81results about How to "Global" patented technology

The invention discloses a self-adaptive identification method and system wherein the system comprises: according to a user's historical corpus, establishing a customized dictionary; clustering the customized words in the user's customized dictionary to obtain the type number of each customized word; constructing language models according to the type numbers of the customized words; during the identification of the inputted information of the user, if the words contained in the information exist in the user's customized dictionary, expanding the de-coding paths according to the type numbers of the customized words corresponding to the words; decoding the information according to the expanded decoding paths to obtain a plurality of selective decoding results; according to the language models, calculating the scores of the language models of the selective decoding results; and selecting the selective decoding result whose language model obtains the highest score as the information identification result. With the method and system of the invention, it is possible to increase the identification accuracy for a user's customized words and the consumption of the system can be reduced.

The invention relates to a human face expression identification method, in particular to a convolutional neural network-based human face expression identification method. The method comprises the steps of firstly obtaining human face images from a video, performing scale normalization operation processing on the obtained human face images to obtain human face images same in size, and performing alignment preprocessing operation on the human face images same in size to obtain preprocessed human face images; and performing feature extraction operation on the preprocessed human face images by using a convolutional neural network to obtain features of the human face images, and performing classified identification operation on the features of the human face images by utilizing a Softmax classifier. A human face expression identification algorithm realized by utilizing the convolutional neural network is an end-to-end process. According to the method, the human face images only need to be simply preprocessed and then fed into the convolutional neural network, the feature extraction is automatically performed, and a classification result is given, so that the accuracy is greatly improved, adjustable parameters are reduced, and intermediate processing steps are simplified to a great extent.

The invention discloses an intelligent identification method for electrocardiogram data based on a residual network. The intelligent identification method comprises the steps of S1, according to the Rwave position, intercepting a map in front and behind the heartbeat signal R wave position of an electrocardiogram, and dividing the map into a training set and a test set proportionally; S2, constructing a deep convolutional neural network structure with a residual structure, and adding dense connection based on the original residual network; S3, inputting the training set into the network structure for iterative training, and conducting iteration continually until the layer of the network structure is increased to 152; S4, saving a model and testing the output. The intelligent identification method has high feature level and overall importance, and achieves intelligent identification of the electrocardiogram data, and the recognition accuracy is high.

The invention discloses a load balancing method based on vector mapping. The method comprises the following steps of: obtaining the load information of physical nodes and virtual machines in a virtual cluster system; obtaining the physical nodes, for which the load condition needs to be regulated, according to the resource information, and calculating a balance degree to obtain the virtual machines to be migrated; converting and normalizing the three-dimensional resource of the virtual machines to be migrated and all the physical nodes, representing the converted and normalized resource in a space coordinate system in the form of three-dimensional vector; projecting the space coordinate system and the resource vector, grouping the projection vectors, and obtaining the group match of the physical nodes and the virtual machines; and for each group match, scheduling with an optimal descending order bin packing virtual machine to get the decision result. The technical scheme of the invention is used in a distributed cluster system, is supportive to differentiated, global and expandable scheduling of multi-dimensional resource, and is featured by high speed of decision making and high accuracy.

The invention discloses a video keyframe extracting method and a video keyframe extracting system. The method comprises the following steps of reading video data which consists of a plurality of picture frames; extracting candidate keyframes from the video data to obtain a candidate keyframe set; classifying the candidate keyframes in the candidate keyframe set to obtain at least two clustering sets; and extracting the keyframe from the clustering sets. By the mode, the video keyframe extracting method and the video keyframe extracting system have global superiority, the computing speed is high, and actual application demands are met.

The invention discloses a method for realizing a KLT (Karhunen Loeve Transform) moving target tracking algorithm based on a multicore DSP (Digital Signal Processor). The method disclosed by the invention comprises the steps of: image smoothing, pyramid graph calculating, image gray value difference calculating, feature point selecting, feature point tracking and feature point consistency checking; and six modules for executing the steps are transplanted into the multicore DSP, wherein one core is allocated to each module, and the processing is carried out in a datastream manner. According to the method, a KLT tracking algorithm is adopted, and the advantage and processing speed of multicore parallel execution of the multicore DSP are used, so that the real-time and accurate tracking of moving targets can be realized.

