139 results about "Point cluster" patented technology

A scene point cloud is processed and a solution to an inverse-function is determined to determine its source objects. A primitive extraction process and a part matching process are used to compute the inverse function solution. The extraction process estimates models and parameters based on evidence of cylinder and planar geometry in the scene. The matching process matches clusters of 3D points to models of parts from a library. A selected part and its associated polygon model is used to represent the point cluster. Iterations of the extraction and matching processes complete a 3D model for a complex scene made up of planes, cylinders, and complex parts from the parts library. Connecting regions between primitives and / or parts are processed to determine their existence and type. Constraints may be used to ensure a connected model and alignment of its components.

Methods of and systems for resolving multiple location estimate conflicts in a WLAN-positioning system are provided. Disclosed are methods to quantify the probability that a particular location estimate of a mobile device made by a Wi-Fi based positioning system is correct to within an arbitrary accuracy. Implementations use observed access point cluster size, age information for access point location determination, and / or cumulative distribution functions that characterize the conditional probability that one or more access points detected by the mobile device have relocated within a specified time interval to make the probability determinations.

Methods of and systems for estimating the probability of movement of access points in a WLAN-based positioning system are provided. Disclosed are methods to quantify the probability that a particular location estimate of a mobile device made by a Wi-Fi based positioning system is correct to within an arbitrary accuracy. Implementations use observed access point cluster size, age information for access point location determination, and / or the probability that one or more access points detected by the mobile device have relocated based on historic information about the movement of a collection of access points to make the probability determinations.

An apparatus and method are provided including a point-to-point cluster link configured to receive a data packet and determines a cyclic redundancy code check for the data packet. The point-to-point cluster link is configured to add a cyclic redundancy code check bit to the data packet transmitted to the point-to-point inter-cluster link, to clear the cyclic redundancy code check bit when the data packet is received, and to sample a cyclic redundancy code window to identify a corrupted data packet.

A suspicious domain name detection method and a device relate to the network safety field; the embodiment of the application can utilize a zombie Trojan system detection log and a DNS log to detect the IP address of a suspicious domain name; the method comprises the following steps: obtaining the IP address of an active zombie Trojan control end according to the zombie Trojan system detection log;obtaining a corresponding relation between the IP address and the domain name in a detection period according to the DNS log; obtaining an IP address with a corresponding domain name number bigger than a second threshold; extracting a key domain name string of each domain name corresponding to the IP address, calculating the appearance frequency of the character in a preset scope, wherein the preset scope comprises key domain name strings of all domain names corresponding to the IP address; using a cluster algorithm to cluster the IP addresses; determining a suspicious point cluster accordingto the number of active zombie Trojan points contained by each point cluster; determining the IP address included by the suspicious point cluster as the IP address corresponding to the suspicious domain name. The method and device are used for detecting suspicious domain names.

The invention discloses a point cloud scene object extracting method. The point cloud scene object extracting method specifically comprises the steps of 1 utilizing the distribution situation of point cloud data on a Gaussian sphere and utilizing a mean shift algorithm to conduct coarse segmentation on the point cloud data; 2 adopting a distance based point clustering method to conduct fine segmentation on the point cloud data subjected to coarse segmentation; 3 utilizing the nature and curvature information of the Gaussian sphere to conduct shape recognition on the point cloud data subjected to fine segmentation; 4 correcting the point cloud data subjected to shape recognition. The point cloud scene object extracting method can be applied to extraction of point cloud scenes including complex objects not just extraction of specific objects in the scenes or object extraction conducted on aviation LiDAR data including few objects.

The invention provides a multipoint identification method and a multipoint identification system for an infrared touch screen. The method comprises the following steps of: acquiring light path data; generating an image / data structure which is in a preset proportion to the size of the infrared touch screen; establishing the attribute that pixels / data elements are correlated with light paths on the infrared touch screen; and extracting touch point information according to the attribute that the pixels / data elements are correlated with the light paths on the infrared touch screen and the light path data. The number of the light paths correlated with the image pixels / data elements is changed through the light path data; and when the number of the light paths correlated with a certain image pixel / data element is changed into zero, the image pixel / data element is marked as a target point, the target points are clustered according to the density of the marked target points, and the target points clustered into the same type form one touch point. By determining the target points according to the orders of magnitude of the image pixels or the data elements, ghost points and noise points can be effectively removed, and high identification precision can be achieved.

