The multitude of airborne point clouds limits the point cloud processing efficiency.Superpoints are grouped based on similar points,which can effectively alleviate the demand for computing resources and improve proces...The multitude of airborne point clouds limits the point cloud processing efficiency.Superpoints are grouped based on similar points,which can effectively alleviate the demand for computing resources and improve processing efficiency.However,existing superpoint segmentation methods focus only on local geometric structures,resulting in inconsistent spectral features of points within a superpoint.Such feature inconsistencies degrade the performance of subsequent tasks.Thus,this study proposes a novel Superpoint Segmentation method that jointly utilizes spatial Geometric and Spectral Information for multispectral point cloud superpoint segmentation(GSI-SS).Specifically,a similarity metric that combines spatial geometry and spectral information is proposed to facilitate the consistency of geometric structures and object attributes within segmented superpoints.Following the formation of the primary superpoints,an intersuperpoint pointexchange mechanism that maximizes feature consistency within the final superpoints is proposed.Experiments are conducted on two real multispectral point cloud datasets,and the proposed method achieved higher recall,precision,F score,and lower global consistency and feature classification errors.The experimental results demonstrate the superiority of the proposed GSI-SS over several state-of-the-art methods.展开更多
The superpoints are the sources （or the destinations） that connect with a great deal of destinations （or sources） during a measurement time interval, so detecting the superpoints in real time is very important to ...The superpoints are the sources （or the destinations） that connect with a great deal of destinations （or sources） during a measurement time interval, so detecting the superpoints in real time is very important to network security and management. Previous algorithms are not able to control the usage of the memory and to deliver the desired accuracy, so it is hard to detect the superpoints on a high speed link in real time. In this paper, we propose an adaptive sampling algorithm to detect the superpoints in real time, which uses a flow sample and hold module to reduce the detection of the non-superpoints and to improve the measurement accuracy of the superpoints. We also design a data stream structure to maintain the flow records, which compensates for the flow Hash collisions statistically. An adaptive process based on different sampling probabilities is used to maintain the recorded IP addresses in the limited memory. This algorithm is compared with the other algorithms by analyzing the real network trace data. Experiment results and mathematic analysis show that this algorithm has the advantages of both the limited memory requirement and high measurement accuracy.展开更多
结构化场景中,存在着低纹理表面为特征的人造环境,基于点特征的SLAM(Simultaneous Localization and Mapping,同时定位与地图构建)算法难以得到足够的匹配点对,从而导致相机估计运动失败。除了点之外,结构化环境提供了大量的几何特征,...结构化场景中,存在着低纹理表面为特征的人造环境,基于点特征的SLAM(Simultaneous Localization and Mapping,同时定位与地图构建)算法难以得到足够的匹配点对,从而导致相机估计运动失败。除了点之外,结构化环境提供了大量的几何特征,例如线和平面。因此,提出一种基于点线面特征融合的SLAM算法。算法将基于深度学习的SuperPoint点特征与传统线面特征相结合,利用结构化场景的特性,将位姿解耦细化。首先,使用线面特征构建MW(Manhattan World,曼哈顿世界)坐标系,利用每一时刻相机与MW坐标系的相对旋转得到相机之间的旋转矩阵;然后,构建点线面特征的重投影误差函数,通过最小化联合误差函数得到平移矩阵;最后,根据结构化环境下平面间相互垂直和平行的特性添加约束函数,同时为弥补环境中出现不严格遵守MW假设的情况,使用关键帧构建的局部地图投影到当前帧进一步优化位姿。在TUM公开数据集上与主流方法对比表明,该算法有效提升了结构化低纹理环境下的SLAM定位精度。展开更多
基金supported by the Youth Project of the National Natural Science Foundation of China(Grant No.62201237)the Yunnan Fundamental Research Projects(Grant Nos.202101BE070001-008 and202301AV070003)+1 种基金the Youth Project of the Xingdian Talent Support Plan of Yunnan Province(Grant No.KKRD202203068)the Major Science and Technology Projects in Yunnan Province(Grant No.202202AD080013)。
文摘The multitude of airborne point clouds limits the point cloud processing efficiency.Superpoints are grouped based on similar points,which can effectively alleviate the demand for computing resources and improve processing efficiency.However,existing superpoint segmentation methods focus only on local geometric structures,resulting in inconsistent spectral features of points within a superpoint.Such feature inconsistencies degrade the performance of subsequent tasks.Thus,this study proposes a novel Superpoint Segmentation method that jointly utilizes spatial Geometric and Spectral Information for multispectral point cloud superpoint segmentation(GSI-SS).Specifically,a similarity metric that combines spatial geometry and spectral information is proposed to facilitate the consistency of geometric structures and object attributes within segmented superpoints.Following the formation of the primary superpoints,an intersuperpoint pointexchange mechanism that maximizes feature consistency within the final superpoints is proposed.Experiments are conducted on two real multispectral point cloud datasets,and the proposed method achieved higher recall,precision,F score,and lower global consistency and feature classification errors.The experimental results demonstrate the superiority of the proposed GSI-SS over several state-of-the-art methods.
基金the National Basic Research Program of China (Grant No. 2003cb314804)
文摘The superpoints are the sources （or the destinations） that connect with a great deal of destinations （or sources） during a measurement time interval, so detecting the superpoints in real time is very important to network security and management. Previous algorithms are not able to control the usage of the memory and to deliver the desired accuracy, so it is hard to detect the superpoints on a high speed link in real time. In this paper, we propose an adaptive sampling algorithm to detect the superpoints in real time, which uses a flow sample and hold module to reduce the detection of the non-superpoints and to improve the measurement accuracy of the superpoints. We also design a data stream structure to maintain the flow records, which compensates for the flow Hash collisions statistically. An adaptive process based on different sampling probabilities is used to maintain the recorded IP addresses in the limited memory. This algorithm is compared with the other algorithms by analyzing the real network trace data. Experiment results and mathematic analysis show that this algorithm has the advantages of both the limited memory requirement and high measurement accuracy.
文摘结构化场景中,存在着低纹理表面为特征的人造环境,基于点特征的SLAM(Simultaneous Localization and Mapping,同时定位与地图构建)算法难以得到足够的匹配点对,从而导致相机估计运动失败。除了点之外,结构化环境提供了大量的几何特征,例如线和平面。因此,提出一种基于点线面特征融合的SLAM算法。算法将基于深度学习的SuperPoint点特征与传统线面特征相结合,利用结构化场景的特性,将位姿解耦细化。首先,使用线面特征构建MW(Manhattan World,曼哈顿世界)坐标系,利用每一时刻相机与MW坐标系的相对旋转得到相机之间的旋转矩阵;然后,构建点线面特征的重投影误差函数,通过最小化联合误差函数得到平移矩阵;最后,根据结构化环境下平面间相互垂直和平行的特性添加约束函数,同时为弥补环境中出现不严格遵守MW假设的情况,使用关键帧构建的局部地图投影到当前帧进一步优化位姿。在TUM公开数据集上与主流方法对比表明,该算法有效提升了结构化低纹理环境下的SLAM定位精度。
