摘要
受环境、硬件设备及成本等因素限制,所生产的视频质量低,存在大量冗余噪声,导致局部区域图像模糊不清,无法对目标个体判定或识别。为解决上述问题,构建一种基于变分正则化的低质视频图像二维增强方法。通过采集先验信息,确定噪声去除和质量增强区域,随后使用ROF经典模型和偏微分对原始图像做平滑及去噪处理,并提出前后扩散方程,使方法能够有效抑制背景区域噪声,并对目标个体边缘做锐化处理,利用变分正则化提升视频图像整体分辨率,使其完成图像二维增强。仿真结果表明,所提方法具有控制噪声与加强图像质量双重优势,有效去除、抑制噪声影响,提升目标个体区域强散射点,且方法复杂程度较低,能够实现低质视频图像的均衡优化。
Traditional method was limited by environment, hardware equipment and cost, leading to low-quality video, massive redundant noise. In addition, the local area image was fuzzy, so it was unable to judge or identify the target. Therefore, a two-dimensional enhancement method for low-quality video images based on variation regularization was proposed. Firstly, the regions of noise removal and quality enhancement were determined by collecting prior information. After that, classical ROF model and partial differential were used to smooth the original image and reduce the noise. Then, the diffusion equation was proposed to effectively suppress the noise in background region. Moreover, the edge of target was sharpened and the resolution of video image was improved by variation regularization. Thus, we completed the two-dimensional enhancement for the image. Simulation results show that the proposed method has the advantages of controlling noise and enhancing image quality, which effectively removes and suppresses the noise. Meanwhile, this method improves the strong scattering point in target area. Due to low complexity, the proposed method is able to achieve the equalization optimization of low-quality video image.
作者
王雪红
WANG Xue-hong(School of Mathematics and Computer Science,Hezhou University,Hezhou Guangxi 542899,China)
出处
《计算机仿真》
北大核心
2020年第4期402-405,445,共5页
Computer Simulation
关键词
变分正则化
图像增强
低质视频
噪声去除
Variation regularization
Image enhancement
Low-quality video
Noise removal
