暗视觉环境下图像增强算法的研究

Research on Image Enhancement Algorithm in Scotopic Vision Environment

针对暗视觉环境下因光源不足等原因使得采集器采集到的图像噪声多、亮度低、边缘模糊等问题,提出了一种改进的多尺度视网膜和皮层(MSR)算法来增强图像原有的信息特征。利用同态滤波算法消减因不均匀光照对原图像的影响,增强图像灰暗区的细节纹理。将目标图像从蓝、绿、红(RGB)空间转换到色调、饱和度、亮度(HSV)空间,在明度通道上对图像进一步处理。传统算法复杂且容易滤除图像边缘信息,使用非局部均值滤波,较大程度上利用图像冗余信息并提高对图像的操作速度。结合Laplace算法,增强图像的边缘特征。仿真结果表明,采用改进算法能够有效提高暗视觉图像的整体图像质量,使得增强结果符合人眼视觉感受。

In the scotopic vision environment,due to the lack of light and other reasons,the image acquired by the collectos is characterized by more noises,low brightness,blurred edges and so on.An improved multi-scale retina and cortex(MSR)algorithm is proposed to enhance the original information features of the image.Firstly,the homomorphic filtering algorithm is used to diminish the effects of uneven illumination on the image and enhance the detailed texture of the gray zone.Secondly,the image is transformed from RGB to HSV color space,and it is further processed on luminance channel.The traditional method is so complicated that it is easy to filter the edge information of the image.Therefore,non-local means(NLM)is used to leverage the redundant information of the image to a great extent and improve the processing speed of the image.Finally,combined with the Laplace algorithm,the edge features of the image are also enhanced.The simulation results show that the proposed algorithm can effectively improve the overall quality of the scotopic vision image,making the enhancement results in line with human visual perception.