| 基于颜色和纹理特征的大麦主要病害识别研究 |
| |
| 引用本文： | 杨倩,高晓阳,武季玲,李红岭,杨占峰,孔彦龙,毛红玉,寇敏瑜.基于颜色和纹理特征的大麦主要病害识别研究[J].中国农业大学学报,2013,18(5):129-135. |
| |
| 作者姓名： | 杨倩 高晓阳 武季玲 李红岭 杨占峰 孔彦龙 毛红玉 寇敏瑜 |
| |
| 作者单位： | 甘肃农业大学 工学院/甘肃省干旱生境作物学重点实验室, 兰州 730070;甘肃农业大学 工学院/甘肃省干旱生境作物学重点实验室, 兰州 730070;甘肃农业大学 生命科学技术学院, 兰州 730070;甘肃农业大学 工学院/甘肃省干旱生境作物学重点实验室, 兰州 730070;甘肃农业大学 工学院/甘肃省干旱生境作物学重点实验室, 兰州 730070;甘肃农业大学 工学院/甘肃省干旱生境作物学重点实验室, 兰州 730070;甘肃农业大学 工学院/甘肃省干旱生境作物学重点实验室, 兰州 730070;甘肃农业大学 工学院/甘肃省干旱生境作物学重点实验室, 兰州 730070 |
| |
| 基金项目： | 国家自然科学基金项目(61164001);甘肃省教育厅高等学校科研计划项目(1102-07);甘肃省干旱生境作物学重点实验室开放基金课题(1102-11) |
| |
| 摘 要： | 为提出一种在自然环境条件下基于采集图像的颜色一阶与二阶矩和纹理LBP算子改进模式综合特征参数的大麦病害识别方法,以甘肃河西地区发生的大麦白粉病、云纹病和条锈病为研究对象,采用颜色矩和LBP算子均匀模式综合特征参数来提取大麦病斑的颜色和纹理特征,并将该特征向量作为输入向量构建以径向基为核函数的支持向量机(SVM)分类器模型。利用SVM分类模型对采集到的355幅病害图像进行实例分析,结果表明当径向基参数时,大麦病害整体识别正确率达84.7458%。本研究为农田大麦病害诊断提供了有效的分析手段,验证分类模型在大麦病害研究中的可行性,并可为其他农作物病害诊断提供借鉴和参考。
|
| 关 键 词： | 大麦病害 颜色矩 纹理特征 LBP算子 支持向量机 |
| 收稿时间： | 2013/3/5 0:00:00 |
Identification of barley diseases based on texture color feature |
| |
| YANG Qian,GAO Xiao-yang,WU Ji-ling,LI Hong-ling,YANG Zhan-feng,KONG Yan-long,MAO Hong-yu and KOU Min-yu.Identification of barley diseases based on texture color feature[J].Journal of China Agricultural University,2013,18(5):129-135. |
| |
| Authors: | YANG Qian GAO Xiao-yang WU Ji-ling LI Hong-ling YANG Zhan-feng KONG Yan-long MAO Hong-yu and KOU Min-yu |
| |
| Institution: | College of Engineering/Gansu Provincial Key Laboratory of Aridland Crop Science, Gansu Agricultural University, Lanzhou 730070, China;College of Engineering/Gansu Provincial Key Laboratory of Aridland Crop Science, Gansu Agricultural University, Lanzhou 730070, China;College of Life Science and Technology, Gansu Agricultural University, Lanzhou 730070, China;College of Engineering/Gansu Provincial Key Laboratory of Aridland Crop Science, Gansu Agricultural University, Lanzhou 730070, China;College of Engineering/Gansu Provincial Key Laboratory of Aridland Crop Science, Gansu Agricultural University, Lanzhou 730070, China;College of Engineering/Gansu Provincial Key Laboratory of Aridland Crop Science, Gansu Agricultural University, Lanzhou 730070, China;College of Engineering/Gansu Provincial Key Laboratory of Aridland Crop Science, Gansu Agricultural University, Lanzhou 730070, China;College of Engineering/Gansu Provincial Key Laboratory of Aridland Crop Science, Gansu Agricultural University, Lanzhou 730070, China |
| |
| Abstract: | This study aims at presenting a method for identifying barley diseases based on the comprehensive characteristic parameters of first and second moment of color and with improving LBP model of texture feature under natural environmental conditions.Barley powdery mildew,moire disease and stripe rust occurred in Hexi corridor of Gansu province and used as the objective samples.Comprehensive characteristic parameters of color moments and LBP uniform model were used to extract color and texture features from diseased regions,which were chosen as inputs to the constructed classifier model in the base of radial basis kernel function for support vector machine (SVM) on barley disease recognition.The 355 collected disease images were analyzed by SVM classification model.The results showed that the overall identification accuracy of barley disease was up to 84.745 8% while RBF parameter.The system provided with an effective analysis means for barley disease diagnosis,verified the feasibility of classification model in barley diseases,and also could be used as a reference for the other crop disease diagnosis. |
| |
| Keywords: | barley diseases color moments texture feature LBP support vector machine |
|
| 点击此处可从《中国农业大学学报》浏览原始摘要信息 |
| 点击此处可从《中国农业大学学报》下载免费的PDF全文 |
|
