一种基于MDARNet的低照度图像增强方法

Low-light Image Enhancement Method Based on MDARNet

由于低照度环境下所采集的图像存在亮度低、对比度差、出现噪声和色彩失衡等低质问题,严重影响其在图像处理应用中的性能.为了提升低照度图像质量,以获得具有完整结构和细节且自然清晰的图像,结合Retinex理论与卷积神经网络,提出了一种基于MDARNet的低照度图像增强方法,并引入Attention机制模块和密集卷积模块以提升性能.首先,MDARNet利用同时包含二维和一维的3个不同尺度卷积核对图像进行初步特征提取,并用像素注意模块对多尺度特征图进行针对性学习;其次,设计跳跃连接结构对图像进行特征提取,使图像特征被最大限度地利用;最后,用通道注意模块和像素注意模块同时对提取到的特征图进行权重学习和照度估计.实验结果表明:MDARNet能够有效提升低照度图像的亮度、对比度、色彩等;且相较于一些经典算法,该方法在视觉效果及客观评价指标(PSNR,SSIM,MS-SSIM,MSE)能够得到更好的效果.

Due to the low-quality problems such as low brightness,poor contrast,noise,and color imbalance,the performance of the images collected in low-illumination environment is seriously affected in the process of image processing applications.The purpose of this paper is to improve the quality of low-illumination images to obtain natural and clear images with complete structure and details.Combining Retinex theory and convolutional neural network,this paper proposes a low-light image enhancement method based on MDARNet,which includes Attention mechanism module and dense convolution module to improve performance.Firstly,MDARNet uses three different scale convolution kernels that contain both two-dimensional kernels and one-dimensional kernels to perform preliminary feature extraction on the image,and the pixel attention module to perform targeted learning on multi-scale feature maps.Secondly,the skip connection structure is designed for feature extraction,so that the features of the image can achieve maximum utilization.Finally,the channel attention module and the pixel attention module are employed to perform weight learning and illumination estimation on the extracted feature maps simultaneously.The experimental results show that MDARNet can effectively improve the brightness,contrast,and color of low-light images.Compared with some classical algorithms,the MDARNet adopted in this thesis can achieve better results in visual effects and objective evaluation(PSNR,SSIM,MS-SSIM,MSE).