侧抑制模型和高斯分布函数,对标准自动色彩均衡算法进行改进。利用非循环侧抑制网络模型计算目标点的修正亮度值,由高斯分布函数选择高相关度的采样点,降低算法复杂度。使用标准自动色彩均衡算法和本文的改进自动色彩均衡算法对三幅存在不同色彩缺陷的图像进行对比实验,经本文算法处理的图像均值更合理且标准差和熵等重要参数均较原算法有所提升,其中标准差提高15%。实验结果表明,本算法具有较好的色彩均衡效果和的视觉效果。

Automatic Color Equalization Algorithm Based on Visual Lateral Inhibition Mechanism

For solving the problems of exposure imbalance and low contrast ratio in photogrammetry and visible light remote sensing, as well as the standard automatic color equalization algorithm, which does not fully take into account the color characteristics of some regions,this paper improves the standard automatic color equalization algorithm based on the non-circulating lateral inhibition model that simulate human visual characteristics and Gaussian distribution. The corrected luminance value of the target point is calculated by using the non-circulating lateral inhibition network model, and the Gaussian distribution is used to select the sampling points with higher correlation, which reduces the complexity of the algorithm. Three images with different color defects were compared, using standard automatic color equalization algorithm and improved ones. The experimental results show that the mean value of the images is more reasonable and the important parameters such as standard deviation and entropy are improved compared with the original algorithm. The standard deviation is about 15% improvement. The experimental results show that the improved algorithm has obvious color balance effect and visual effect.