Recently, there have been some attempts of Transformer in 3D point cloud classification. In order to reduce computations, most existing methods focus on local spatial attention,but ignore their content and fail to est...Recently, there have been some attempts of Transformer in 3D point cloud classification. In order to reduce computations, most existing methods focus on local spatial attention,but ignore their content and fail to establish relationships between distant but relevant points. To overcome the limitation of local spatial attention, we propose a point content-based Transformer architecture, called PointConT for short. It exploits the locality of points in the feature space(content-based), which clusters the sampled points with similar features into the same class and computes the self-attention within each class, thus enabling an effective trade-off between capturing long-range dependencies and computational complexity. We further introduce an inception feature aggregator for point cloud classification, which uses parallel structures to aggregate high-frequency and low-frequency information in each branch separately. Extensive experiments show that our PointConT model achieves a remarkable performance on point cloud shape classification. Especially, our method exhibits 90.3% Top-1 accuracy on the hardest setting of ScanObjectN N. Source code of this paper is available at https://github.com/yahuiliu99/PointC onT.展开更多
The classification of point cloud data is the key technology of point cloud data information acquisition and 3D reconstruction, which has a wide range of applications. However, the existing point cloud classification ...The classification of point cloud data is the key technology of point cloud data information acquisition and 3D reconstruction, which has a wide range of applications. However, the existing point cloud classification methods have some shortcomings when extracting point cloud features, such as insufficient extraction of local information and overlooking the information in other neighborhood features in the point cloud, and not focusing on the point cloud channel information and spatial information. To solve the above problems, a point cloud classification network based on graph convolution and fusion attention mechanism is proposed to achieve more accurate classification results. Firstly, the point cloud is regarded as a node on the graph, the k-nearest neighbor algorithm is used to compose the graph and the information between points is dynamically captured by stacking multiple graph convolution layers;then, with the assistance of 2D experience of attention mechanism, an attention mechanism which has the capability to integrate more attention to point cloud spatial and channel information is introduced to increase the feature information of point cloud, aggregate local useful features and suppress useless features. Through the classification experiments on ModelNet40 dataset, the experimental results show that compared with PointNet network without considering the local feature information of the point cloud, the average classification accuracy of the proposed model has a 4.4% improvement and the overall classification accuracy has a 4.4% improvement. Compared with other networks, the classification accuracy of the proposed model has also been improved.展开更多
Ground-based cloud classification is challenging due to extreme variations in the appearance of clouds under different atmospheric conditions. Texture classification techniques have recently been introduced to deal wi...Ground-based cloud classification is challenging due to extreme variations in the appearance of clouds under different atmospheric conditions. Texture classification techniques have recently been introduced to deal with this issue. A novel texture descriptor, the salient local binary pattern (SLBP), is proposed for ground-based cloud classification. The SLBP takes advantage of the most frequently occurring patterns (the salient patterns) to capture descriptive information. This feature makes the SLBP robust to noise. Experimental results using ground-based cloud images demonstrate that the proposed method can achieve better results than current state-of-the-art methods.展开更多
This paper presents the automated pixel-scale neural network classification methods being developed at National Satellite Meteorological Center(NSMC)of China to classify clouds by using NOAA/AVHRR and GMS-5 satellite ...This paper presents the automated pixel-scale neural network classification methods being developed at National Satellite Meteorological Center(NSMC)of China to classify clouds by using NOAA/AVHRR and GMS-5 satellite imageries.By using Terra satellite MODIS imageries,an automated pixel-scale threshold technique has been developed to detect and classify clouds.The study focuses on applications of these cloud classification techniques to the Huaihe River and the Changjiang(Yangtze)River drainage basin.The different types of clouds show more clearly on this cloud classification image than single band image.The results of the cloud classifications are the basis of studying cloud amount,cloud top height and cloud top pressure.Cloud mask methods are widely used in SST,LST,and TPW retrieval schemes.Some case studies about cloud mask and cloud classification in satellite imageries,which are related with the study of Global Energy and Water Cycle Experiment(GEWEX)in the Huaihe River and the Changjiang River drainage basin are illustrated.展开更多
It is thought that satellite infrared (IR) images can aid the recognition of the structure of the cloud and aid the rainfall estimation. In this article, the authors explore the application of a classification metho...It is thought that satellite infrared (IR) images can aid the recognition of the structure of the cloud and aid the rainfall estimation. In this article, the authors explore the application of a classification method relevant to four texture features, viz. energy, entropy, inertial-quadrature and local calm, to the study of the structure of a cloud cluster displaying a typical meso-scaie structure on infrared satellite images. The classification using the IR satellite images taken during 4-5 July 2003, a time when a meso-scale torrential rainstorm was occurring over the Yangtze River basin, illustrates that the detailed structure of the cloud cluster can be obviously seen by means of the neural network classification method relevant to textural features, and the relationship between the textural energy and rainfall indicates that the structural variation of a cloud cluster can be viewed as an exhibition of the convection intensity evolvement. These facts suggest that the scheme of following a classification method relevant to textural features applied to cloud structure studies is helpful for weather analysis and forecasting.展开更多
