由于低照度图像的整体亮度比较暗、动态范围低、噪声大等特点，提出一种基于亮度传播图的低照度图像增强算法。考虑到低照度图像增强的同时也会放大噪声，因此在增强图像之前对图像进行去噪处理。使用BM3 D在YCbCr空间对图像进行去噪之后，在HSI空间对图像进行增强，利用亮度分量估计亮度传播图，利用物理模型还原低照度图像。实验表明该方法能够快速有效地提高低照度图像的整体亮度和对比度，增强图像的细节并减少噪声，得到视觉效果良好的图像。%Low illumination images always have lower overall brightness,low dynamic range and high noise.Because of these characteristics,we proposed an enhancement algorithm for low illumination image which is based on luminance transmission map.Considering that while enhancing the low illumination image the noise will be amplified as well,it needs to reduce image noise before image enhancement. We first used BM3D (block matching 3D)for image denoising in YCbCr space,then we enhanced the image in HSI space,and used luminance component to estimate the luminance transmission map,finally we used physical model to restore the low illumination image. Experiments showed that the method could rapidly and effectively improve the overall brightness and contrast of the low illumination image, enhance the detail of the image and reduce noise as well,and get an image with good visual effect.
展开▼
