CAJ | 학술논문

为了在无需人工干预的情况下自适应地处理多种不同特征的降质图像，有效恢复图中的细节信息，提出一种双通道局部处理的图像增强方法。首先在 HSV色彩空间内分别在亮、暗双通道对亮度分量进行局部分析，计算出与之相适应的增强函数，对图像进行增强；然后对双通道的增强结果进行混合高斯加权合并，得到增强后的亮度分量；最后分析增强前后亮度分布的差异，计算色彩补偿函数，矫正增强过程中引入的色彩失真。此外，还提出一种快速滑窗技术，以有效地降低运算的时间复杂度。实验结果表明，该方法能够灵活地处理如曝光不足、曝光过度、逆光以及雾霾影响等不同种类的图像，甚至是综合了以上多种特性的复杂图像，在处理效果和自适应能力上优势明显。
위료재무수인공간예적정황하자괄응지처리다충불동특정적강질도상，유효회복도중적세절신식，제출일충쌍통도국부처리적도상증강방법。수선재 HSV색채공간내분별재량、암쌍통도대량도분량진행국부분석，계산출여지상괄응적증강함수，대도상진행증강；연후대쌍통도적증강결과진행혼합고사가권합병，득도증강후적량도분량；최후분석증강전후량도분포적차이，계산색채보상함수，교정증강과정중인입적색채실진。차외，환제출일충쾌속활창기술，이유효지강저운산적시간복잡도。실험결과표명，해방법능구령활지처리여폭광불족、폭광과도、역광이급무매영향등불동충류적도상，심지시종합료이상다충특성적복잡도상，재처리효과화자괄응능력상우세명현。
In order to adaptively process various kinds of degraded images and effectively recover their de-tails without manual intervention, an image enhancement algorithm is proposed that locally processes im-ages in double channel. Firstly, in HSV color space, images are analyzed locally to calculate the appropriate enhancement functions in two channels, and then to be enhanced using these functions; then, enhancement results in two channels are weightedly merged using mixed Gaussian functions into enhanced intensity components; finally, a color compensation method is employed that considers intensity transformation to calculate compensation function and correct the color distortion resulting from the enhancement. Further-more, a fast sliding window technique is applied to significantly reduce the computational cost. Experimen-tal results show that the proposed algorithm can flexibly process images with different characteristics such as underexposed, overexposed, backlight, and misty images, as well as complex images of all the mentioned characteristics, the performance and adaptive ability are much better than other methods.