A convective and stratiform cloud classification method for weather radar is proposed based on the density-based spatial clustering of applications with noise(DBSCAN)algorithm.To identify convective and stratiform clo...A convective and stratiform cloud classification method for weather radar is proposed based on the density-based spatial clustering of applications with noise(DBSCAN)algorithm.To identify convective and stratiform clouds in different developmental phases,two-dimensional(2D)and three-dimensional(3D)models are proposed by applying reflectivity factors at 0.5°and at 0.5°,1.5°,and 2.4°elevation angles,respectively.According to the thresholds of the algorithm,which include echo intensity,the echo top height of 35 dBZ(ET),density threshold,andεneighborhood,cloud clusters can be marked into four types:deep-convective cloud(DCC),shallow-convective cloud(SCC),hybrid convective-stratiform cloud(HCS),and stratiform cloud(SFC)types.Each cloud cluster type is further identified as a core area and boundary area,which can provide more abundant cloud structure information.The algorithm is verified using the volume scan data observed with new-generation S-band weather radars in Nanjing,Xuzhou,and Qingdao.The results show that cloud clusters can be intuitively identified as core and boundary points,which change in area continuously during the process of convective evolution,by the improved DBSCAN algorithm.Therefore,the occurrence and disappearance of convective weather can be estimated in advance by observing the changes of the classification.Because density thresholds are different and multiple elevations are utilized in the 3D model,the identified echo types and areas are dissimilar between the 2D and 3D models.The 3D model identifies larger convective and stratiform clouds than the 2D model.However,the developing convective clouds of small areas at lower heights cannot be identified with the 3D model because they are covered by thick stratiform clouds.In addition,the 3D model can avoid the influence of the melting layer and better suggest convective clouds in the developmental stage.展开更多
In this paper,improvement on man-computer interactive classification of clouds based on hispeetral satellite imagery has been synthesized by using the maximum likelihood automatic clustering(MLAC)and the unit feature ...In this paper,improvement on man-computer interactive classification of clouds based on hispeetral satellite imagery has been synthesized by using the maximum likelihood automatic clustering(MLAC)and the unit feature space classification(UFSC)approaches.The improved classification not only shortens the time of sample-training in UFSC method,but also eliminates the inevitable shortcomings of the MLAC method.(e.g.,1.sample selecting and training is confined only to one cloud image:2.the result of clustering is pretty sensitive to the selection of initial cluster center:3.the actual classification basically can not satisfy the supposition of normal distribution required by MLAC method;4.errors in classification are difficult to be modified.) Moreover,it makes full use of the professionals'accumulated knowledge and experience of visual cloud classifications and the cloud report of ground observation,having ensured both the higher accuracy of classification and its wide application as well.展开更多
Topic modeling is a mainstream and effective technology to deal with text data, with wide applications in text analysis, natural language, personalized recommendation, computer vision, etc. Among all the known topic m...Topic modeling is a mainstream and effective technology to deal with text data, with wide applications in text analysis, natural language, personalized recommendation, computer vision, etc. Among all the known topic models, supervised Latent Dirichlet Allocation (sLDA) is acknowledged as a popular and competitive supervised topic model. How- ever, the gradual increase of the scale of datasets makes sLDA more and more inefficient and time-consuming, and limits its applications in a very narrow range. To solve it, a parallel online sLDA, named PO-sLDA (Parallel and Online sLDA), is proposed in this study. It uses the stochastic variational inference as the learning method to make the training procedure more rapid and efficient, and a parallel computing mechanism implemented via the MapReduce framework is proposed to promote the capacity of cloud computing and big data processing. The online training capacity supported by PO-sLDA expands the application scope of this approach, making it instrumental for real-life applications with high real-time demand. The validation using two datasets with different sizes shows that the proposed approach has the comparative accuracy as the sLDA and can efficiently accelerate the training procedure. Moreover, its good convergence and online training capacity make it lucrative for the large-scale text data analyzing and processing.展开更多
基金supported in part by the Nationa Natural Science Foundation of China (61876011)the National Key Research and Development Program of China (2022YFB4703700)+1 种基金the Key Research and Development Program 2020 of Guangzhou (202007050002)the Key-Area Research and Development Program of Guangdong Province (2020B090921003)。
文摘Recently, there have been some attempts of Transformer in 3D point cloud classification. In order to reduce computations, most existing methods focus on local spatial attention,but ignore their content and fail to establish relationships between distant but relevant points. To overcome the limitation of local spatial attention, we propose a point content-based Transformer architecture, called PointConT for short. It exploits the locality of points in the feature space(content-based), which clusters the sampled points with similar features into the same class and computes the self-attention within each class, thus enabling an effective trade-off between capturing long-range dependencies and computational complexity. We further introduce an inception feature aggregator for point cloud classification, which uses parallel structures to aggregate high-frequency and low-frequency information in each branch separately. Extensive experiments show that our PointConT model achieves a remarkable performance on point cloud shape classification. Especially, our method exhibits 90.3% Top-1 accuracy on the hardest setting of ScanObjectN N. Source code of this paper is available at https://github.com/yahuiliu99/PointC onT.
