Building and using maps is a fundamental issue for bionic robots in field applications. A dense surface map, which offers rich visual and geometric information, is an ideal representation of the environment for indoor...Building and using maps is a fundamental issue for bionic robots in field applications. A dense surface map, which offers rich visual and geometric information, is an ideal representation of the environment for indoor/outdoor localization, navigation, and recognition tasks of these robots. Since most bionic robots can use only small light-weight laser scanners and cameras to acquire semi-dense point cloud and RGB images, we propose a method to generate a consistent and dense surface map from this kind of semi-dense point cloud and RGB images. The method contains two main steps： （1） generate a dense surface for every single scan of point cloud and its corresponding image（s） and （2） incrementally fuse the dense surface of a new scan into the whole map. In step （1） edge-aware resampling is realized by segmenting the scan of a point cloud in advance and resampling each sub-cloud separately. Noine within the scan is reduced and a dense surface is generated. In step （2） the average surface is estimated probabilistically and the non-coincidence of different scans is eliminated. Experiments demonstrate that our method works well in both indoor and outdoor semi-structured environments where there are regularly shaped objects.展开更多
High-quality 3D reconstruction is an important topic in computer graphics and computer vision with many applications,such as robotics and augmented reality.The advent of consumer RGB-D cameras has made a profound adva...High-quality 3D reconstruction is an important topic in computer graphics and computer vision with many applications,such as robotics and augmented reality.The advent of consumer RGB-D cameras has made a profound advance in indoor scene reconstruction.For the past few years,researchers have spent significant effort to develop algorithms to capture 3D models with RGB-D cameras.As depth images produced by consumer RGB-D cameras are noisy and incomplete when surfaces are shiny,bright,transparent,or far from the camera,obtaining highquality 3D scene models is still a challenge for existing systems.We here review high-quality 3D indoor scene reconstruction methods using consumer RGB-D cameras.In this paper,we make comparisons and analyses from the following aspects:(i)depth processing methods in 3D reconstruction are reviewed in terms of enhancement and completion,(ii)ICP-based,feature-based,and hybrid methods of camera pose estimation methods are reviewed,and(iii)surface reconstruction methods are reviewed in terms of surface fusion,optimization,and completion.The performance of state-of-the-art methods is also compared and analyzed.This survey will be useful for researchers who want to follow best practices in designing new high-quality 3D reconstruction methods.展开更多
基金Project supported by the National Natural Science Foundation of China （Nos. 61075078 and 61473258）
文摘Building and using maps is a fundamental issue for bionic robots in field applications. A dense surface map, which offers rich visual and geometric information, is an ideal representation of the environment for indoor/outdoor localization, navigation, and recognition tasks of these robots. Since most bionic robots can use only small light-weight laser scanners and cameras to acquire semi-dense point cloud and RGB images, we propose a method to generate a consistent and dense surface map from this kind of semi-dense point cloud and RGB images. The method contains two main steps： （1） generate a dense surface for every single scan of point cloud and its corresponding image（s） and （2） incrementally fuse the dense surface of a new scan into the whole map. In step （1） edge-aware resampling is realized by segmenting the scan of a point cloud in advance and resampling each sub-cloud separately. Noine within the scan is reduced and a dense surface is generated. In step （2） the average surface is estimated probabilistically and the non-coincidence of different scans is eliminated. Experiments demonstrate that our method works well in both indoor and outdoor semi-structured environments where there are regularly shaped objects.
基金National Key R&D Program of China under Grant No.2018YFC2000600Open Projects Program of National Laboratory of Pattern Recognition under Grant No.202100009+1 种基金National Natural Science Foundation of China under Grant No.72071018Fundamental Research Funds for Central Universities under Grant No.2021TD006。
文摘High-quality 3D reconstruction is an important topic in computer graphics and computer vision with many applications,such as robotics and augmented reality.The advent of consumer RGB-D cameras has made a profound advance in indoor scene reconstruction.For the past few years,researchers have spent significant effort to develop algorithms to capture 3D models with RGB-D cameras.As depth images produced by consumer RGB-D cameras are noisy and incomplete when surfaces are shiny,bright,transparent,or far from the camera,obtaining highquality 3D scene models is still a challenge for existing systems.We here review high-quality 3D indoor scene reconstruction methods using consumer RGB-D cameras.In this paper,we make comparisons and analyses from the following aspects:(i)depth processing methods in 3D reconstruction are reviewed in terms of enhancement and completion,(ii)ICP-based,feature-based,and hybrid methods of camera pose estimation methods are reviewed,and(iii)surface reconstruction methods are reviewed in terms of surface fusion,optimization,and completion.The performance of state-of-the-art methods is also compared and analyzed.This survey will be useful for researchers who want to follow best practices in designing new high-quality 3D reconstruction methods.
