331 results about "Spatial clustering" patented technology

The invention discloses a hard disk failure prediction method for a cloud computing platform. The hard disk failure predication method comprises the following steps: marking SMART log data of a hard disk as a normal hard disk sample and a faulted hard disk sample according to a hard disk maintenance record in a prediction time window; then, dividing the denoised normal hard disk sample into k non-intersected subsets by adopting a K-means clustering algorithm; combining the k non-intersected subsets with the faulted hard disk sample respectively; generating k groups of balance training sets according to an SMOTE (Synthetic Minority Oversampling Technique) so as to obtain k support vector machine classifiers for predicting the faulted hard disk. In the prediction stage, test sets can be clustered by using a DBSCAN (Density-based Spatial Clustering Of Applications With Noise), a sample in a clustered cluster is predicted as the normal hard disk sample, a noise sample is predicted by each classifier obtained by training, and further a final prediction result is obtained by voting. According to the method disclosed by the invention, hard disk fault prediction is carried out by using the SMART data of the hard disk, and relatively high fault recall ratio and overall performance can be obtained.

The invention discloses a target detection method based on 3D laser radar and image data. The method comprises the following steps: acquiring 3D point cloud data and camera images of the surrounding environment by using 3D laser radar and camera, and pre-processing the 3D point cloud data; filtering out ground points in the 3D point cloud data, performing a spatial clustering on the remaining non-ground points, and extracting the 3D interest region of the target; calibrating the external parameters of the coordinates of the 3D laser radar and the camera, and mapping the 3D interest region of the target to the corresponding camera image according to the calibrated parameters, and extracting the corresponding 2D interest region in the camera image; performing a feature extraction on the 2D interest region using a deep convolution network to locate and identify objects in the 2D interest region. According to the target detection method based on 3D laser radar and image data, the complementarity between the 3D laser radar and the camera data is fully utilized, the accuracy and timeliness of the target positioning and classification recognition of the scene are improved, and the invention can be used for real-time detection of the target in an unmanned vehicle.

The invention discloses a dynamic road segment division based vehicle route guidance method, which is characterized in that a dynamic rod network connected graph is built through dynamic road segment division for the purpose of searching optimal path search and realizing dynamic navigation. The method specifically comprises the following steps: first, acquiring vehicle real-time information through a vehicle networking technology by the traffic center, and utilizing an algorithm of Density-based Spatial Clustering Of Applications With Noise (DBSCAN) to regularly and dynamically divide the regional rods, so as to generate the dynamic rod network connected graph; secondly, sending the position and destination of a vehicle itself to a traffic information center for asking for the optimal path; and finally, generating the optimal path on the dynamic rod network connected graph through utilizing a shortest path algorithm by the traffic information center according to the position and destination of the vehicle, and sending the information to the vehicle and realizing path guidance. The method has the advantages that the generated dynamic rod network connected graph which is accurate and real-time can provide the optimal path guidance for a traveler, thereby alleviating city traffic jam and improving running efficiency.

The invention discloses a significant object detection method based on sparse subspace clustering and low-order expression. The method comprises the steps of: 1, carrying out super pixel segmentation and clustering on an input image; 2, extracting the color, texture and edge characteristics of each super pixel in clusters, and constructing cluster characteristic matrixes; 3, ranking all super pixel characteristics according to the magnitude of color contrast, and constructing a dictionary; according to the dictionary, constructing a combined low-order expression model, solving the model and decomposing the characteristic matrixes of the clusters so as to obtain low-order expression coefficients, and calculating significant factors of the clusters; and 5, mapping the significant value of each cluster into the input image according the spatial position, and obtaining a significant map of the input image. According to the invention, the significant objects relatively large in size in the image can be completely and consistently detected, the noise in a background is inhibited, and the robustness of significant object detection of the image with the complex background is improved. The significant object detection method is applicable to image segmentation, object identification, image restoration and self-adaptive image compression.

The invention discloses an extraction method for dynamic crowd gathering characteristics and belongs to the technical field of intelligent monitoring of computers. According to the extraction method, a nuclear density spatial clustering algorithm is introduced into the 'external characteristics analysis process of crowd gathering', so that extraction and quantitative study and judgment of the crowd gathering characteristics for a mass disturbance are effectively achieved. As the crowd movement mode is recognized by the adoption of the method of crowd barycenter motion tracking, the current situation that at present, mass disturbance early warning is performed only according to crowd density grades, other external characteristics, such as the crowd gathering shape, the crowd movement speed and the crowd growth rate, of crowd gathering fail to be fully considered, and the false alarm rate and the missed alarm rate are high is changed. The extraction method can be applied to other safety-sensitive crowded places, thereby having broad popularization and application prospects.

