由于河流图像中常常存在光学噪音，且水污染分布呈现出复杂和多变的特征，使得特征获取难度较大。为此，提出一种城市周边流域河流有机污染物分布特征提取方法。采用自适应边缘相似度非局部均值(Adaptive Non-local Means Denoising with Edge Similarity，ANLM-ES)图像去噪方法，利用两个像素之间的高斯加权距离，获取复合图像块相似性权重函数，通过加权平均，经过计算确定中心像素，将图像去噪处理。分割城市周边流域河流图像，通过最小二乘法确定偏析线方向，采用拟合直线投影图像像素点，根据城市周边流域河流有机污染物分布特征，实现不同特征提取。通过仿真分析表明，所提方法去噪均方误差仅为 0。025，可以获取良好的城市周边流域河流有机污染物分布特征提取结果。
Extraction of Different Distribution Characteristics of Organic Pollutants in Surrounding Urban River Basins
Generally,there is optical noise in river images,and the distribution of water pollution presents complex and variable characteristics,making the feature acquisition more difficult.To address this,a method of extrac-ting the organic pollutant distribution feature in river basins around the city was presented.Firstly,a method of Adap-tive Non-local Means Denoising with Edge Similarity(ANLM-ES)was adopted,and then the Gaussian weighted dis-tance between two pixels was used to obtain a weight function of composite image block similarity.Through the weigh-ted mean and calculation,the central pixel was determined.Meanwhile,the image was denoised.Moreover,the image of the river basin around the city was segmented.Furthermore,the direction of the segregation line was determined by the least squares method.Finally,using the image pixel points projected by the fitting straight line,different features were extracted based on the distribution characteristics of organic pollutants in the river basin around the city.Simula-tion results show that the denoising mean square error is only 0.025,so the proposed method can get better extraction results of distribution features of organic pollutants.
NETL
NSTL
万方数据