Various aspects of the disclosure relate to distributed multiple-input multiple-output (MIMO) communication such as coordinated beamforming or Joint MIMO. In some aspects, distributed MIMO is used to support communication in a cluster of wireless nodes (e.g., access points). A distributed MIMO scheduling scheme as taught herein is used to schedule the wireless nodes (e.g., access points and / or stations) operating within the cluster. For example, stations may be scheduled across basis services sets of the access points for a downlink transmission and / or an uplink transmission.

Disclosed are an external clustering method and a system for large-scale text data, applied to the information technical field; the designed clustering system comprises a candidate analyzer, a relationship generator, a relationship selection component and a clustering component; each sample point serves as a cluster; a candidate related candidate point cluster is selected for each sample through search technology; the relationship between the sample and the candidate relationship sample is calculated by the relationship generator and outputted to the external storage space by an increasing or decreasing order. The method mainly comprises the following steps: to pre-process the input text set and generate the inverted index and the eigenvector of the text set; to use the search technology to retrieve the candidate relationship set of each file; to calculate the relations of the files with candidate relationship through a relation calculation method; to orderly output the calculated results which are greater than a certain threshold value; then according to the orderly outputted results, the clustering algorithm repeatedly and iteratively combines a text pair with a first direct relationship, so as to eventually achieve the text set clustering output. The device has novel and scientific design, occupies small space in the clustering process and uses large-capacity external memories to make differentiated treatment to the treatment process.

The invention provides an automatic unstructured road boundary line extraction method based on a vehicle-mounted laser point cloud. The method comprises steps: scanning line segmentation is carried out; a topology relation network is built; a moving window method is adopted to acquire ground points; clustering is carried out through a K nearest neighbor clustering method, a pseudo ground point cluster is eliminated according to the number of laser points in the point cluster, and road surface laser point cloud data are acquired; the outer contour point of the road surface laser point cloud data is set to be at the position of the road boundary line, outer contour boundary point extraction is carried out on each scanning line, two ground points with the farthest distance in each scanning line are acquired, and outer contour points at two sides of the road are finally acquired; clustering analysis is carried out, and a left contour point set and a right contour point set of the road areacquired; and boundary line tracking optimization is carried out to acquire the optimal boundary line. Extraction on an unstructured road and a road boundary can be quickly realized based on the laserpoint cloud data direction, an effective solution is provided for automatic unstructured road boundary line extraction, and the automatic road boundary line extraction degree is improved.

A method and apparatus for partitioning a computer cluster through coordination point devices that provide enhanced split brain arbitration during I / O fencing. In one embodiment, a method for using coordination point devices to support Input / Output fencing in a computer cluster comprises processing at least one arbitration request from at least one client computer upon an occurrence of a split brain within a computer cluster, wherein each arbitration request indicates at least one established connection between at least one coordination point device and each client computer of the at least one client computer and each client sub-cluster membership, wherein the at least one coordination point device form a coordination point cluster membership and selecting a client computer of the at least one client computer to continue to be a portion of the computer cluster based on the arbitration request.

The invention provides a single-plant tree point cloud automatic extraction method based on vehicle-mounted laser scanning data, and the method mainly comprises two big steps: obtaining point clusters, called as a candidate tree clusters, from the vehicle-mounted laser scanning data based on the rough classification of cluster features, wherein the point clusters may comprise trees; extracting a single-plant tree point cloud from each candidate tree cluster based on two-stage voxel regional growth. The method combines the overall and local geometrical characteristics of trees, can automatically recognize the seeds of single-plant trees in a complex environment, respectively sets different growth rules for the growing of tree trunks and crowns, and achieves the extraction of the single-plant trees. Meanwhile, compared with other single-plant tree extraction algorithm, the method achieves greater improvement on the extraction integrity and accuracy of trees, the extraction precision of a single-plant tree point cloud and the extraction in the complex environment, and can be used for further information extraction of trees.

The invention discloses a smart-phone track chain-cluster identification method considering sequential DBSCAN. The method comprises that data collection and preprocessing are carried out; the sequential angle offset of behavior track points is calculated; the sequential distance offset of the behavior track points is calculated; chains and non-chains are identified and gathered according to rules; and whether a non-chain segment belongs to a point cluster is determined by using DBSCAN, and outputs an analysis result. The logical process of the behavior track identified by the human eyes is integrated with the DBSCAN by providing the two indexes, namely the sequential angle offset and the sequential distance offset, the disadvantage that a traditional algorithm cannot identify sequential features is overcome, and the accuracy of behavior track chain-cluster identification is improved.