A pipe network node flow measuring and dispatching method based on pressure monitoring comprises the steps that node pressure of a predetermined node in a pipe network is measured actually, and according to the measured node pressure, based on the discrete point spatial interpolation computing method, node water heads of all nodes are worked out; the pipe section flow of a predetermined pipe section is measured actually, coefficients of friction resistance of all flow measuring pipe sections are determined according to the actually measured pipe section flow and the worked-out node water heads, pipelines are grouped through the spatial clustering algorithm according to the coefficients of friction resistance of the pipe sections, and coefficients of friction resistance of all pipe sections are determined; according to the worked-out node water heads, the coefficients of friction resistance of the pipe sections, pipe lengths and pipe diameters, the pipe section flows of all the pipe sections are worked out; according to the material equilibrium principle, the node flows of all the nodes are worked out, and therefore pipe network model updating and checking, pipe network dispatching, leakage quantity analysis, leaking point positioning and other pipeline maintenance and operation management are conducted conveniently. By applying the method, energy consumption and loss due to leakage can be reduced substantially, and the pipeline management level is improved.

Controlling drones and vehicles in package delivery, in one aspect, may include routing a delivery vehicle loaded with packages to a dropoff location based on executing on a hardware processor a spatial clustering of package destinations. A set of drones may be dispatched. A drone-to-package assignment is determined for the drones and the packages in the delivery vehicle. The drone is controlled to travel from the vehicle's dropoff location to transport the assigned package to a destination point and return to the dropoff location to meet the vehicle. The delivery vehicle may be alerted to speed up or slow down to meet the drone at the return location, for example, without the delivery vehicle having to stop and wait at the dropoff location while the drone is making its delivery.

The invention discloses an indoor passive positioning method based on channel state information and a support vector machine. The method comprises the following steps: firstly preprocessing the acquired channel state information data, performing de-noising and smoothness through the adoption of a density-based spatial clustering of applications with noise and a weight-based moving average algorithm, and then using the principal component analysis algorithm to extract the features. The data after the preprocessing and feature-extracting can reflect the signal change more accurately and the dimension is greatly reduced. The passive positioning adopts two-stage positioning. In the training stage, the large positioning space is divided into sub-regions, the support vector machine classification and regression model is established for each sub-region so as to acquire a statistic model for accurately representing the nonlinear relationship between the position and the signal. The two-stage positioning firstly determines the sub-regions through the classification of the support vector machine, and the precision position is determined in the sub-region through the regression of the support vector machine. The method disclosed by the invention has the beneficial effects that the passive positioning can be performed in the absence of the active participation of the target, and the indoor positioning precision is improved to sub-meter level.

The invention discloses an adaptive spatial clustering method, comprising the following steps of: (1) preprocessing spatial data and selecting features; (2) creating a Delaunay triangulation network according to spatial attribute; (3) performing clustering analysis operations according to the spatial attribute; (4) turning to a step (5) if a spatial solid obstacle is needed to be further considered, and turning to a step (6) if a thematic attribute is needed to be considered, otherwise, ending the spatial clustering operations; (5) introducing a spatial obstacle layer, performing overlap analysis on the spatial obstacle and the side length of the Delaunay triangulation network between the entities in each spatial cluster, and breaking the side length if the spatial obstacle is intersected with the side length; (6) performing the thematic attribute clustering by an improved density-based spatial clustering method; (7) visualizing the clustering result, and outputting the clustering result. The adaptive spatial clustering method is simple and convenient to operate, high in degree of automation, high in calculation efficiency, perfect in functions, strong in applicability and the like, and can effectively improve capability of spatial clustering analysis to excavate deep-seated geoscience rules.

