[Objectives] This study was conducted to improve the genetic character of Epipremnum aureum and provide a new way for genetic breeding of E. aureum. [Methods] In the paper,protoplasts were isolated from tender leaves ...[Objectives] This study was conducted to improve the genetic character of Epipremnum aureum and provide a new way for genetic breeding of E. aureum. [Methods] In the paper,protoplasts were isolated from tender leaves of E. aureum and Scindapsus aureus var. wilcoxii by using 1% cellulase and 1% peetinase. Then the protoplasts were purified and dynamically identified,and fusion of protoplasts of E. aureum and S. aureus was induced by PEG. [Results]A large quantity of protoplasts were obtained from tender leaves of E. aureum and S. aureus successfully,and the activity of protoplasts was over 90%. Fusion product was obtained by PEG induction,and the fusion rate was 16. 3%. [Conclusions]Protoplasts were isolated from tender leaves of E. aureum and S. aureus successfully,and a stable fusion system was built by PEG induction.展开更多
为了能够低成本、自动化批量获取植物叶片的外部表型参数,同时解决自然生长条件下的植物叶片存在遮挡而无法获取完整的外部表型数据的问题,该研究以绿萝叶片为研究对象,基于曲面参数方程建立叶片几何模型,提出一种基于特征分层的动态图(...为了能够低成本、自动化批量获取植物叶片的外部表型参数,同时解决自然生长条件下的植物叶片存在遮挡而无法获取完整的外部表型数据的问题,该研究以绿萝叶片为研究对象,基于曲面参数方程建立叶片几何模型,提出一种基于特征分层的动态图(Dynamic Graph CNN based on Feature Layering,FL-DGCNN)和堆栈编码器模型的绿萝叶片外部表型参数估测算法。通过多层组合的编码-解码器模型对残缺点云进行形状补全,将不同尺度下的点云通过多层感知机提取分组点不同层的特征向量融合后获取特征信息,以决定系数和均方根误差评价模型结果。结果表明:多层组合的编码模型对残缺点云补全的鲁棒性更高,特征分层的动态图模型估测结果的叶长、叶宽、叶面积的决定系数分别为0.92、0.93和0.94,叶长、叶宽的均方根误差分别为0.37、0.34cm,叶面积的均方根误差为3.01cm^(2)。该方法对叶类植物叶片的外部表型参数估测效果较好,具有实用性。展开更多
针对在自然场景中,由于遮挡、视角限制和操作不当等问题,导致传感器获取的植物或器官点云不完整,提出了一种基于多尺度特征提取模块结合点云金字塔解码器(Multi-scale feature extraction model with point cloud pyramid decoder,MSF-P...针对在自然场景中,由于遮挡、视角限制和操作不当等问题,导致传感器获取的植物或器官点云不完整,提出了一种基于多尺度特征提取模块结合点云金字塔解码器(Multi-scale feature extraction model with point cloud pyramid decoder,MSF-PPD)的叶片形状补全网络。首先,采用多尺度特征提取模块实现不同维度特征信息的全局提取和融合,其次,通过点云金字塔解码器进行叶片点云的多阶段生成补全,最终得到完整的目标叶片形状。使用曲面参数方程构建绿萝叶片仿真模型库,并将其离散成点云作为网络模型训练的训练集和验证集,使用Kinect v2相机获取绿萝叶片点云作为网络模型补全性能评估的测试集。试验结果表明,本文网络结构对叶片点云补全的效果理想,证明本文方法能够对遮挡情况下的绿萝叶片进行高效、完整的补全。展开更多
文摘[Objectives] This study was conducted to improve the genetic character of Epipremnum aureum and provide a new way for genetic breeding of E. aureum. [Methods] In the paper,protoplasts were isolated from tender leaves of E. aureum and Scindapsus aureus var. wilcoxii by using 1% cellulase and 1% peetinase. Then the protoplasts were purified and dynamically identified,and fusion of protoplasts of E. aureum and S. aureus was induced by PEG. [Results]A large quantity of protoplasts were obtained from tender leaves of E. aureum and S. aureus successfully,and the activity of protoplasts was over 90%. Fusion product was obtained by PEG induction,and the fusion rate was 16. 3%. [Conclusions]Protoplasts were isolated from tender leaves of E. aureum and S. aureus successfully,and a stable fusion system was built by PEG induction.
文摘为了能够低成本、自动化批量获取植物叶片的外部表型参数,同时解决自然生长条件下的植物叶片存在遮挡而无法获取完整的外部表型数据的问题,该研究以绿萝叶片为研究对象,基于曲面参数方程建立叶片几何模型,提出一种基于特征分层的动态图(Dynamic Graph CNN based on Feature Layering,FL-DGCNN)和堆栈编码器模型的绿萝叶片外部表型参数估测算法。通过多层组合的编码-解码器模型对残缺点云进行形状补全,将不同尺度下的点云通过多层感知机提取分组点不同层的特征向量融合后获取特征信息,以决定系数和均方根误差评价模型结果。结果表明:多层组合的编码模型对残缺点云补全的鲁棒性更高,特征分层的动态图模型估测结果的叶长、叶宽、叶面积的决定系数分别为0.92、0.93和0.94,叶长、叶宽的均方根误差分别为0.37、0.34cm,叶面积的均方根误差为3.01cm^(2)。该方法对叶类植物叶片的外部表型参数估测效果较好,具有实用性。
文摘针对在自然场景中,由于遮挡、视角限制和操作不当等问题,导致传感器获取的植物或器官点云不完整,提出了一种基于多尺度特征提取模块结合点云金字塔解码器(Multi-scale feature extraction model with point cloud pyramid decoder,MSF-PPD)的叶片形状补全网络。首先,采用多尺度特征提取模块实现不同维度特征信息的全局提取和融合,其次,通过点云金字塔解码器进行叶片点云的多阶段生成补全,最终得到完整的目标叶片形状。使用曲面参数方程构建绿萝叶片仿真模型库,并将其离散成点云作为网络模型训练的训练集和验证集,使用Kinect v2相机获取绿萝叶片点云作为网络模型补全性能评估的测试集。试验结果表明,本文网络结构对叶片点云补全的效果理想,证明本文方法能够对遮挡情况下的绿萝叶片进行高效、完整的补全。
