大数据处理警示性图像颜色纹理特征选取仿真

详细信息    查看全文 | 推荐本文 |

英文篇名：Big Data Processing Warning Image Color Texture Feature Selection Simulation 作者：张子栋 ; 张杰敏 ; 茅剑 英文作者：ZHANG Zi-dong;ZHANG Jie-min;MAO Jian;Computer Engineering College, Jimei University; 关键词：大数据处理 ; 警示性图像 ; 颜色纹理 ; 特征选取 英文关键词：Big data processing;;Warning image;;Color texture;;Feature selection 中文刊名：JSJZ 英文刊名：Computer Simulation 机构：集美大学计算机工程学院; 出版日期：2019-05-15 出版单位：计算机仿真 年：2019 期：v.36 基金：福建省自然科学基金(2017J01762);; 福建省教育厅(JA15278) 语种：中文; 页：JSJZ201905089 页数：5 CN：05 ISSN：11-3724/TP 分类号：440-443+476

摘要

警示性图像颜色纹理特征选取的有效性直接影响着多个领域的发展。针对当前方法存在的警示性图像颜色纹理特征选取准确率低的问题,提出了一种基于维纳滤波和增强矩阵的警示性图像颜色纹理特征选取方法,采用维纳滤波对警示性图像中的噪声进行消除,噪声消除后利用横向增量与纵向增量计算警示性图像中像素点的灰度值,求解每个像素点灰度值所对应的横纵坐标轴的二阶偏导数之和,根据这个偏导数推导出警示性图像的颜色纹理特征矩阵,利用该矩阵得到了一个警示性图像的颜色纹理特征模型,通过这个模型实现了对大数据处理下警示性图像颜色纹理特征的选取。仿真结果表明,所提方法对警示性图像的噪声具有很好的抑制效果,可以较好的保留警示性图像中的信息,并且所提方法能够准确的在大数据处理环境下对警示性图像颜色纹理特征的选取。
        Current feature selection method of color texture of warning image has many disadvantages, such as poor selection accuracy. In order to overcome the disadvantages, this research proposed a new feature selection method based on Wiener filtering and enhanced matrix. Noise of the warning image was eliminated via the Wiener filtering. After that, gray value of pixel in the warning image was calculated via horizontal increment and vertical increment. Sum of second partial derivative of abscissa axis and ordinate axis corresponding to the gray value of each pixel was solved. According to the partial derivative, Feature matrix of the color texture was deduced. A model of feature of the color texture was obtained via the matrix. The feature selection was achieved via the model. Simulation results show that the method has excellent inhibitory effect for noise of the warning image. It can reserve information of the image preferably and select the feature of color texture under big data processing accurately.

引文

[1] 张琴,等.混合智能优化算法的SAR图像特征选择[J].遥感学报,2016,20(1):73-79.
    [2] 肖志云,刘洪.小波域马铃薯典型虫害图像特征选择与识别[J].农业机械学报,2017,48(9):24-31.
    [3] 曲智国,等.图像特征点选取方法[J].信号处理,2017,33(8):1027-1033.
    [4] 李钦,等.图像深度层次特征提取算法[J].模式识别与人工智能,2017,30(2):127-136.
    [5] 孙继平,杨坤.一种煤岩图像特征提取与识别方法[J].工矿自动化,2017,43(5):1-5.
    [6] 袁春琦,等.顾及分类器参数的全极化SAR图像特征选择与分类[J].测绘科学技术学报,2016,33(5):507-512.
    [7] 郑春红,等.基于组合特征选择的目标识别算法研究[J].上海航天,2017,34(5):59-64.
    [8] 李响.多分辨率图像模糊特征快速抽取方法仿真[J].计算机仿真,2017,34(9):226-229.
    [9] 谢娟英,屈亚楠,王明钊.基于密度峰值的无监督特征选择算法[J].南京大学学报(自然科学),2016,52(4):735-745.
    [10] 魏国安.西安北动车段垃圾填埋场地基处理措施研究[J].地震工程学报,2017,39(5):939-945.