文摘The classification of point cloud data is the key technology of point cloud data information acquisition and 3D reconstruction, which has a wide range of applications. However, the existing point cloud classification methods have some shortcomings when extracting point cloud features, such as insufficient extraction of local information and overlooking the information in other neighborhood features in the point cloud, and not focusing on the point cloud channel information and spatial information. To solve the above problems, a point cloud classification network based on graph convolution and fusion attention mechanism is proposed to achieve more accurate classification results. Firstly, the point cloud is regarded as a node on the graph, the k-nearest neighbor algorithm is used to compose the graph and the information between points is dynamically captured by stacking multiple graph convolution layers;then, with the assistance of 2D experience of attention mechanism, an attention mechanism which has the capability to integrate more attention to point cloud spatial and channel information is introduced to increase the feature information of point cloud, aggregate local useful features and suppress useless features. Through the classification experiments on ModelNet40 dataset, the experimental results show that compared with PointNet network without considering the local feature information of the point cloud, the average classification accuracy of the proposed model has a 4.4% improvement and the overall classification accuracy has a 4.4% improvement. Compared with other networks, the classification accuracy of the proposed model has also been improved.
基金Supported by the National Natural Science Foundation of China (61172103, 60933010, and 60835001)
文摘Ground-based cloud classification is challenging due to extreme variations in the appearance of clouds under different atmospheric conditions. Texture classification techniques have recently been introduced to deal with this issue. A novel texture descriptor, the salient local binary pattern (SLBP), is proposed for ground-based cloud classification. The SLBP takes advantage of the most frequently occurring patterns (the salient patterns) to capture descriptive information. This feature makes the SLBP robust to noise. Experimental results using ground-based cloud images demonstrate that the proposed method can achieve better results than current state-of-the-art methods.
基金the National Natural Science Foundation of China(49794030).
文摘This paper presents the automated pixel-scale neural network classification methods being developed at National Satellite Meteorological Center(NSMC)of China to classify clouds by using NOAA/AVHRR and GMS-5 satellite imageries.By using Terra satellite MODIS imageries,an automated pixel-scale threshold technique has been developed to detect and classify clouds.The study focuses on applications of these cloud classification techniques to the Huaihe River and the Changjiang(Yangtze)River drainage basin.The different types of clouds show more clearly on this cloud classification image than single band image.The results of the cloud classifications are the basis of studying cloud amount,cloud top height and cloud top pressure.Cloud mask methods are widely used in SST,LST,and TPW retrieval schemes.Some case studies about cloud mask and cloud classification in satellite imageries,which are related with the study of Global Energy and Water Cycle Experiment(GEWEX)in the Huaihe River and the Changjiang River drainage basin are illustrated.
基金This work was supported by the National Natural Science Foundation of China under Grant Nos. 40405009 and 40575022, by the Jiangsu Natural Science Foundation Program through Grant No. BK2005141.
