低照度环境下图像非均匀纹理增强算法

Image Non-uniform Texture Enhancement Algorithm in Low Illumination Environment

传统增强算法针对图像中非均匀纹理部分的处理过程中容易丢失纹理细节,导致增强后图像峰值信噪比与结构相似性的量化值偏低。为此,研究了一种针对低照度环境下的图像非均匀纹理增强算法。通过图像中非均匀纹理的合成情况,计算图像像素区域置信度和数据项,进而计算图像中非均匀纹理的增强优先权系数,确定增强先后顺序。建立基于卷积自编码的低照度图像增强模型,划分并融合增强图像照度,完成增强算法的设计。实验结果表明,该算法得到的增强图像峰值信噪比与结构相似性的量化值均高于传统算法,不同像素下的处理耗时总体优于传统算法。

Traditional enhancement algorithms are easy to lose texture details during the processing of non-uniform texture parts in the image,resulting in a low quantized value of the peak signal-to-noise ratio and structural similarity of the enhanced image.To this end,a non-uniform texture enhancement algorithm for images in low-illuminance environments is studied.Through the synthesis of the non-uniform texture in the image,the image pixel area confidence and data items are calculated,and then the enhancement priority coefficient of the non-uniform texture in the image is calculated,and the order of enhancement is determined.A low illumination image enhancement model based on convolutional self-coding is established,and the enhanced image illumination is divided and fused to complete the enhancement algorithm design.The experimental results show that the quantized value of the enhanced image peak signal-to-noise ratio and structural similarity obtained by the algorithm are higher than that of the traditional algorithm,and the processing time under different pixels is generally better than that of the traditional algorithm.