The invention belongs to computer visual pattern recognition and a video picture processing method. The computer visual pattern recognition and the video picture processing method comprise the steps that establishing a three-dimensional space-time sub-frame cube in a video human action reorganization model, establishing a human action characteristic space, conducting the clustering processing, updating labels, extracting the three-dimensional space-time sub-frame cube in the video human action reorganization model and the human action reorganization from monitoring video, extracting human action characteristic, confirming category of human sub-action in each video and classifying and merging on videos with sub-category labels. According to the computer visual pattern recognition and the video picture processing method, the highest identification accuracy is improved by 16.5% compared with the current international Hollywood2 human action database. Thus, the video human action reorganization method has the advantages that human action characteristic with higher distinguishing ability, adaptability, universality and invariance property can be extracted automatically, the overfitting phenomenon and the gradient diffusion problem in the neural network are lowered, and the accuracy of human action reorganization in a complex environment is improved effectively; the computer visual pattern recognition and the video picture processing method can be applied to the on-site video surveillance and video content retrieval widely.

The invention discloses a traffic travel origin and destination identification method based on a space-time clustering analysis algorithm. A mobile phone GPS positioning instrument is used to collect the complete travel space-time positioning data of a resident in a day. The collected data is subjected to preprocessing, abnormal data is removed, and missing data is repaired. The space-time clustering analysis algorithm based on density is used to identify traffic travel origin and destination. A final traffic travel origin and destination identification result is formed according to different user and user travel time sequence statistics. The characteristics of high precision and continuous tracking of a travel path of mobile phone GPS positioning technology is fully utilized, the identification advantage of the space-time clustering analysis algorithm based on density is developed, the disadvantage of a traditional space-time clustering algorithm in identifying an actual travel endpoint is solved, and the resident traffic travel origin and destination information intelligent identification with the use of mobile phone positioning data is realized. The method can be used for large-scale automated resident traffic travel origin and destination information collection.

The invention relates to a nerve network clustering method based on iteration. The method comprises steps that 1, parameters of an over-limit learning machine model are initiated; 2, samples having the same number as the required clustering number are randomly selected, each sample represents one clustering, an initial model sample set is formed, and an initial over-limit learning machine model is acquired through training; 3, clustering grouping for the samples is carried out by utilizing the initial over-limit learning machine model, and a clustering result is acquired; 4, for each clustering group, multiple samples are selected according to rules as models of the clustering groups; 5, the model samples of each clustering group acquired in the step 4 are utilized to update the over-limit learning machine model; and 6, the process returns to the step 3 for iteration till the clustering groups are stable or iteration frequency requirements are satisfied, and the clustering groups are acquired and outputted. The method solves a high-dimensional non-linear data space clustering problem and further solves problems of large memory consumption and long operation time.

A method for perceive and processing multidimensional information including multidimensional data acquisition, multidimensional data joint processing and after-processing data application, The multi-dimensional data collection includes bayonet passing data, wifi probe detection data, multi-dimensional perception door collection data, face data taken by human face equipment, rfid record data, electronic fence collection data, etc., and collects 110 police information, police force, case, population data, Internet cafe hotel, hotel accommodation, intelligent police kiosk, smart community and other data; Multi-dimensional data joint processing includes spatial clustering analysis and relational map analysis. The application of processed data includes: portrait early warning, accompanying personnel analysis, foothold analysis, lingering vehicle analysis, time-space intersection analysis, peer analysis, daytime and nighttime analysis, daily routine analysis, vehicle decoration analysis, deployment analysis, and crime pattern analysis.

The invention discloses a population identifying method based on cellular signal data and belongs to the field of traffic planning data analysis. The method comprises: (1) preprocessing cellular signaling data to quickly extract available fields so as to form a storage format that uses a user ID (identifier) as a key field; (2) sorting by time, mobile positions of users, and denoising by a speed and angle anomaly judging method; (3) using the DBSCAN (density-based spatial clustering of applications with noise) algorithm to form a dense point, and identifying each user's all daily retention points and retention start and stop times; (4) classifying the retention points monthly to identify permanent population, employed population and short-term floating population. The method of the invention makes full use of cellular signaling data and quickly and accurately calculates, via the above steps, the permanent population, employed population and short-term floating population, thereby providing data support for urban population monitoring and traffic planning.

