传统的三维重建技术在面对移动物体干扰时难以有效完成场景重建任务。针对该问题,本文提出一种基于SLAM(Simultaneous Localization and Mapping)、TSDF(Truncated Signed Distance Function)和SCNet(Sample Consistency Networks)实例...传统的三维重建技术在面对移动物体干扰时难以有效完成场景重建任务。针对该问题,本文提出一种基于SLAM(Simultaneous Localization and Mapping)、TSDF(Truncated Signed Distance Function)和SCNet(Sample Consistency Networks)实例分割网络的三维重建方法ORBTSDF-SCNet。该方法采用深度相机或双目相机获取重建物体及场景的深度图与RGB图,且基于ORB_SLAM2实时获取位姿信息;采用基于结构化点云数据的表面重建算法TSDF与深度图相结合,实现在线三维模型重建;为了消除场景中移动物体对场景三维重建的干扰,提出采用SCNet实例分割网络检测和分割移动物体,并结合优化策略减小检测和实例分割误差以及深度图和RGB图对准误差。通过抠除移动物体,保证了重建场景的完整性。在ICL-NUIM、TUM数据集上的实验表明了本文所提方法的有效性。展开更多
由于原始TSDF(Truncated Signed Distance Function,TSDF)模型仅考虑相邻时间上的关联,误差将不可避免的累积到下一时刻,无法构建全局一致的地图。为了实时精确的建立大场景稠密3D地图,对TSDF模型进行了改进。首先,构筑相机位姿模型和...由于原始TSDF(Truncated Signed Distance Function,TSDF)模型仅考虑相邻时间上的关联,误差将不可避免的累积到下一时刻,无法构建全局一致的地图。为了实时精确的建立大场景稠密3D地图,对TSDF模型进行了改进。首先,构筑相机位姿模型和加权融合3D点截断信息的TSDF模型,用于准确表示创建物体的表面。其次,提出一种改进的回环检测方法,并将其与随机蕨类彩色图像编码化相结合,进而优化TSDF模型,即混合优化位姿模型。最后,使用g2o图优化库解算约束函数,建立数据集间的优化边。实验结果表明:混合优化位姿模型能识别曾到达区域,特别在较大场景下使用可以得到更加准确的相机轨迹和地图。采用TUM数据集中的fr1/xyz、fr1/room、fr1/desk对所提算法进行检验,结果表明该方法能够使相机轨迹的均方根误差分别下降0.59cm,3.14cm,0.94cm。在室内环境和公开数据集上的实验结果证明了所提算法的有效性和准确性。展开更多
A geo-environmental investigation is carried out to identify the suitability for treatment,storage and disposal facility(TSDF) in the industrial area at Perundurai,Tamilnadu(India).State industries promotion corporati...A geo-environmental investigation is carried out to identify the suitability for treatment,storage and disposal facility(TSDF) in the industrial area at Perundurai,Tamilnadu(India).State industries promotion corporation of Tamilnadu(SIPCOT), Perundurai is a fast growing industrial centre therefore,needs a common utility i.e.TSDF site for safe management of the industrial wastes.展开更多
针对小型机器人在复杂环境中进行实时定位与建图时,存在机载端CPU(Central Processing Unit)计算资源不足,建图精度差、探索效率低的问题.本文提出一种基于同时定位与建图(Simultaneous Localization and Mapping,SLAM)、截断符号距离函...针对小型机器人在复杂环境中进行实时定位与建图时,存在机载端CPU(Central Processing Unit)计算资源不足,建图精度差、探索效率低的问题.本文提出一种基于同时定位与建图(Simultaneous Localization and Mapping,SLAM)、截断符号距离函数(Truncated Signed Distance Function,TSDF)实时的三维重建方法 .该方法基于深度相机或双目相机获取重建目标及场景的RGB(Red Green Blue)图和深度图,同时基于ORB_SLAM2获取位姿信息;采用基于特征点云数据的表面重建算法TSDF与深度图相结合,实现一种实时三维场景重建;为了降低三维重建模型与真实场景的误差,提出一种采用光线碰撞检测融合特征点的方法,并结合优化策略减小光线投影距离与体素到物体表面距离的误差.通过优化后的TSDF值,保证了重建场景的完整性.在ASL(Autonomous Systems Lab)开源数据集上,相比于Voxblox、Voxfield和VDBblox,该系统三维重建模型的均方根误差分别下降了55.6%、47.11%、21.7%,相比于Voxblox、Voxfield,系统地图更新速率分别提升了9.7%和12.9%.最后,将该系统用于室内场景实验,地图平均每帧更新速率为7.35 ms/帧,验证了所提系统的可行性和有效性.展开更多
口腔医学专业本科生第五年进入实习阶段,牙周临床实习内容与实验课相比有差距,实习质量有待提高。以案例为基础,利用浙江大学医学院附属口腔医院医学影像管理系统(picture archiving and communication system,PACS)查看X线片和口内照...口腔医学专业本科生第五年进入实习阶段,牙周临床实习内容与实验课相比有差距,实习质量有待提高。以案例为基础,利用浙江大学医学院附属口腔医院医学影像管理系统(picture archiving and communication system,PACS)查看X线片和口内照及牙周系统检查表电子化,在学生标准化病人(student as standardized patient,SSP)培训过程中予以辅助,保障SSP的使用。在牙周实习"讲授-示范-操作-反馈"(tell-showdo-feedback,TSDF)椅旁教学过程中,利用数字化辅助SSP,循序渐进、从简到难教学,并结合口腔数字示教系统教学超声洁治,逐步提高实习生病史采集能力、诊疗能力、医患沟通能力和超声洁治技能,保障实习生对常见牙周疾病的掌握,提高临床实践能力,为将来住院医师规范化培训打好基础。展开更多
