基于生成对抗网络的平面视觉图像色彩增强系统设计  

Design of plane vision image color enhancement system based on generative adversarial network

作　　者：窦红英 DOU Hongying(North University of China,Taiyuan 030000,China)

机构地区：[1]中北大学,山西太原030000

出　　处：《现代电子技术》2022年第23期66-70,共5页Modern Electronics Technique

基　　金：山西省文化和旅游厅2020年度规划课题(2020G055)。

摘　　要：平面图像在光线不足情况下容易出现色彩减弱问题,为了提升图像色彩视觉效果,设计了基于生成对抗网络的平面视觉图像色彩增强系统。使用CMOS传感器采集图像信息,并将其传输至色彩增强模块中。构建基于生成对抗的图像色彩增强网络,对图像色彩自适应调整,在灰度级局部化情况下利用函数映射公式对直方图均衡化处理。根据直方图非线性自适应特征,对不同情况下平面视觉图像亮度进行增强处理。根据不同区域像素间的关联性,增强局部对比度,完整存储图像边缘细节,通过判别公式计算对抗损失,实现扰动攻击性对抗处理。由实验结果可知,该系统差异色彩x坐标集中范围为5002~5006 mm,y坐标集中范围为3501~3503 mm,平面视觉亮度和清晰度均得到显著提高。Plane images are prone to color reduction in the case of insufficient light.In order to improve the visual effect of image color,a plane vision image color enhancement system based on generative adversarial network(GAN)is designed.The CMOS sensor is used to collect image information and transmit it to the color enhancement module.An image color enhancement network based on generative adversarial is built to adjust the image color adaptively.The function mapping formula is used to equalize the histogram in the case of gray level localization.According to nonlinear adaptive feature of the histogram,the brightness of plane vision image under different conditions is enhanced.According to the correlation between pixels in different regions,local contrast is enhanced,image edge details are stored completely,and the adversarial loss is calculated by the discrimination formula to achieve the disturbance attack adversarial processing.It can be seen in the experimental results that the concentration range of the difference color x coordinate of the system is within 5002~5006 mm,and the concentration range of the difference color y coordinate is within 3501~3503 mm.The brightness and clarity of the plane vision have been improved significantly.

关 键 词：生成对抗网络 平面视觉 图像色彩增强 图像亮度 局部对比度 色彩参数 

分 类 号：TN911.73-34[电子电信—通信与信息系统] TP274[电子电信—信息与通信工程]