基于奇异值分解的降质图像视觉效果增强方法  

Visual Effect Enhancement Method of Degraded Image Based on Singular Value Decomposition

作　　者：万方[1] 雷光波[1] 徐丽[1] WAN Fang;LEI Guang-bo;XU Li(School of Computer,Hubei University of Technology,Wuhan Hubei 430068,China)

机构地区：[1]湖北工业大学计算机学院,湖北武汉430068

出　　处：《计算机仿真》2022年第10期219-223,共5页Computer Simulation

基　　金：湖北省教育厅2019年科学技术研究计划指导性项目(B2019049)。

摘　　要：针对图像质量低,细节信息丢失与视觉效果低下问题,提出一种基于奇异值分解的降质图像视觉效果增强方法,计算降质图像的透射率,依靠软抠图算法与导向滤波器,精细化透射率,恢复图像内景深突变处光晕效应。在图像频域中,反变换傅里叶系数至初始空间域内,增强图像对比度,同时对比像素调整明暗,使结果映射到合适视觉范围内,以优化图像的视觉亮度。最后按照奇异值矩阵的唯一性,刻画图像矩阵数据分布式特征,完成视觉效果的加强。实验证明,所提方法能够有效提高图像质量,边缘细节清晰可见,视觉亮度和效果都得到优化。Due to low image quality, loss of detail information and low visual effect, this paper puts forward a method to enhance the visual effect of degraded images based on singular value decomposition(SVD). Firstly, the transmissivity of degraded images was calculated, and then the soft matting algorithm and the guide filter were adopted to refine the transmissivity and restore the halation at the abrupt change position of the depth of field in images. In the image frequency domain, Fourier coefficients were inversely transformed into the initial spatial domain to enhance image contrast. In the meanwhile, pixels were compared and their brightness was also adjusted, so that the result was mapped onto a suitable visual range, thus optimizing the visual brightness of the image. According to the uniqueness of the singular value matrix, the distributed characteristics of data in the image matrix were described to complete the enhancement for visual effect. Experimental results show that the proposed method can improve image quality effectively and get clear and visible edge details. In addition, the visual brightness and effect are optimized.

关 键 词：奇异值分解 降质图像 视觉效果 亮度程度 透射率 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]