为了解决室内动态环境下移动机器人的准确定位问题,提出了一种融合运动检测算法的半直接法RGB-D视觉SLAM(同时定位与地图创建)算法,它由运动检测、相机位姿估计、基于TSDF (truncated signed distance function)模型的稠密地图构建3个...为了解决室内动态环境下移动机器人的准确定位问题,提出了一种融合运动检测算法的半直接法RGB-D视觉SLAM(同时定位与地图创建)算法,它由运动检测、相机位姿估计、基于TSDF (truncated signed distance function)模型的稠密地图构建3个步骤组成.首先,通过最小化图像光度误差,利用稀疏图像对齐算法实现对相机位姿的初步估计.然后,使用视觉里程计的位姿估计对图像进行运动补偿,建立基于图像块实时更新的高斯模型,依据方差变化分割出图像中的运动物体,进而剔除投影在图像运动区域的局部地图点,通过最小化重投影误差对相机位姿进行进一步优化,提升相机位姿估计精度.最后,使用相机位姿和RGB-D相机图像信息构建TSDF稠密地图,利用图像运动检测结果和地图体素块的颜色变化,完成地图在动态环境下的实时更新.实验结果表明,在室内动态环境下,本文算法能够有效提高相机位姿估计精度,实现稠密地图的实时更新,在提升系统鲁棒性的同时也提升了环境重构的准确性.展开更多
As a universal pathogen leading to neonatal defects and transplant failure,human cytomegalovirus(HCMV)has strict species specificity and this has prevented the development of a suitable animal model for the pathogenes...As a universal pathogen leading to neonatal defects and transplant failure,human cytomegalovirus(HCMV)has strict species specificity and this has prevented the development of a suitable animal model for the pathogenesis study.The mechanism of cross-species barrier remains elusive and there are so far no non-human cell culture models that support HCMV replication.The Chinese tree shrew(Tupaia belangeri chinensis)is a small laboratory animal and evolutionary closely related with primates.We investigated the susceptibility of primary tree shrew dermis fibroblasts(TSDF)to HCMV infection.Infection with a GFP-expressing HCMV virus resulted in green fluorescence in infected cells with the expression of IE1,UL44 and pp28.The titers of cell-free viruses reached 10^3 PFU/mL at 96 hpi,compared to titers of 10^4 PFU/mL observed in primary human foreskin fibroblasts.Our results suggested that TSDF was semi-permissive for HCMV infection.The TSDF model could be further used to investigate key factors influencing cross-species multiplication of HCMV.展开更多
This paper presents a handheld 3D vision-based scanner for small objects by using Kinect. It is different from the previous color-glove-based approaches which require segmenting the target object. First, we eliminate ...This paper presents a handheld 3D vision-based scanner for small objects by using Kinect. It is different from the previous color-glove-based approaches which require segmenting the target object. First, we eliminate the noises and the outliers caused by holding hands. Second, we apply Kinect-fusion algorithm and truncated signed distance function （TSDF） to represent 3D surfaces. Third, we propose a modified integration strategy to eliminate the hand effect. Fourth, we take advantage of the parallel computation of GPUs for real-time operation. The major contributions of this paper are （1） the registration precision is improved, （2） the oflline amendment and loop closure operation are not required, and （3） concave 3D object reconstruction is feasible.展开更多
