We present a method of 3D image mosaicing for real 3D representation of roadside buildings, and implement a Web-based interactive visualization environment for the 3D video mosaics created by 3D image mosaicing. The 3...We present a method of 3D image mosaicing for real 3D representation of roadside buildings, and implement a Web-based interactive visualization environment for the 3D video mosaics created by 3D image mosaicing. The 3D image mo- saicing technique developed in our previous work is a very powerful method for creating textured 3D-GIS data without excessive data processing like the laser or stereo system. For the Web-based open access to the 3D video mosaics, we build an interactive visualization environment using X3D, the emerging standard of Web 3D. We conduct the data preprocessing for 3D video mosaics and the X3D modeling for textured 3D data. The data preprocessing includes the conversion of each frame of 3D video mosaics into concatenated image files that can be hyperlinked on the Web. The X3D modeling handles the representation of concatenated images using necessary X3D nodes. By employing X3D as the data format for 3D image mosaics, the real 3D representation of roadside buildings is extended to the Web and mobile service systems.展开更多
This paper presents a full view video navigation platform and the approach to buildup such a platform.Unlike the traditional full view navigation system,which always bases on image panorama,our system is based on full...This paper presents a full view video navigation platform and the approach to buildup such a platform.Unlike the traditional full view navigation system,which always bases on image panorama,our system is based on full view video created by video mosaics.Neither the current concept of video mosaics,which means posting together video frames to get a wide view,while these frames are from the same video.Video mosaics in this paper mean stitching different videos together to get a full view video.Videos,together recording 360° of the scenery,are captured through a certain route by digital cameras arranged in a circle around the view point.Then break up these videos into frame sequence.Frame coming from different videos with the same timestamp are stitched into a panoramic image.Consider these panoramic images as frames;montage these panoramic frames into a full view video,by sequence of time.While displaying the full view video people could decide to go forward,or stop to see around,or proceed in a certain direction.While talking about full view navigation,we are mostly talking about image-based cylindrical or spherical panorama [2].One way to create such a panorama is,after a series of translation,stitching together a group of images taken by panning camera [1].Each of these images presents part of the scene and with certain degrees of overlap to adjacent ones.After stitching,we get a panorama image with width is much bigger than the raw images,map it onto a cylinder or sphere surface is a better way to show it.The mapped one is called cylindrical panorama or spherical panorama.While navigating,taking sphere panorama as an example,people were supposed to standing at the center of the sphere,by rotating the sphere,different view coming into eyes.This kind of navigation makes people feel as the same as standing in the real world.But one limitation is that the step is stuck to one point,could see around but not move forward or back(there is moving forward and back by zooming in and zooming out respectively,but it's not truly move).In our work,videos are used as the stitching source instead of images.Video,or could say a sequence of images taken at continuous view point(actually it is the definition of 'video',and is the saying we are going to take advantage of in this paper),are captured by a camera circle,and are stitched into a full view video.Frames coming from different video with the same timestamp are stitched into a panorama frame.Transform the processed panorama frames into a video.展开更多
针对大场景视频拼接技术在汽车环视系统等领域的应用需求,并为了适应嵌入式系统快速发展的要求,提出了基于FPGA片上可编程系统(system on a programmable chip,SOPC)技术来实现多个摄像头视频数据的传输和拼接的大视场视频解决方案。系...针对大场景视频拼接技术在汽车环视系统等领域的应用需求,并为了适应嵌入式系统快速发展的要求,提出了基于FPGA片上可编程系统(system on a programmable chip,SOPC)技术来实现多个摄像头视频数据的传输和拼接的大视场视频解决方案。系统的硬件平台的构建以Xilinx软核处理器Microblaze为核心,主要完成了视频数据的采集、存储、处理以及传输等工作,图像拼接部分采用频域相位相关算法,对待拼接图像进行配准,并通过融合算法得到具有360°全景视频信息的大视场图像。通过实验对系统进行测试,验证了系统的可行性。展开更多
