A new type of computational three-dimensional (3d) reconstruction in the condition of low light level has been proposed using a planar catadioptric method based on the principle of computational photon counting integr...
详细信息
ISBN:
(纸本)9781509062904
A new type of computational three-dimensional (3d) reconstruction in the condition of low light level has been proposed using a planar catadioptric method based on the principle of computational photon counting integral imaging technique. Only one camera and one shot are employed for photon counting integral imaging, rather than multiple sensors or dynamic movement in previous works. A 3d planar catadioptric photon counting method has been demonstrated under photon staved condition while the 3d scenes can be successfully reconstructed.
data visualization is a science andtechnology research on data visual form, which is originated in the 50's of the twentieth Century. It went from scientific visualization, information visualization to data visua...
详细信息
data visualization is a science andtechnology research on data visual form, which is originated in the 50's of the twentieth Century. It went from scientific visualization, information visualization to data visualization. Based on computer graphics and pictures data visualization display and reveal the main information contained in data. data visualization contains data acquisition, data analysis, data processing anddata modeling. Based on psychology the motivation of data visualization is to expand the scope of visual perception which is easier to get information than other biological perceptions while the effect is *** science andtechnology greatly extends the human sensory "arms".Visual perception has very long "arm" and often "robs" other perceptions' "busi-ness". This paper introduces the basic concept and main content of data visuali-zation, reveals the essence of data visualization. Through the "Agricultural plan-ning visualization" case, the technology process of data visualization applied in production has been established, which can be described as 4 steps operated by order. The first step is to determine anddecompose a target. The second step is data acquisition and processing. The third step is models designing and expres-sion. The last step is video production.
In 3dimage generation, depth map plays an important role in 3dperception of viewers. due to low capability of depth acquisitiondevices as well as ready-made depth maps, depth map upsampling is often needed. In this...
详细信息
In 3dimage generation, depth map plays an important role in 3dperception of viewers. due to low capability of depth acquisitiondevices as well as ready-made depth maps, depth map upsampling is often needed. In this paper, we investigate the relation between objective quality assessments of upsampleddepth map and3d subjective evaluation of stereoscopic image. For the former, full-reference and no-reference image quality assessments are applied to measure the objective quality of upsampleddepth maps. Subjective evaluation is carried out by dSCQS test on the stereoscopic images. We use several upsampling methods to achieve different depth map qualities as our experimental samples. Finally, the quantitative value of each measurement is compared with a subjective assessment using three correlation coefficients. The experiment shows promising results that could help to select the most appropriate objective quality assessment tool(s) for stereoscopic image quality measurement.
The major task of a robotic vision system is to understand the real world by a robot using its visual perception units.2dimage based vision systems have been extensively investigated for several ***,the world where w...
The major task of a robotic vision system is to understand the real world by a robot using its visual perception units.2dimage based vision systems have been extensively investigated for several ***,the world where we are living is a three-dimensional（3d） physical space.2dimages captured from a scene are lack of 3d geometric information and are sensitive to variations in illumination,scale and *** contrast,3d point clouds provide more geometrical information and are more robust to various *** most of existing robotic vision systems work on 2dimages,it is clear that future robotic vision systems will increasingly reply on 3d point ***,the rapid technological development of low-cost 3dacquisition systems（e.g.,Microsoft Kinect,Intel Realsense,Google Tango） makes point clouds more ***,advances in computing devices have further enabled the computationally intensive point cloud processing algorithms to run in a fairly acceptable ***,the release and progress of open-source Point Cloud Library（PCL） has significantly boosted the academic and application research in this *** these combined factors have contributed to the recent flourishment in point cloud processing and3d robotic *** this talk,I will first present the advantages of 3d point clouds compared to traditional 2dimages.I will then review the recent development of 3d point cloudacquisition technologies and systems,including structured light sensors,laser scanners,stereo vision cameras,and time-of-flight ***,I will list a number of key research points in the area of point cloud processing and3d robotic vision,including 3d mapping,localization and navigation,3d object detection and recognition,and human-machine interaction.I will also talk about various potential applications in robotics industry,including domestic service robots,battlefield robots,unmanned aerial vehicles,and underwater robots.
暂无评论