摘要
为了提高黄瓜叶部病害检测与染病程度的分级的准确率和效率,综合运用图像处理技术、人工神经网络技术,实现黄瓜叶部病害检测与染病程度分级,并主要对发病率高且危害严重的黄瓜霜霉病、白粉病和病毒病进行试验研究。首先,通过接种方法获得了纯正的黄瓜病害样本,并采集染病植株的样本图像。利用基本图像处理的方法对黄瓜叶部病害图像进行处理,综合运用二次分割、形态学滤波得到病斑区域。其次,提取三种特征包括22个特征参数,采用BP算法训练的多层前向人工神经网络对黄瓜病害进行分类。实验证明,检测系统的黄瓜叶部病害平均识别精度为95.31%,并能够快速准确地对黄瓜叶片病害的染病程度进行分级。
To improve accuracy of cucumber leaf disease detection and leave disease severity classification, Computer vision and artificial neural network technology were synthetically applied to automatically identify and classify cucumber leaf disease and the new approach is aimed at cucumber frosty mildew, powdery mildew and viral disease which were seriously harmed and at a high rate of incidence in this paper. First, the pure cucumber disease samples have been obtained and gathered. Then to process cucumber with leaf disease applying three methods and second time division method was adopted and morphology filter operations were used to eliminate noises and then disease spots were obtained. Second, characteristic analysis of cucumber leaf disease was carried and multilayer forward artificial neural network trained with BP algorithms was employed to classify cucumber leaf disease through three kinds of characteristics including 22 characteristic parameters. The experiments showed that the rate of testing precision was not less than 95.31%.
出处
《东北农业大学学报》
CAS
CSCD
北大核心
2012年第5期69-73,共5页
Journal of Northeast Agricultural University
基金
黑龙江省博士后基金(LBH-Z10249)
东北农业大学博士启动基金(2010)
关键词
图像处理
BP算法
黄瓜
病害
识别
image processing
BP algorithm
cucumber
disease
identification