Reconstructing 3D models for single objects with complex backgrounds has wide applications like 3D printing,AR/VR,and so on.It is necessary to consider the tradeoff between capturing data at low cost and getting high-...Reconstructing 3D models for single objects with complex backgrounds has wide applications like 3D printing,AR/VR,and so on.It is necessary to consider the tradeoff between capturing data at low cost and getting high-quality reconstruction results.In this work,we propose a voxel-based modeling pipeline with sparse RGB-D images to effectively and efficiently reconstruct a single real object without the geometrical post-processing operation on background removal.First,referring to the idea of VisualHull,useless and inconsistent voxels of a targeted object are clipped.It helps focus on the target object and rectify the voxel projection information.Second,a modified TSDF calculation and voxel filling operations are proposed to alleviate the problem of depth missing in the depth images.They can improve TSDF value completeness for voxels on the surface of the object.After the mesh is generated by the MarchingCube,texture mapping is optimized with view selection,color optimization,and camera parameters fine-tuning.Experiments on Kinect capturing dataset,TUM public dataset,and virtual environment dataset validate the effectiveness and flexibility of our proposed pipeline.展开更多
文摘传统的三维重建技术在面对移动物体干扰时难以有效完成场景重建任务。针对该问题,本文提出一种基于SLAM(Simultaneous Localization and Mapping)、TSDF(Truncated Signed Distance Function)和SCNet(Sample Consistency Networks)实例分割网络的三维重建方法ORBTSDF-SCNet。该方法采用深度相机或双目相机获取重建物体及场景的深度图与RGB图,且基于ORB_SLAM2实时获取位姿信息;采用基于结构化点云数据的表面重建算法TSDF与深度图相结合,实现在线三维模型重建;为了消除场景中移动物体对场景三维重建的干扰,提出采用SCNet实例分割网络检测和分割移动物体,并结合优化策略减小检测和实例分割误差以及深度图和RGB图对准误差。通过抠除移动物体,保证了重建场景的完整性。在ICL-NUIM、TUM数据集上的实验表明了本文所提方法的有效性。
文摘由于原始TSDF(Truncated Signed Distance Function,TSDF)模型仅考虑相邻时间上的关联,误差将不可避免的累积到下一时刻,无法构建全局一致的地图。为了实时精确的建立大场景稠密3D地图,对TSDF模型进行了改进。首先,构筑相机位姿模型和加权融合3D点截断信息的TSDF模型,用于准确表示创建物体的表面。其次,提出一种改进的回环检测方法,并将其与随机蕨类彩色图像编码化相结合,进而优化TSDF模型,即混合优化位姿模型。最后,使用g2o图优化库解算约束函数,建立数据集间的优化边。实验结果表明:混合优化位姿模型能识别曾到达区域,特别在较大场景下使用可以得到更加准确的相机轨迹和地图。采用TUM数据集中的fr1/xyz、fr1/room、fr1/desk对所提算法进行检验,结果表明该方法能够使相机轨迹的均方根误差分别下降0.59cm,3.14cm,0.94cm。在室内环境和公开数据集上的实验结果证明了所提算法的有效性和准确性。
文摘A geo-environmental investigation is carried out to identify the suitability for treatment,storage and disposal facility(TSDF) in the industrial area at Perundurai,Tamilnadu(India).State industries promotion corporation of Tamilnadu(SIPCOT), Perundurai is a fast growing industrial centre therefore,needs a common utility i.e.TSDF site for safe management of the industrial wastes.
文摘针对小型机器人在复杂环境中进行实时定位与建图时,存在机载端CPU(Central Processing Unit)计算资源不足,建图精度差、探索效率低的问题.本文提出一种基于同时定位与建图(Simultaneous Localization and Mapping,SLAM)、截断符号距离函数(Truncated Signed Distance Function,TSDF)实时的三维重建方法 .该方法基于深度相机或双目相机获取重建目标及场景的RGB(Red Green Blue)图和深度图,同时基于ORB_SLAM2获取位姿信息;采用基于特征点云数据的表面重建算法TSDF与深度图相结合,实现一种实时三维场景重建;为了降低三维重建模型与真实场景的误差,提出一种采用光线碰撞检测融合特征点的方法,并结合优化策略减小光线投影距离与体素到物体表面距离的误差.通过优化后的TSDF值,保证了重建场景的完整性.在ASL(Autonomous Systems Lab)开源数据集上,相比于Voxblox、Voxfield和VDBblox,该系统三维重建模型的均方根误差分别下降了55.6%、47.11%、21.7%,相比于Voxblox、Voxfield,系统地图更新速率分别提升了9.7%和12.9%.最后,将该系统用于室内场景实验,地图平均每帧更新速率为7.35 ms/帧,验证了所提系统的可行性和有效性.
