速生阔叶树苗的缺水识别与三维可视化

项目名称： 速生阔叶树苗的缺水识别与三维可视化

项目编号： No.31300471

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 农业科学

项目作者： 胡春华

作者单位： 南京林业大学

项目金额： 23万元

中文摘要： 本课题主要采用计算机视觉技术对育苗期的多种速生阔叶树缺水进行识别研究。人工培育多种速生阔叶树苗，人为控制供水量。采用CCD摄像机对培育期的正常与缺水速生阔叶树叶进行样本采集，并采用深度摄像机对相应叶片的空间尺寸进行测量，建立样本库；对采集的样本图片首先采用基于直觉模糊集叶图像分割法分割出叶片样本，然后采用各向异性核扩散流形学习降维方法进行特征提取，最后应用SVM（Support Vector Machine）算法建立多种速生阔叶树苗缺水识别分类器；根据叶片空间位置信息、纹理特征、颜色特征以及形态特征对相应的叶片进行三维可视化重构，构建多类阔叶树苗缺水判别可视化界面。

中文关键词： 速生阔叶苗；缺水判别；三维可视化；自动灌溉；

英文摘要： In this study, an interface for discriminating water shortage of fast-growing hardwood seedlings is constructed. Several kinds of fast-growing hardwood seedlings are planted and water is controlled. To establish the samples library, CCD camera is utilized to take the normal growth and water shortage fast-growing hardwood seedlings leaves, and the depth camera is used to collect the corresponding leaves space measurement.An image segmentation method based on intuitionistic fuzzy set is first proposed to obtain the samples leaves, then anisotropic diffusion map manifold learning method is used for leaves features extraction, and then SVM (Support Vector Machine) is adopted to construct recognition classifier. According to the leaves space sizes, texture features, color characteristics and morphological characteristics of the corresponding leaves, 3D visualization reconstruction interface is constructed, which can discriminate the water deficiency of fast-growing hardwood seedlings on line.

英文关键词： fast growing hardwoods seedlings；water shortage identification；3D visualization；automatic irrigation；