| 本文已被：浏览 885次 下载 721次 |
码上扫一扫！ |
| 基于PCA和BP神经网络的万寿菊黑斑病孢子识别 |
|
刘惠1, 冀荣华2, 祁力钧1,3, 马伟1, 高春花1
|
|
|
| (1.中国农业大学 工学院, 北京 100083;2.中国农业大学 信息与电气工程学院, 北京 100083;3.现代农业装备优化设计北京市重点实验室, 北京 100083) |
|
| 摘要: |
| 针对万寿菊黑斑病难于防治的问题,采用基于主成分分析和BP神经网络的识别方法,对万寿菊黑斑病病原菌(Alternaria tagetica)无侵染力和有侵染力的孢子进行精确识别。首先利用图像处理技术对病原菌孢子显微图像进行分割,选取3个颜色特征(R、G和V)、5个形状特征 (Hu不变矩中的H2、H3、H4、H5和H6),以及3个纹理特征(R、G、B 3个分量的对比度)共11个特征用于病原菌孢子分类识别。为提高识别速度和精度,利用主成分分析法(PCA)对11个特征进行优化和筛选,采用基于L-M算法的BP神经网络对万寿菊黑斑病病原菌的孢子进行分类识别。试验结果表明,经主成分分析后得到的第一、第二主成分能够有效减少BP网络训练时间和提高识别准确率,平均识别准确率达到98%。该方法能够精准识别万寿菊黑斑病病菌有侵染力和无侵染力的孢子。 |
| 关键词: 万寿菊黑斑病 孢子识别 图像处理 主成分分析 BP神经网络 |
| DOI：10.11841/j.issn.1007-4333.2015.06.34 |
| 投稿时间：2015-01-07 |
| 基金项目:农业部行业科技专项资助项目(201203025) |
|
| Spores of marigold black spot identification based on PCA and BP neural network |
|
LIU Hui1, JI Rong-hua2, QI Li-jun1,3, MA Wei1, GAO Chun-hua1
|
| (1.College of Engineering, China Agricultural University, Beijing 100083, China;2.College of Information and Electrical Engineering, China Agricultural University, Beijing 100083, China;3.Beijing Key Laboratory of Optimization Design of Modern Agricultural Equipment, Beijing 100083, China) |
| Abstract: |
| A method based on Principal Component Analysis (PCA) and BP neural network was developed to precisely classify noninfectious and infectious spores of the pathogen of marigold black spot(Alternaria tagetica).The micro spore images were segmented before feature analysis by image processing.Three color features (R、G、V),5 shape features (H2、H3、H4、H5、H6) and 3 texture features (R、G、B-components of contrast) were selected for the spot classification.In order to improve the speed and accuracy of recognition,the 11 features was optimized by principal component analysis and then BP neural network based on L-M algorithm was used to classify the spores.It was found that the first and second principal components could reduce BP nural network training time and increase classification accuracy effectively.The average correct classification rate reached 98%.The data proved that the proposed classification method could accurately classify noninfectious and infectious spores of the pathogen of marigold black spot. |
| Key words: marigold black spot spore identification principal component analysis BP neural network |