The invention discloses a trace point clustering method for clustering trace points of radar targets, and belongs to the field of radar signal processing. The trace point clustering method for clustering the trace points of the radar targets can be used for detecting and recording targets of low-resolution pulse search radars. By using the trace point clustering method for clustering the trace points of the radar targets, a radar detecting video is operated, morphological characteristics of echoes of the radar targets are sufficiently considered, the trace points serve as basic units, a basic principle that 'if a trace point belongs to a known target, another trace point in the next adjacent direction also belongs to the known target, and the distance between the trace point and the known target is equal to the distance between the other trace point in the next adjacent direction and the known target' is followed, a series of phenomena that new targets occur, trace points of the same target are broken, and the shapes of echoes are special are avoided according to distribution of the trace points on the periphery of a testing point, and whether the testing point belongs to a known target or not is judged. By using the trace point clustering method for clustering the trace points of the radar targets, the algorithm complexity is reduced during traditional trace point clustering, echoes of the two targets which are adhered to each other are distinguished into two trace points, the processing speed is increased, and the processing precision is improved.

The invention relates to a point-to-point calling method in cluster system based on LTE technology, a terminal thereof and a system thereof. The method comprises the following steps: through a base station, a core network, an IMS domain and an application server, a source terminal establishes a conversation channel with a target terminal; through the conversation channel, transmission of voice data with the target terminal is carried out. Based on a present air interface protocol, through not changing anything as possible or carrying out function expansion as small as possible, a point-to-point cluster communication mechanism is completed, by employing IMS+PoC configuration, function expansion is carried out on partial network elements in present network configuration, and through the base station, the core network, the IMS domain and the application server, the conversation channel between the source terminal and the target terminal is established. Transmission of the voice data with the target terminal is carried out through the conversation channel, a point-to-point service of the cluster system in an LTE network is realized, and by utilizing the cluster system, compatible and fusion with a present LTE network are realized.

The invention discloses an image splicing method based on feature point clustering four-fork division and local transformation matrixes, and belongs to the field of image processing and computer vision. SIFT (scale-invariant feature transformation) features of two images to be spliced are extracted and matched, one image is selected as a reference image, and matched feature points extracted from the other image are clustered by a k-means method. The non-reference image is subjected to layer-by-layer four-fork division according to clustering results, so that each subspace only contains one type of feature points. The transformation matrix corresponding to each subspace divided in the non-reference image is obtained by an MDLT (mobile data link terminal) method, each pixel value of each subspace divided in the non-reference image is subjected to coordinate transformation by the transformation matrix, and the non-reference image is spliced with the reference image in an aligned manner tofinally obtain a result map. By the aid of inherent properties and distribution conditions of the feature points, the number of the transformation matrixes is greatly decreased, a picture conformingto human eye vision cognition is still obtained, and the calculation efficiency of the splicing process is integrally improved.

A method for segmenting the data of a 3D sensor produced in the presence of aerosol clouds and for increasing the situation awareness and the location detection of obstacles involves transforming the sensor data are transformed into a 3D measurement point cloud, determining related subsets as measurement point clusters from the 3D measurement point cloud of an single measurement cycle of the 3D sensor based on the local measurement point density, determining at least one of the characteristic parameters of the individual measurement point clusters, the characteristic parameters including position, orientation in space, and shape, and determining a time variation of the characteristic parameters using the recorded parameters calculated from subsequent measurement cycles, from which the association of a measurement point cluster with a real obstacle or with the aerosol cloud results.

The invention provides an interest point processing method and device based on artificial intelligence, electronic equipment and a storage medium. The interest point processing method based on artificial intelligence comprises the following steps: performing named entity identification on an address included in acquired interest point data to obtain a candidate region name; when the candidate region names meet the screening condition, determining the candidate region names as to-be-screened region names; determining interest point data including the names of the areas to be screened, and determining longitude and latitude coding areas according to longitude and latitude positions included in the interest point data; clustering the interest point data related to the longitude and latitude coding area to obtain an interest point cluster; determining a to-be-screened region name meeting an occurrence frequency condition in the interest point cluster as a cluster region name; determining acentral longitude and latitude position of the interest point cluster and target interest point data; and establishing an association relationship between the cluster region name and the target interest point data. Through the invention, the manual labeling cost can be reduced, and the interest point processing efficiency is improved.

The present invention discloses a method for determining the mineral content represented by the entire SEM-EDS dataset, including initially unknown data points. SEM-EDS data points are taken and compared to a set of known data points. Any data point that is not sufficiently similar to the known data point is classified as unknown and clustered with like unknown data points. After all data points are analyzed, any clusters of unknown data points with a sufficient number of data points are further analyzed to determine their characteristics. All clusters of unknown data points with an insufficient number of data points to allow further analysis are considered outliers and discarded.

