53results about How to "Improve association accuracy" patented technology

The invention provides a text information associating and clustering collecting processing method based on a domain knowledge model. The method comprises the following steps that a text information training set is searched, stemming preprocessing is conducted, and feature word vectors of a text participle sequence of the information training set are extracted through Chinese named entity identification and domain dictionary query modes; representative feature words of a target event are extracted through topic graph model learning training, and a weighted value of topic associating affiliation is calculated; a feature word set is built according to the topic associating affiliation weighted value, calculated through training, of the feature words, and an event topic word template is built; feature word vectors of a participle sequence accessed to text in real time are extracted through the Chinese named entity identification and domain dictionary query modes; the similarity distance of the feature word vectors and all the target event knowledge templates is calculated; the association relationship of multiple texts to the same topic target event is determined according to the similarity threshold, and classification reorganization is conducted by means of a similarity distance ordering rule.

The invention discloses a method for linking accounts in a plurality of OSNs (On-line Social Networks) on the basis of node similarity. The friend relationship is extracted from OSN website accounts for linking the plurality of accounts, belonging to one real user, in different OSN platforms. The method comprises five parts: a data preprocessing part performs preprocessing on an account node relationship diagram of the OSNs; a node sequence extraction part obtains an account node sequence set through random walk; an account vector representation part generates a vector model of each account through a word to vector tool word2vec; a linear conversion matrix calculation part obtains a linear conversion matrix W from one OSN to another OSN by a gradient descent method; a linking account obtaining part maps the account in one OSN to a coordinate space of another OSN; and through similarity measure and threshold value screening, the linking accounts or a candidate set corresponding to all of the accounts are obtained. Errors of the account linking result caused by unreal feature information of the accounts are avoided, so that the account linking robustness is improved.

The invention discloses a flight path association method and an electronic device. The method comprises the following steps: acquiring N sets of sensing data for representing the flight path of the flight equipment by M sensors arranged on the flight equipment, wherein M is an integer more than or equal to 2 and N is an integer more than 0; dividing the N sets of sensing data into common data and identity data, wherein the common data are the data collected by M sensors and the identity data are used for representing the type of M sensors; and respectively utilizing the common data and the identity data to associate the flight path with the path in a path library, wherein at least one flight path of at least one flight equipment is included in the path library. The technical scheme provided by the invention is used for solving the technical problem of insufficient utilization for the identity information of the path association in the prior art.

The invention discloses a weighted nearest-neighbor data association method for a centralized multi-radar data processing process. The scores of seven parts including a gate, a course angle/angular velocity, attribute similarity, accelerated speed, response times, vertical distance and track status are associated, and the scores of the seven parts are subjected to the calculation of trace point and track relevancy based on the weighted scoring algorithm. In this way, the updated value of a final score is obtained. During the score associating process, after the calculation of trace point and track relevancy, the smaller the updated value of the final score of the associated values, the stronger the matching relationship. When the updated value of the final score of the associated values is larger than 0x9FFFF, the matching requirement is not met. In the condition of limited hardware resources and high requirement on real-time performance, the above data associated processing method is relatively efficient and accurate. Meanwhile, on the basis of an original nearest-neighbor associated processing method, other associated reference factors are introduced at the same time. Therefore, an optimal matching group between trace points and the track can be gradually found out.

The invention discloses a multi-target association method and device. The method includes obtaining azimuth observation sequences corresponding to targets under a first observation platform and a second observation platform; performing track invariant information estimation by utilizing a Kalman filtering algorithm and the azimuth observation sequences so that course information of a target can beobtained; and utilizing the azimuth observation sequences and the track invariant information to calculate degrees of association, performing ordering on the degrees of association, determining the association of the track of the target observed by the first observation platform and the track of the corresponding target observed by the second observation platform when the largest degree of association is larger than a preset threshold value, and outputting association results. Compared with traditional association algorithms only using azimuth information, the method and device can have higher association accuracy.

The invention provides a document and label word semantic association method and a device thereof.The method comprises the following steps that S100, documents are obtained and serve as a document corpus, wherein the documents belong to the field relevant to label words and have timeliness; S200, a syntactic pattern is constructed, syntactic pattern matching is conducted on the document corpus, and results according with the syntactic pattern are merged to serve as a candidate mention relevant word set C; S300, the candidate mention relevant word set C is filtered through a trained Word2vec pattern, and a mention relevant word set V is obtained; S400, according to mention relevant words and the label words, relevancy degrees of multiple documents and the label words are calculated, and a relevancy degree database of the label words and the documents is constructed.The document and label word semantic association method dynamically constructs semantic association on the basis of the real-time document corpora and does not depend on a synonym base which is static and high in maintenance cost.