The invention relates to a load power consumption mode identification method. The load power consumption mode identification method includes the steps: acquiring the electrical load at a sampling time interval T, and obtaining L daily load curves corresponding to L days of time; performing spatial clustering based on density on the obtained daily load curves, and obtaining a classical load power consumption mode; extracting characteristics describing the power consumption behavior of a user in different time scale; and utilizing a gravitation search algorithm to cluster the obtained power consumption characteristics of the user; repeating clustering, utilizing a cluster evaluation index to evaluate the clustering result, and selecting the optimal clustering result, that is, the identification result of the load power consumption mode. The gravitation search algorithm used by the load power consumption mode identification method has high searching capability and high convergence speed, and is not easy to fall into local optimal solution, and is better than a traditional clustering algorithm on the identification effect, so that identification of the load power consumption mode can be effectively realized and powerful guidance for design of the demand side response scheme, analysis of load characteristics and high-accuracy prediction can be provided.

The invention relates to a multi-line lidar-based obstacle clustering method. According to the method, an obstacle clustering system is adopted; the obstacle clustering system comprises a data processing module; and the data processing module is connected with a data acquisition module through an Ethernet, performs analysis and algorithm processing on point cloud data acquired by the data acquisition module, clusters obstacles and outputs the sizes and positions of the obstacles in real time. The multi-line lidar-based obstacle clustering method is based on an existing clustering algorithm; grid filtering and adaptive neighborhood parameters are adopted; density-based spatial clustering of application with noise (DBSCAN) is used in combination; and therefore, the accuracy and real-time performance of obstacle clustering recognition can be improved.

This invention provides one product cloud shape property analysis method based on data dynamic memory module, which comprises the following steps: reading the output data from digital device into memory and establishing linear memory structure for data; then adopting 3D R*-tree for rapid space polymer sorting on data and to establish data dynamic memory; then in the browsing process, using current data point as aim point to get the local reference data set by use of Bezier curve accuracy near to point set to estimate curve rate ; finally mapping the data curve rate as RGB color mode.

The invention discloses an automatic target tracking method for aerially photographed videos and aims to solve the technical problem that an existing aerially photographed video target tracking method based on stable image detection and data association is high in tracking error rate. According to the technical scheme, firstly, a pyramid optical flow method is used for extracting robust optical flow track features rather than simple motion foreground segmentation; secondly, motion constraint, position constraint and existence duration constraint are utilized for effective management of an optical flow track; lastly, clustering in time dimension is conducted according to a spatial clustering result of a plurality of continuous frames to effectively solve the problem of spatial clustering errors under the conditions of target intersection and a close range between targets. According to the automatic target tracking method for the aerially photographed videos, a target tracking result is obtained on the basis of spatial clustering, and therefore the tracking error rate is lowered. Tests prove that the tracking error rate is lowered by 10% from 18% in the background art to 8%.

The invention provides a multi-person motion capture method based on three-dimensional hypothesis spatial clustering. The method can be used for unmarked human motion capture. The method comprises thesteps of associating two-dimensional joint point candidate points between different views, reconstructing three-dimensional joint point candidate points, and carrying out three-dimensional attitude analysis and attitude tracking. Under the condition that a human body model is not utilized or any human body prior knowledge is assumed, stable and credible two-dimensional and global three-dimensional human body posture estimation can be carried out on a plurality of persons with different figures and unfixed number of people. The generated posture meets the multi-view geometric constraint and the human body bone length constraint, and robust and credible human body posture estimation under challenging scenes such as multi-person mutual shielding and close interaction is achieved.

The invention provides a vehicular point cloud clustering method and system. The vehicular point cloud clustering method comprises carrying out de-noising and filtering on point clouds, removing scattered noise points on the point clouds, carrying out filtering processing on the point clouds, and distinguishing between ground points and non-ground points; segmenting the point clouds to generate super voxels, segmenting the non-ground points by means of a spatial clustering algorithm based on the density to generate the super voxels, and guaranteeing that different surface features do not adhere to each other; clustering the point clouds which have spatial context linkage, analyzing features and the spatial context linkage of the super voxels, carrying out region growing on the super voxels by integrating multifactor weights, and completing point cloud clustering. According to the invention, the problem that during the vehicular point cloud clustering process, over-segmentation or insufficient segmentation occurs is improved, the demand of obtaining 3d spatial information quickly is satisfied, and the vehicular point cloud clustering method and system have an important market value.

A background removal apparatus comprises a color normalization module that normalizes an original image, and an image segmentation module that segments the normalized image into several segmented regions. A background mesh generation module simulates the variation of pixel colors in background and generates an interpolated background mesh. A comparison module compares the normalized image and the interpolated background mesh to form a background mask by extracting the coherent regions between them. A refinement module uses a refined rule to determine a final background mask, and a background removal module through which a pure foreground image is obtained. The apparatus combines the efficiency of both color and spatial clustering, and improves the capabilities of current image segmentation method to perform background removal.