The invention belongs to the field of medical image processing, and particularly relates to a DSA vascular image segmentation method based on SIFT feature point clustering and a Boolean different operation. The method comprises the following steps: corresponding DSA mask images and live images are inputted; under the condition of a same threshold value, geometric feature points of the mask images and the live images are extracted by utilizing an SIFT algorithm; adjustment of local position of the geometric feature points is performed by adopting a method based on gray gradient values; the geometric feature points clustering and the Boolean different operation are performed based on Euclidean distance; and vascular image segmentation is performed by adopting the method based on an Otsu threshold value. The process of obtaining vascular images of patients in DSA subtraction has advantages of being high in robustness and convenient in calculation process so that accurate vascular image data can be provided to clinical operations based on DSA interventional therapy.

The present invention provides a micro-facial expression detection method. The method comprises: extracting N feature points from each image frame in a to-be-detected image sequence and aligning human faces; selecting K key feature points from the N feature points and dividing the K key feature points into M feature point clusters; selecting to-be-detected clusters from the M feature point clusters; selecting a base frame; calculating key point vectors of the to-be-detected clusters in each image frame; taking the key point vector of the base frame as a base vector, and calculating a deformation vector of each image frame; taking D times a maximum deformation vector as a deformation threshold; adding all image frames with the deformation vectors greater than the deformation threshold into the quasi-micro-facial expression frame sequence; and when there are consecutive frames in the quasi-micro-facial expression frame sequence, and the number of frames is greater than or equal to the preset frame threshold, taking the continuous frames as the micro-expression frame sequence. According to the method provided by the present invention, the micro-facial expression required to be detected is detected from the to-be-detected image sequence, and the recognition capability of the micro-facial expression and the robustness of the detection method are greatly improved.

The invention provides a method and a system for measuring eye movement fixation points based on density. The method comprises a trajectory point classification stage and a trajectory point clustering stage; at the trajectory point classification stage, the eye movement trajectory point data of a test picture are input, and the trajectory points belonging to fixation points are classified into different categories; the trajectory point clustering stage comprises executing each category, calculating the Euclidean distances between the trajectory points to form a distance matrix, counting the accumulated concern time density of each point, confirming a serial number set of trajectory points having the density equal to the maximum density value, and calculating the central position of the fixed points by solving the center of mass of the set of trajectory points having the density equal to the maximum density value. The clustering center generated by use of the method is more fitted with a user concern center and the emphasis of user concern is taken into full account; besides, the method is simple and efficient, and has higher anti-interference capability.

A method for counting high density population based on an SURF characteristi comprises the following steps: 1) collecting real time monitoring data by a camera, and pretreating images; 2) extracting feature points of moving people from the pretreated images; 3) feature points clustering: using an MST-DBSCAN algorithm which is based on a traditional DBSCN algorithm; 4) counting of people flow: according to the result of clustering in step 3), calculating a real time feature vector T of the crowd by (4), obtaining (5) using a support vector regression, and providing a predicted number of people by (5). The invention provides an effective method for counting high density population based on the SURF characteristic, and the method is suitable for scenes that have a large flow of people or a high density population.

The invention discloses a method for recognizing a clamping plate furcal hole through an intelligent forklift. The method comprises the steps that a laser scanner acquires the data of the surrounding environment of a fork arm, and passes the environment data back to a software program in the intelligent forklift in real time; the software program analyzes the received environment data, extracts relevant characteristic values of a clamping plate, carries out wave filtering on consecutive points, and filters discrete points; interference point clusters in point clusters generated after the consecutive points are subjected to wave filtering according to the relevant characteristic values of the clamping plate are filtered, cargo pier sets are found, and the positional relationship among the cargo pier sets is analyzed; and the position of the furcal hole of the clamping plate is found and tracked, and the intelligent forklift is guided to complete the furcal hole entering action. The invention further relates to a device for achieving the method. The method and the device for recognizing the clamping plate furcal hole through the intelligent forklift have the beneficial effects that the furcal hole entering accuracy of the intelligent forklift can be improved, and the intelligent forklift can be automatically guided to complete the furcal hole entering action.