The invention discloses a nuclear K-mean value track correlation method based on KMDL criteria. The method comprises the following steps that firstly, a typical track correlation scene is established; secondly, the number of target tracks is determined based on the KMDL criteria; thirdly, the nuclear K-mean value algorithm is used for conducting correlation on the track observation scene. According to the nuclear K-mean value track correlation method based on the KMDL criteria, the KMDL criteria and the nuclear K-mean value algorithm are combined to solve the multi-target track correlation problem under the complex environment where clutters are dense, targets are near and the number of the targets is not known. According to the method, the movement state information of the targets is fully utilized, the correlation accuracy rate is effectively increased, the correlation criteria are simple and practicable, the calculated amount is small, the correlation accuracy rate is high, it is not sensitive for target intersection, and the method is suitable for track correlation under the environment where the targets are dense and intersect and applicable to engineering realization.

The invention provides a method and device for searching for an enterprise logo and a computer readable storage medium, and relates to the technical field of computers. The method comprises the stepsof acquiring a target associated enterprise associated with a current enterprise; performing word segmentation processing on an enterprise name of the current enterprise to obtain a first processing result, and performing word segmentation processing on an enterprise name of the target associated enterprise to obtain a second processing result, wherein the first processing result includes brand information of the current enterprise, and the second processing result includes brand information of the target associated enterprise; comparing the first processing result with the second processing result to determine the enterprise logo information of the current enterprise according to a comparison result. The technical problem that there is difficulty in obtaining the enterprise logo information of enterprises without exclusive logos by using the prior art is solved.

The invention relates to a site enforcement recorder and mobile terminal backward code scanning bundled ternary alliance uploading system. The system comprises a site enforcement recorder, a data acquisition workstation, a concentrated storage server, a response alarm platform server, a mobile terminal, a police service private network server, an isolation gatekeeper, a ternary alliance server anda management computer, wherein the police service private network server, the data acquisition workstation, the concentrated storage server, the response alarm platform server, the ternary alliance server and the management computer are connected through a private network; the mobile terminal is wirelessly connected with the police service private network server through a mobile police service wireless private network; and the police service private network server is connected with the private network through the isolation gatekeeper. According to the system, audios, videos and photo data ofthe site enforcement recorder can be automatically associated to a response alarm platform and automatic uploading of recorded data is realized, so that the intelligence degree is improved, the workload of front-line policemen is lightened, the association efficiency is improved, the association correctness is enhanced and the law enforcement and case handling are standardized.

The invention discloses a low and slow small-target track filtering method based on interactive multi-model-Kalman filtering. The method comprises the steps of 1, obtaining a target initial state estimated value X0 and an initial state estimation covariance matrix P0 through a track initial algorithm; 2, setting a target motion model set, and obtaining a corresponding state transfer matrix and a radar measuring matrix according to a motion characteristic; 3, calculating a model predication probability and mixing probability which correspond with a time point k; 4, calculating the mixing stateand the mixing covariance at the time point k; 5, calculating the one-step predication value and the predicated covariance matrix of each model mixing state at a time point k+1 and the measured one-step predication value and the measured predicated variance matrix; and 6, calculating the target state estimated value and the estimated covariance matrix at the time point k+1, and calculating a likelihood function and a model probability; adding one to the value of k, and executing the step 3 again. The method improves point trace association success rate and finally improves radar detecting precision.

The invention relates to an IMSI and face depth fitting method, device, equipment and system. The method comprises the following steps: acquiring face equipment data and electric surrounding base station equipment data; performing position analysis on the geographic position of the face acquisition equipment and the geographic position of the electric surrounding equipment, and determining a spatial association relationship between the face acquisition equipment and the electric surrounding equipment; determining a time association relationship between the personnel and the IMSI according to the face data and the IMSI data; obtaining a space-time association relationship between the personnel and the IMSI according to the space association relationship and the time association relationship; carrying out deep fitting processing on the space-time association relationship by utilizing a pre-trained score increase model and a score attenuation model to obtain an association relationship weight score of the personnel and the IMSI; and determining a depth fitting relationship between the personnel and the IMSI according to the association relationship weight score. According to the method, the association accuracy of the personnel and the IMSI is greatly improved, and the association relationship can be updated in time.

