基于同态高低通滤波与多尺度Retinex的低照度彩色图像增强

针对低照度条件下彩色图像的细节丢失与视觉效果不佳等问题,提出一种改进的低照度彩色图像增强算法。对三个颜色分量图像分别进行同态高低通滤波处理,在提取高频成分的同时,保留部分低频成分。对其中高频图像进行多尺度Retinex算法、指数变换及量化等处理,加强图像细节;对低频图像进行量化和线性伸缩变换处理,避免图像过度增强。MATLAB仿真结果表明,该算法不仅改善了低照度图像的视觉效果,而且在信息熵、峰值信噪比和对比度等性能指标上也有较大提升。

ENHANCEMENT OF LOW ILLUMINATION COLOR IMAGE BASED ON HOMOMORPHIC HIGH-LOW PASS FILTERING AND MULTISCALE RETINEX

To solve the problem of detail loss and poor visual effect of color image under low illumination,an improved low illumination color images enhancement algorithm is proposed.The three color component images were respectively processed by homomorphic high and low-pass filtering,and some low-frequency components were retained while extracting high-frequency components.The high-frequency images were processed by multi-scale Retinex algorithm,exponential transformation and quantization to enhance the image details,while the low-frequency images were processed by quantization and linear scaling transformation to avoid excessive image enhancement.MATLAB simulation results show that our algorithm not only improves the visual effect of low illumination images,but also improves the performance indexes such as information entropy,peak signal-to-noise ratio and contrast.