The invention provides an armored equipment damage rule analysis method based on a simulation experiment. The method comprises the following steps: obtaining a hit position in which each fight happens in a combat simulation test based on a missile and target space intersection model of an equipment unit; establishing a battle effect simulation model of typical ammunition to a typical part of typical equipment, correcting the real equipment battle effect by means of live ammunition, and determining the reasonability of the model and parameters; establishing a target non-uniform segmentation model based on a K-Means spatial clustering algorithm, optimizing the experimental design of damage analysis of parts in the equipment, and determining damaged parts and equipment damage level under each battle condition; and establishing a damage database matching model, and establishing finite-state off-line damage simulation and an infinite-state on-line real-time matching relationship by adopting a database mapping method, so that the application query is facilitated.

The invention provides a method for identifying traveling OD nodes and extracting a path between nodes in a big data environment. According to the method, traveling path data of massive individuals ismined by using spatial activity data sets of individuals of mobile terminals in a specified time range, and fitting interpolation is performed on the traveling path data, so as to acquire an individual traveling time-space sequence of an equal time interval; a possible cluster region is searched in the individual traveling time-space sequence through a spatial clustering method, intersection angle differences between a center point of the cluster region and external nodes of the cluster region are compared so as to determine whether an extracted cluster point is an OD point, and the travelingtime-space sequence of a user is split. Through adoption of the method, the traveling time-space sequence of massive individuals in a specified time range can be acquired conveniently and automatically at low cost, node regions with an OD feature can be found rapidly through a spatial clustering algorithm and a weighted averaging method, and OD points are determined according to rules, so that ODnode-based road section segmentation is performed on the traveling time-space sequence of the user conveniently and efficiently.

The invention discloses a class center compression transformation-based text clustering method in a search engine. The method comprises the following steps of: by using an improved tf-idf formula, calculating word weight of each file in a text set, calculating an initial class center, mining a synonym word set and a concurrent high-frequency word set, calculating a word center and performing primary classification according to similarity of the initial class center with each file; compressing the center word according to information such as title word, article length, synonyms and concurrent associated words, thereby guaranteeing that the same word only occurs in some class centers with high similarity with the word; clustering the file by using a new cluster center again; calculating core similarity of each class; splitting the biggest class; combining smaller classes to produce a new class; iterating compression, clustering and split operation until the number of the classes converges; and guaranteeing that the similarity of the text in the same class with the cluster center reaches a certain threshold value. The clustering accuracy is obviously higher than those of the conventional methods such as KMeans and DBSCAN (Density-based Spatial Clustering of Applications with Noise).

The present disclosure relates to system, method and apparatuses for identifying load volatility of a power customary and a tangible computer readable medium therefor. In an embodiment of the present disclosure, the system comprises at least one processor; and at least one memory storing computer executable instructions. The at least one memory and the computer executable instructions are configured to, with the at least one processor, cause the system to determine boundary points for splitting a load curve of a power customer automatically, through performing density-based spatial clustering on data points of the load curve of the power customer; and detect tendency turning points of the load curve by means of the determined boundary points, so as to identify the load volatility of the power customer. With embodiments of the present disclosure, the boundary points for splitting the load curve may be determined automatically based on load data of each power customer instead of using a predetermined threshold and thus the load volatility of the load curve, which could provide a solution of self-adapted auto-identification for load volatility.

The invention provides a spatial clustering mining PSE (Problem Solving Environments) system, which comprises a data layer, a functional layer and a user layer, wherein the data layer comprises at least one spatial database for providing basic spatial data; the functional layer is used for packaging a spatial clustering mining module and providing a uniform interface to realize the issuance, the discovery and the call of spatial clustering module service, and is used for visually displaying and returning a spatial clustering analysis result; and the user layer is used for providing the interface for a user to input parameters and select the module service. According to the invention, the spatial clustering mining module is constructed, and an OGC WPS (Web Processing Service) standard is utilized to package the mining module service, so that service sharing is realized on any system and application platform; and a portal architecture is applied, so that the expandability is good, effective support is provided for the discovery and the extraction of data useful in a decision-making process from massive data related to positions, and the application hierarchy and quality of the spatial data mining module is greatly improved and broadened.