基于多源波谱成像信息的茄科蔬菜灰霉病早期检测方法研究

项目名称： 基于多源波谱成像信息的茄科蔬菜灰霉病早期检测方法研究

项目编号： No.61075017

项目类型： 面上项目

立项/批准年度： 2011

项目学科： 金属学与金属工艺

项目作者： 冯雷

作者单位： 浙江大学

项目金额： 10万元

中文摘要： 灰霉病是一种严重威胁保护地蔬菜生产的真菌病害，不仅在植株生长期间严重发生，而且在产后的储藏、运输过程中也可继续造成严重危害。本研究提出以城乡居民主要消费蔬菜之一的番茄和茄子作为研究对象，以引致蔬菜灰霉病的半知菌亚门葡萄孢属作为病原物，用光谱技术分析蔬菜染病后的光谱反射特征，确定能反映植物病害的最优波长范围，并用多光谱图像、高光谱图像、热红外图像等多源波谱成像检测技术提取并理解植株染病后叶面、冠层及果实的图像特征信息，并与正常样本的测试结果进行对比试验，通过结合化学计量学和植物病理学等方法，实现染病植株和果实病症未出现情况下的早期检测，建立能准确反映植物灰霉病害的检测模型和病害程度的定量描述模型，形成对植物灰霉病进行早期、准确、非破坏性诊断的一种新理论和新方法。这对提高我国植物抗病机制的研究，控制保护地蔬菜灰霉病的危害，促进植物病理学发展具有重要的意义。

中文关键词： 植物保护；多光谱图像；高光谱图像；热红外图像；灰霉病

英文摘要： Botrytis cinerea Pers. is a ubiquitous fungus that is distributed worldwide. It is a severe threat to the productivity of greenhouse-grown plants. B. cinerea infects leaves, flowers, stems, and fruits by direct penetration or through wounds that are caused by cultivation practices or transportation. As there are some disadvantages and problems on the information requirement of naked eyes or molecule technique, this project studied on the detection of disease grade and early detection by using Multi-spectral imaging/Hyper-spectral imaging/Infrared thermal imaging technique. Visible and near-infrared reflectance spectroscopy was applied to the early detection of Botrytis cinerea on tomato and eggplant before symptoms appeared.After that,Multi-spectral image/Hyper-spectral image/Infrared thermal image technology were used to build the prediction image model of vegetable leave and fruit based on the spectral reflectance data. Through analyze the image of leaf and canopy in different stage of vegetable growing,a fast,reliable,and untouched detection and analyse method for detecting gray mold in the early stage during the plant growing season would be discovered.The infection pattern approach of gray mold compare with Multi-spectral image/Hyper-spectral image/Infrared thermal image also would be investigated in this work.

英文关键词： plant protection;Multi-spectral image;Hyper-spectral image;Infrared thermal image;Grey Mold