针对云台网络摄像机监控系统,提出一种基于摄像机视频流的全景图生成算法,以构建更大的监控场景。根据帧间重叠区域的大小选取关键帧,进行柱面投影,利用计算性能优越的SURF(Speeded Up Robust Features,加速鲁棒性特征)算法对所选取的...针对云台网络摄像机监控系统,提出一种基于摄像机视频流的全景图生成算法,以构建更大的监控场景。根据帧间重叠区域的大小选取关键帧,进行柱面投影,利用计算性能优越的SURF(Speeded Up Robust Features,加速鲁棒性特征)算法对所选取的关键帧进行特征点提取,使用基于哈希映射的特征点匹配算法加快特征点的匹配,并结合RANSAC(RANdom SAmple Consensus,随机抽样一致)算法剔除误匹配,估计关键帧之间的变换关系。实验结果表明,该方法能较好实现视频序列的快速拼接,鲁棒性强,具有较高的实用价值。展开更多
文摘We present a method of 3D image mosaicing for real 3D representation of roadside buildings, and implement a Web-based interactive visualization environment for the 3D video mosaics created by 3D image mosaicing. The 3D image mo- saicing technique developed in our previous work is a very powerful method for creating textured 3D-GIS data without excessive data processing like the laser or stereo system. For the Web-based open access to the 3D video mosaics, we build an interactive visualization environment using X3D, the emerging standard of Web 3D. We conduct the data preprocessing for 3D video mosaics and the X3D modeling for textured 3D data. The data preprocessing includes the conversion of each frame of 3D video mosaics into concatenated image files that can be hyperlinked on the Web. The X3D modeling handles the representation of concatenated images using necessary X3D nodes. By employing X3D as the data format for 3D image mosaics, the real 3D representation of roadside buildings is extended to the Web and mobile service systems.
文摘This paper presents a full view video navigation platform and the approach to buildup such a platform.Unlike the traditional full view navigation system,which always bases on image panorama,our system is based on full view video created by video mosaics.Neither the current concept of video mosaics,which means posting together video frames to get a wide view,while these frames are from the same video.Video mosaics in this paper mean stitching different videos together to get a full view video.Videos,together recording 360° of the scenery,are captured through a certain route by digital cameras arranged in a circle around the view point.Then break up these videos into frame sequence.Frame coming from different videos with the same timestamp are stitched into a panoramic image.Consider these panoramic images as frames;montage these panoramic frames into a full view video,by sequence of time.While displaying the full view video people could decide to go forward,or stop to see around,or proceed in a certain direction.While talking about full view navigation,we are mostly talking about image-based cylindrical or spherical panorama [2].One way to create such a panorama is,after a series of translation,stitching together a group of images taken by panning camera [1].Each of these images presents part of the scene and with certain degrees of overlap to adjacent ones.After stitching,we get a panorama image with width is much bigger than the raw images,map it onto a cylinder or sphere surface is a better way to show it.The mapped one is called cylindrical panorama or spherical panorama.While navigating,taking sphere panorama as an example,people were supposed to standing at the center of the sphere,by rotating the sphere,different view coming into eyes.This kind of navigation makes people feel as the same as standing in the real world.But one limitation is that the step is stuck to one point,could see around but not move forward or back(there is moving forward and back by zooming in and zooming out respectively,but it's not truly move).In our work,videos are used as the stitching source instead of images.Video,or could say a sequence of images taken at continuous view point(actually it is the definition of 'video',and is the saying we are going to take advantage of in this paper),are captured by a camera circle,and are stitched into a full view video.Frames coming from different video with the same timestamp are stitched into a panorama frame.Transform the processed panorama frames into a video.
文摘针对大场景视频拼接技术在汽车环视系统等领域的应用需求,并为了适应嵌入式系统快速发展的要求,提出了基于FPGA片上可编程系统(system on a programmable chip,SOPC)技术来实现多个摄像头视频数据的传输和拼接的大视场视频解决方案。系统的硬件平台的构建以Xilinx软核处理器Microblaze为核心,主要完成了视频数据的采集、存储、处理以及传输等工作,图像拼接部分采用频域相位相关算法,对待拼接图像进行配准,并通过融合算法得到具有360°全景视频信息的大视场图像。通过实验对系统进行测试,验证了系统的可行性。
