面向全变分图像复原的增广拉格朗日方法综述

图像复原旨在根据退化图像重建高品质原始图像,其复原的质量和速度问题一直都是图像处理领域研究的重要方向。由于其图像边缘保持特性,全变分(TV)最小化模型在图像复原领域取得了很大的成功。然而,全变分图像复原是一个典型的非光滑优化问题,需要发展相应的快速优化算法,而增广拉格朗日方法(ALM)则是近年来发展起来的一类代表性方法。结合相关进展,综述了全变分图像复原模型,变量分裂(VS)法和典型ALM算法,并通过实验从CPU运行时间、峰值信噪比(PSNR)和品质评价等方面分析了不同的变量分裂和ALM方法对图像复原性能的影响。     

大气光估计对单幅图像去雾复原的影响

对单幅图像进行去雾复原过程中,需要根据大气光进一步估计计算透射函数值,从而实现复原图像的求解。大气光估计过程中存在偏差,会直接影响透射函数的求解。针对这一问题,对户外图像去雾复原结果进行分析,通过综合比较不同大气光估计值对复原结果的影响,给出基于RGB条件对大气光进行估计的方法。实验结果表明,适当选取大气光估计的区域,能更加准确获得大气光的估计值,使得复原图像具有更好的视觉外观。     

总变分图像复原方程的离散化方法

基于总变分的图像重建和复原导致一类最小化问题,它归结为解一个非线性椭圆型偏微分方程。为了得到数值解,必须将问题线性化和离散化。C.R.Vogel和M.E.Oman等人(1996)的定点迭代是一个良好的线性化方法。然而文献中已报道的离散化方法需要微分方程数值解的工具,比较繁杂。该文提出一种新的离散化方法,它只需要图像处理中的常规技术。图像反降晰和噪声抑制的实验结果表明该文的结果不亚于文献中报道的结果。     

电信级分组传送网T-MPLS网络生存性技术

在介绍T—MPLS网络生存性的基础上,从路径保护和子网连接保护2个方面着重阐述了T—MPLS线性保护倒换机制,最后介绍了T—MPLS共享保护环和恢复及快速重路由技术。     

罐体顶升充砂法处理大型储罐倾斜工程实践

由于浮顶罐基础不均匀下沉,致使罐体倾斜,并导致油品泄漏。采用“罐体顶升充砂法”,较好地解决了这个问题。检修后,罐体的水平度、椭圆度偏差明显减小,使原储罐能继续使用。     

Assessing impaired benthic communities using sediment toxicity and contaminant concentrations from reference sites inside the Niagara River Area of Concern

Anthropogenically degraded benthic-macroinvertebrate communities (benthos) are one of seven beneficial use impairments (BUIs) in the Niagara River Area of Concern (AOC). Over the last 50 years, upgrades to waste-water treatment, industry closures, and sediment remediations reduced contaminant levels throughout the system. Improvements in benthic communities and sediment toxicity, however, were difficult to assess because there are no comparable reference reaches outside the AOC. A multi-phase study was initiated in 2015 to determine if data from inside the AOC could identify reference conditions, if toxicity and benthic-community data from these sites differed from other AOC sites, and if further remediation efforts were warranted in parts of the AOC. Concentrations or quotients of PAHs, PCBs, dioxins and furans, pesticides, and most metals were below their New York Sediment Class A Guidance Values at 10 sites that were subsequently designated as reference sites. Survival and growth data from Chironomus dilutus and Hyalella azteca bioassays indicated that sediments from only a few individual AOC-impact sites were toxic or significantly different from reference sites, and that mean toxicity at pooled AOC-impact and reference sites did not differ significantly. Similarly, New York Biological Assessment Profile scores and chironomid mentum deformity scores at only a few individual sites differed significantly from corresponding indices at reference sites, but neither metric differed significantly in comparisons between pooled AOC-impact and reference sites. Most analyses indicated that benthic communities were unimpaired and that removal criteria for the benthos BUI were largely met in much of the upper Niagara River AOC.