The invention relates to a multi-target tracking algorithm based on multi-feature matching, which comprises the steps of firstly, establishing a feature matrix of all targets in allusion to a problemthat the number of matching features is too small when a measuring point is associated with a target, wherein the feature matrix includes test statistic features and Doppler frequency features of alltargets; secondly, comparing features of all association points with the feature matrix of the target when the optimal association point is searched so as to find out an association point with the highest matching degree with the target; and finally, determining the motion state estimation value, test statistics and Doppler frequency features of the target at the current tracking moment to estimate the motion state of the target after the current tracking moment. The multi-target tracking method solves a problem of inaccurate target trajectory tracking when the trajectories of multiple targetsare intersected and overlapped through establishing the feature matrix and finding the association point which is the most matched with the feature matrix of the target.

The invention discloses an extended target tracking method based on an automobile radar. The method comprises the following steps: establishing a random finite set model of an actual target automobileaccording to a state set; establishing a random finite set model of a measurement target automobile according to the measurement set; correcting abnormal Doppler velocities; after azimuth-Doppler velocity pre-partitioning processing, using Gibbs sampling for data association; under the condition of successful association, updating the posterior density of multi-target PMBM; predicting the multi-target PMBM prior density at the next moment; determining whether a new target enters under the condition of unsuccessful association, predicting the multi-target PMBM priori density of the new targetat the next moment; calculating an output result by combining a complete state vector; otherwise, removing the clutter. Therefore, the tracking efficiency and the tracking precision are substantiallyimproved by increasing a small amount of time complexity.

The present invention provides a multi-sensor data fusion method, a device and a vehicle. The invention aims to solve the problem of low association accuracy of the data of a plurality of sensors in the prior art. According to the multi-sensor data fusion method, the device and the vehicle, a plurality of sensors are used for detecting data of a plurality of dimensions of objects in the same area.The method includes the following steps that: a corresponding data interval is set for each dimension of a first object according to the data of a plurality of dimensions of the first object detectedby a first sensor, wherein the first object is any object detected by the first sensor; whether a unique second object exists in all objects detected by a second sensor is judged, and the data of each dimension of the second object are all in the data intervals set correspondingly set for the first object; and if the unique second object exists in all the objects detected by the second sensor, and the first object and the second object are associated to correspond to the same object in the area.

The invention discloses a simultaneous scanning and framing X-ray measurement system. The system comprises a dual-channel X-ray imaging system, a transmission type X-ray coding plate, an X-ray streakcamera and a decoder connected with the X-ray streak camera. The simultaneous scanning and framing X-ray measurement system can simultaneously record a one-dimensional continuous change process imageand a two-dimensional framing image of a to-be-measured target by adopting the same recording system from the same visual angle, has the advantages of compact structure, small occupied space and hightime correlation precision of the scanning image and the framing image, and has a wide application prospect in laser inertial confinement fusion physical experiment research.

The invention provides a traffic object fusion association method and device in a driving environment and edge calculation equipment, and relates to an intelligent traffic technology and an automatic driving technology. The method comprises the steps: obtaining first environment information and second environment information; the first environment information and the second environment information are collected through different sensors; determining first lane information of a first traffic object in the first environment information according to the first environment information, and determining second lane information of a second traffic object in the second environment information according to the second environment information; According to the information of the first traffic object in the first environment information, the information of the second traffic object in the second environment information, the first lane information and the second lane information, determining whether the first traffic object and the second traffic object have an association relationship. According to the scheme provided by the invention, the information of the lane where the traffic object is located can be combined to determine whether the traffic object has the association relationship, so that the fusion association accuracy of the traffic object can be improved.

The invention provides a calculation method and device for the association degree of words and a web page.Word segmentation and pretreatment are performed to the title of the web page and the text content, headwords and text words are used for constructing a word connection set to be used for calculating the TextRank score of the words, and the TextRank score is taken as the association degree of the words and the web page and saved in a database.According to the method, the TextRank score is taken as the association degree, the association relationship of the words and the web page can be effectively reflected, the words of the title are used for amending the word connection set constructed on the basis of the text words, and the amended word connection set is used for calculating the TextRank score of the words, the significance of the title in the web page information is fully considered, and the association precision can be improved favorably.