基于小波的弱光照图像增强算法仿真  

Simulation of Weak Illumination Image Enhancement Algorithm Based on Wavelet

作　　者：李雪娇 宋建萍[1] 郑晓东[1] LI Xue-jiao;SONG Jian-ping;ZHENG Xiao-dong(College of Science and Technology of China Three Gorges University,Yichang Hubei 443000,China)

机构地区：[1]三峡大学科技学院,湖北宜昌443000

出　　处：《计算机仿真》2023年第3期210-213,223,共5页Computer Simulation

摘　　要：针对通过人工参数调整图像,容易出现细节纹理缺失或失真等问题,提出一种基于双尺度图像分解的弱光照图像增强算法。将小波变换函数与软阈值结合,通过小波包频带变换明确图像频带值,分解复数小波包,提取去噪处理后原始弱光照图像中的像素点,计算目标点邻域范围,并实施双尺度矩阵分割。利用同态滤波算法分析图像中像素点对弱光照的透射特性,在待处理图像上覆盖“雾”形成伪雾,分析去雾后图像透射点的分布情况,自动调整对应增强阈值。分析透射率与亮度分量之间的正向关联,得出图像透射率与亮度分量的变化规律,完成图像增强分析。仿真结果证明,所提算法鲁棒性强,在满足弱光照增强效果的同时还能保证图像的细节,增强后的图像细节更丰富,视觉观感更好。Adjusting the image by artificial parameters is likely to lead to the detail texture missing or textural distortion.In this article,an algorithm for enhancing weak-light images based on two-scale image decomposition was proposed.Firstly,we combined the wavelet transform function with the soft threshold,and used wavelet packet band transformation to determine the image frequency band value.Secondly,we decomposed complex wavelet packets and extracted pixels in denoising weak-light images.Thirdly,we calculated the neighborhood area of the target point and performed the two-scale matrix segmentation.After that,we used the homomorphic filtering algorithm to analyze the transmission characteristic of the pixel for weak light.Meanwhile,we covered the image with"fog"and thus formed a pseudo-fog.Based on the analysis result on the distribution of image transmission points after defogging,the corresponding enhancement threshold was automatically adjusted.Finally,we analyzed the positive correlation between transmissivity and brightness components and got the changing law of transmissivity and brightness components,thus completing the analysis of image enhancement.Simulation results show that the proposed algorithm has strong robustness.And this algorithm can ensure the details of a image while satisfying the enhancement effect of weak light.Moreover,the enhanced image has richer details and better visual perception.

关 键 词：弱光照图像 双尺度图像分解 小波变换函数 伪雾覆盖 同态滤波算法 

分 类 号：TP367[自动化与计算机技术—计算机系统结构]