Abstract: The haze image has the problems of low contrast, low illumination, color distortion and blurry edge, aiming at these problems, this paper proposes a haze image enhancement method based on human visual properties. By using the method of Bidimensional Empirical Mode Decomposition(BEMD), an image is decomposed, and multiple Intrinsic Mode Function(IMF) image layers with different resolution and a trend layer are obtained, then the layers are adjusted according to the proposed vision adaptability model to make it agree with human visual properties. Experimental results show that the enhanced image has an appropriate illumination and contrast, more comfortable to people’s eyes.

Key words: image enhancement, Bidimensional Empirical Mode Decomposition(BEMD), Intrinsic Mode Function(IMF), human visual properties, HSV space

摘要： 针对云雾图像对比度低、照度低、色彩失真、边缘模糊的问题，提出一种基于人眼视觉特性的云雾图像增强方法。采用二维经验模式分解方法，得到多个不同尺度的内蕴模式函数图像和一个趋势图像，根据视觉适应模型对分解所得的内蕴模式函数图像和趋势图像进行调整，将人眼视觉特性融入其中。实验结果表明，增强后的图像照度、对比度适中，更适合人眼观察。

关键词: 图像增强, 二维经验模式分解, 内蕴模式函数, 人眼视觉特性, HSV空间

CLC Number: 

• TP391.41