基于自适应Retinex算法的煤矿低光照图像增强研究

Study on Low-light Image Enhancement in Coal Mine Based on Adaptive Retinex Algorithm

针对煤矿井下图像由于光照不均以及粉尘等因素影响,造成采集图像亮度、对比度不足等问题,提出了一种基于自适应Retinex算法的煤矿低光照图像增强算法。首先将图像分解为光照图和反射图,并采用Unet网络提取多尺度图像特征,引入控制因子抵消多尺度带来的影响,有效提高了图像质量。然后将图像梯度作为调节因子,自适应调整多尺度算子,并引入重构损失和去噪模块抑制图像增强过程中产生的噪声。试验结果表明:所提算法在峰值信噪比和结构相似性方面优于部分现有算法,充分证明了该方法的有效性。

In order to solve the problem of insufficient brightness and contrast in coal mine images due to the influence of uneven illumination and dust,an adaptive Retinex algorithm based on low illumination image enhancement algorithm is proposed.Firstly,the image is decomposed into light map and reflection map,and Unet network is used to extract multi-scale image features,and control factors are introduced to offset the influence of multi-scale,which effectively improves the image quality.Then,the image gradient is used as the adjusting factor,the multi-scale operator is adjusted adaptively,and the reconstruction loss and denoising module are introduced to suppress the noise generated in the process of image enhancement.Experimental results show that the proposed algorithm performs better than the existing methods in terms of peak signal-to-noise ratio and mechanism similarity,which fully proves the effectiveness of the proposed method.