文摘It is thought that satellite infrared (IR) images can aid the recognition of the structure of the cloud and aid the rainfall estimation. In this article, the authors explore the application of a classification method relevant to four texture features, viz. energy, entropy, inertial-quadrature and local calm, to the study of the structure of a cloud cluster displaying a typical meso-scaie structure on infrared satellite images. The classification using the IR satellite images taken during 4-5 July 2003, a time when a meso-scale torrential rainstorm was occurring over the Yangtze River basin, illustrates that the detailed structure of the cloud cluster can be obviously seen by means of the neural network classification method relevant to textural features, and the relationship between the textural energy and rainfall indicates that the structural variation of a cloud cluster can be viewed as an exhibition of the convection intensity evolvement. These facts suggest that the scheme of following a classification method relevant to textural features applied to cloud structure studies is helpful for weather analysis and forecasting.
基金funded by the Key-Area Research and Development Program of Guangdong Province(Grant No.2020B1111200001)the Key project of monitoring,early warning and prevention of major natural disasters of China(Grant No.2019YFC1510304)+1 种基金the S&T Program of Hebei(Grant No.19275408D)the Scientific Research Projects of Weather Modification in Northwest China(Grant No.RYSY201905).
文摘A convective and stratiform cloud classification method for weather radar is proposed based on the density-based spatial clustering of applications with noise(DBSCAN)algorithm.To identify convective and stratiform clouds in different developmental phases,two-dimensional(2D)and three-dimensional(3D)models are proposed by applying reflectivity factors at 0.5°and at 0.5°,1.5°,and 2.4°elevation angles,respectively.According to the thresholds of the algorithm,which include echo intensity,the echo top height of 35 dBZ(ET),density threshold,andεneighborhood,cloud clusters can be marked into four types:deep-convective cloud(DCC),shallow-convective cloud(SCC),hybrid convective-stratiform cloud(HCS),and stratiform cloud(SFC)types.Each cloud cluster type is further identified as a core area and boundary area,which can provide more abundant cloud structure information.The algorithm is verified using the volume scan data observed with new-generation S-band weather radars in Nanjing,Xuzhou,and Qingdao.The results show that cloud clusters can be intuitively identified as core and boundary points,which change in area continuously during the process of convective evolution,by the improved DBSCAN algorithm.Therefore,the occurrence and disappearance of convective weather can be estimated in advance by observing the changes of the classification.Because density thresholds are different and multiple elevations are utilized in the 3D model,the identified echo types and areas are dissimilar between the 2D and 3D models.The 3D model identifies larger convective and stratiform clouds than the 2D model.However,the developing convective clouds of small areas at lower heights cannot be identified with the 3D model because they are covered by thick stratiform clouds.In addition,the 3D model can avoid the influence of the melting layer and better suggest convective clouds in the developmental stage.
文摘In this paper,improvement on man-computer interactive classification of clouds based on hispeetral satellite imagery has been synthesized by using the maximum likelihood automatic clustering(MLAC)and the unit feature space classification(UFSC)approaches.The improved classification not only shortens the time of sample-training in UFSC method,but also eliminates the inevitable shortcomings of the MLAC method.(e.g.,1.sample selecting and training is confined only to one cloud image:2.the result of clustering is pretty sensitive to the selection of initial cluster center:3.the actual classification basically can not satisfy the supposition of normal distribution required by MLAC method;4.errors in classification are difficult to be modified.) Moreover,it makes full use of the professionals'accumulated knowledge and experience of visual cloud classifications and the cloud report of ground observation,having ensured both the higher accuracy of classification and its wide application as well.
基金This work was supported in part by the National Natural Science Foundation of China under Grant Nos. 61572226 and 61876069, and the Key Scientific and Technological Research and Development Project of Jilin Province of China under Grant Nos. 20180201067GX and 20180201044GX.
文摘Topic modeling is a mainstream and effective technology to deal with text data, with wide applications in text analysis, natural language, personalized recommendation, computer vision, etc. Among all the known topic models, supervised Latent Dirichlet Allocation (sLDA) is acknowledged as a popular and competitive supervised topic model. How- ever, the gradual increase of the scale of datasets makes sLDA more and more inefficient and time-consuming, and limits its applications in a very narrow range. To solve it, a parallel online sLDA, named PO-sLDA (Parallel and Online sLDA), is proposed in this study. It uses the stochastic variational inference as the learning method to make the training procedure more rapid and efficient, and a parallel computing mechanism implemented via the MapReduce framework is proposed to promote the capacity of cloud computing and big data processing. The online training capacity supported by PO-sLDA expands the application scope of this approach, making it instrumental for real-life applications with high real-time demand. The validation using two datasets with different sizes shows that the proposed approach has the comparative accuracy as the sLDA and can efficiently accelerate the training procedure. Moreover, its good convergence and online training capacity make it lucrative for the large-scale text data analyzing and processing.