文摘口腔医学专业本科生第五年进入实习阶段,牙周临床实习内容与实验课相比有差距,实习质量有待提高。以案例为基础,利用浙江大学医学院附属口腔医院医学影像管理系统(picture archiving and communication system,PACS)查看X线片和口内照及牙周系统检查表电子化,在学生标准化病人(student as standardized patient,SSP)培训过程中予以辅助,保障SSP的使用。在牙周实习"讲授-示范-操作-反馈"(tell-showdo-feedback,TSDF)椅旁教学过程中,利用数字化辅助SSP,循序渐进、从简到难教学,并结合口腔数字示教系统教学超声洁治,逐步提高实习生病史采集能力、诊疗能力、医患沟通能力和超声洁治技能,保障实习生对常见牙周疾病的掌握,提高临床实践能力,为将来住院医师规范化培训打好基础。
文摘为了解决室内动态环境下移动机器人的准确定位问题,提出了一种融合运动检测算法的半直接法RGB-D视觉SLAM(同时定位与地图创建)算法,它由运动检测、相机位姿估计、基于TSDF (truncated signed distance function)模型的稠密地图构建3个步骤组成.首先,通过最小化图像光度误差,利用稀疏图像对齐算法实现对相机位姿的初步估计.然后,使用视觉里程计的位姿估计对图像进行运动补偿,建立基于图像块实时更新的高斯模型,依据方差变化分割出图像中的运动物体,进而剔除投影在图像运动区域的局部地图点,通过最小化重投影误差对相机位姿进行进一步优化,提升相机位姿估计精度.最后,使用相机位姿和RGB-D相机图像信息构建TSDF稠密地图,利用图像运动检测结果和地图体素块的颜色变化,完成地图在动态环境下的实时更新.实验结果表明,在室内动态环境下,本文算法能够有效提高相机位姿估计精度,实现稠密地图的实时更新,在提升系统鲁棒性的同时也提升了环境重构的准确性.
基金supported by the National Key Research and Development Project of China (2017YFC1309302)the Applied Basic Research Projects of Yunnan Province (2018FD034)the Yunnan Provincial Department of Education Science Research Fund Project (1405189906)
文摘As a universal pathogen leading to neonatal defects and transplant failure,human cytomegalovirus(HCMV)has strict species specificity and this has prevented the development of a suitable animal model for the pathogenesis study.The mechanism of cross-species barrier remains elusive and there are so far no non-human cell culture models that support HCMV replication.The Chinese tree shrew(Tupaia belangeri chinensis)is a small laboratory animal and evolutionary closely related with primates.We investigated the susceptibility of primary tree shrew dermis fibroblasts(TSDF)to HCMV infection.Infection with a GFP-expressing HCMV virus resulted in green fluorescence in infected cells with the expression of IE1,UL44 and pp28.The titers of cell-free viruses reached 10^3 PFU/mL at 96 hpi,compared to titers of 10^4 PFU/mL observed in primary human foreskin fibroblasts.Our results suggested that TSDF was semi-permissive for HCMV infection.The TSDF model could be further used to investigate key factors influencing cross-species multiplication of HCMV.
基金supported by the Ministry of Science and Technology of Taiwan under Grant No.MOST103-2221-E-468-006–MY1
文摘This paper presents a handheld 3D vision-based scanner for small objects by using Kinect. It is different from the previous color-glove-based approaches which require segmenting the target object. First, we eliminate the noises and the outliers caused by holding hands. Second, we apply Kinect-fusion algorithm and truncated signed distance function （TSDF） to represent 3D surfaces. Third, we propose a modified integration strategy to eliminate the hand effect. Fourth, we take advantage of the parallel computation of GPUs for real-time operation. The major contributions of this paper are （1） the registration precision is improved, （2） the oflline amendment and loop closure operation are not required, and （3） concave 3D object reconstruction is feasible.
基金supported by the Key Technological Innovation Projects of Hubei Province,China(No.2018AAA062)the National Natural Science Foundation of China(No.61972298)+1 种基金the Ministry of Education of Humanities and Social Sciences Project,China(No.17YJC760124)the Scientific Research Project of Department of Education of Hubei Province,China(No.B2021278).
文摘Reconstructing 3D models for single objects with complex backgrounds has wide applications like 3D printing,AR/VR,and so on.It is necessary to consider the tradeoff between capturing data at low cost and getting high-quality reconstruction results.In this work,we propose a voxel-based modeling pipeline with sparse RGB-D images to effectively and efficiently reconstruct a single real object without the geometrical post-processing operation on background removal.First,referring to the idea of VisualHull,useless and inconsistent voxels of a targeted object are clipped.It helps focus on the target object and rectify the voxel projection information.Second,a modified TSDF calculation and voxel filling operations are proposed to alleviate the problem of depth missing in the depth images.They can improve TSDF value completeness for voxels on the surface of the object.After the mesh is generated by the MarchingCube,texture mapping is optimized with view selection,color optimization,and camera parameters fine-tuning.Experiments on Kinect capturing dataset,TUM public dataset,and virtual environment dataset validate the effectiveness and flexibility of our proposed pipeline.