The present invention relates to a method for scheduling an access point of a wireless local area network. The method comprises: within a current scheduling cycle of a network running state, a first controller receiving statistical performance data reported by a plurality of access points in a first access point cluster within the last scheduling cycle, wherein the first access point cluster comprises the first controller and a plurality of access points connected to the first controller; according to statistical performance information, generating a scheduling solution of the plurality of access points within the next scheduling cycle; and broadcasting the scheduling solution to the plurality of access points, for configuring a sending parameter within the next scheduling cycle by the plurality of access points within the current scheduling cycle according to the scheduling solution, and communicating with a user equipment according to the sending parameter, wherein the sending parameter comprises an antenna direction, an upper limit value of a transmission power, and a threshold value of an estimated parameter of an idle channel. By means of the embodiments of the present invention, that the effect of different access points on other access points when communicating with the user equipment (UE) is reduced is realized in a local area network with a plurality of access points, thereby improving the access rate of the wireless local area network.

The invention provides a feature point matching method for a close-range shot stereoscopic image. The feature point matching method for the close-range shot stereoscopic image comprises the steps that feature point positive matching is conducted on homonymous images with subregions as constrain conditions and with triangles as constraint conditions in sequence so that a positive homonymous image point cluster can be obtained, feature point reversal matching is conducted on the homonymous images with the subregions as the constrain conditions and with the triangles as the constraint conditions in sequence so that a reversal homonymous image point cluster can be obtained, matched homonymous image point results in the positive homonymous image point cluster and in the reversal homonymous image point cluster are reserved, and a final matched homonymous image point result is obtained. According to the feature point matching method for the close-range shot stereoscopic image, feature point detection is conducted with the subregions as the constraint conditions and with the triangles as the constraint conditions in sequence, the time efficiency of feature point detection is greatly improved, the number of correct matched points is improved, and the feature point matching method has high application value on the aspect of close-range shot stereoscopic image matching.

The invention relates to a biological sample intelligent identification method based on a molecular map. An existing classification method is limited to a limited biological sample set, a universal technology and means capable of analyzing biological samples of different types and different sources in a high-throughput and high-accuracy mode are lacked, and the classification and recognition requirements of biological large samples are difficult to meet. The invention provides an intelligent identification method based on a molecular map. According to the invention, spatial characteristics such as point clusters and shapes hidden in a biological sample molecular map are utilized to carry out high-efficiency identification on the biological sample. Points with high density are clustered into point clusters through clustering, high-precision scanning and matching between unknown samples and standard samples are carried out through fine space information of the point clusters, and recognition is carried out by comparing the matching degrees; the result shows that the classification effect is obviously superior to that of the reported method based on the spatial characteristics contained in the molecular map.

The invention discloses a method and system for electronic commerce user purchasing power classification. The method includes the steps that based on price and sale volume distribution, the level of commodities of the same type is determined to be the high level, the level of other commodities of the same type is determined to be the non-high level, and the commodities at the non-high level are divided into x levels from high to low sequentially according to the price from high to low, wherein x is a preset natural number larger than or equal to 1; the ratio of the commodities at each level in overall commodities purchased by each user is calculated, the ratios are vectorized based on all the users, purchasing power vectors of all the users are acquired and are x+1 dimension vectors, and each dimension corresponds to one level; the purchasing power vectors are clustered, x+1 point clusters related to the purchasing power vectors are acquired, each point cluster corresponds to one level, and the level corresponding to the point cluster where the corresponding purchasing power vector of each user is located serves as the purchasing power level of the user. By the adoption of the method and system, user purchasing power classification is more accurate.

The invention provides a modulation format recognition method, which adopts high-order cumulant of signals in a Stokes space and clustering center points clustered based on a DENCLUE algorithm as characteristic parameters for recognizing modulation format signals; a simple decision tree classifier is adopted; the method has low computational complexity and is insensitive to optical fiber channel damage such as polarization rotation, carrier frequency deviation, carrier phase noise and the like; recognition with relatively high accuracy can be realized under the condition of low signal-to-noiseratio; the modulation format recognition method is high in fusion degree with an existing commercial optical fiber communication system and is easy to upgrade from an existing system.

The invention provides a topology segmentation method and device oriented to a closed space, the method comprises: acquiring target point cloud data of a closed space to be processed; and extracting point cloud data, complying with a wall characteristic, from the target point cloud data by a density projection and characteristic value method; performing 3D to 2D projection on the point cloud datacomplying with the wall characteristic, rasterizing the projected data to extract multiple line segments; clustering the multiple line segments by an iterative clustering method to obtain the contourof the closed space to be processed; clustering the grids in the contour by using a point clustering method with multi-directional scanning based on the contour, to obtain a topological segmentation map of the closed space to be processed. The method improves the accuracy of topology segmentation, can effectively segment the closed space, and solves a technical problem that an existing topology segmentation method has high error rate, cannot achieve effective segmentation, and cannot determine the topological space of any location in the space.