The invention discloses a mycobacterium M. tuberculosis holoprotein chip and an application thereof. The protein chip has at least one of the following functions (1) to (7); the following 4,168 proteins are connected to the protein chip, and each protein independently forms a detection point. Experiments show that the mycobacterium M. tuberculosis holoprotein chip consists of a substrate and proteins containing more than 95 percent of mycobacterium M. tuberculosis genome codes. The holoprotein chip is obviously significant for looking for a serum biomarker, a substrate interacting with a micromolecular compound c-di-GMP, a kinase substrate interacting with protein kinase and a substrate interacting with macrophage. Compared with an existing research means, the holoprotein chip has global and high-throughput screening characteristics, research design and operation processes are simplified, and the research efficiency is remarkably improved.

The invention provides a distributed high concurrent real-time message pushing method and a distributed high concurrent real-time message pushing system. The method comprises the steps of building a message connection between a first user and a web server, receiving a message sent by a second user to the web server after the second user logs in a web, verifying whether the message is valid, if no,ending an illegal access, if yes, performing high speed concurrent treatment on redis data, and performing kafka partition message treatment; synchronously consuming a kafka message via the web servers, and synchronously pushing same data via the multiple web servers to corresponding grouped users of a websocket connection, and receiving the message sent by the web server via the first user. According to the method and system provided by the invention, the retardance of the websocket instant messaging under the high concurrent scene can be reduced, and the expandability of the same can be improved.

Aiming at a problem that measurement and evaluation of network service and application are currently in a rough comparison state and evaluation indicator systems and accurate measurement index values are unavailable, the invention discloses a network safety evaluation method based on attack-defense utility. The method includes steps of 1, determining scene differential manifold; 2, calculating a behavior path; 3, calculating behavior utility; 4, performing network attach-defense evaluation based on behavior utility.

The invention discloses a convolutional long-term and short-term memory network space-time sequence prediction method improved by utilizing an attention mechanism. The method relates to the field of computer prediction, and specifically comprises the following steps: (1) extracting spatial features by an asymmetric convolution block high-dimensional feature extractor; (2) a ConvLSTM encoder decoder architecture embedded with an attention module predicting extrapolation features; (3) reversely reconstructing a feature result; (4) L1 and L2 regularization optimization algorithms being carried out; and (5) predicting a space-time sequence image. According to the method, high-dimensional features of spatio-temporal sequence data can be well extracted through the multi-layer convolutional neural network, and the high-dimensional features are used as input of the model, so that the problem that high dimensions cannot be calculated is solved, and spatial key information is emphasized; according to the improved ConvLSTM, the spatial and temporal features can be better learned to realize more accurate extrapolation; the method is suitable for all sequential images.

The invention provides a measurement device for surface roughness, comprising a measurement mechanism and a measurement workbench, wherein the measurement workbench consists of an upper-layer rack and a lower-layer rack, each layer of the rack comprises a linear guide sliding pair, each linear guide sliding pair comprises a fixed guide rail and an air-floating guide rail, each air-floating guide rail consists of an upper air-floating block, two side air-floating blocks fixedly connected with two sides of the bottom of the upper air-floating block and two bottom air-floating blocks which are fixedly connected with the bottoms of the two side air-floating blocks and oppositely extend respectively, the inner surface of each air-floating guide rail consists of air outlet surface of each air-floating block, air outlet holes are arranged on the air outlet surfaces which are opposite to the fixed guide rails, and the air outlet holes of the air-floating blocks are communicated with one another in the air-floating guide rails by gas passageways and jointly use one air inlet hole; and the invention further provides the measurement workbench. The air-floating guide rails formed into the measurement workbench can move along with the fixed guide rails without friction, so that the problem that the measurement workbench in the prior art is slow in response speed, low in moving speed and low in control precision can be solved.

The invention discloses a two-dimensional code anti-fake method and system based on a domain name. Two-dimensional codes of enterprise domain name information, commodity information and passwords are generated and printed on anti-fake tags arranged on commodities, the two-dimensional codes of the anti-fake tags arranged on the commodities are shot by handheld intelligent terminals, a first level of anti-fake protection is formed by judging whether the two-dimensional codes are matched or not, a second level of anti-fake protection is formed by shifting into a domain name website webpage which the production enterprise of the consumed commodities belongs to through the two-dimensional domain name information on the anti-fake tags, and a third level of anti-fake protection is formed by verifying the true commodity snapshots obtained on the webpage through verification codes on the anti-fake tags and obtaining the commodity information. Thus, the authenticity of the commodities is judged, and the most direct, convenient and effective method is provided for consumers to identify the authenticity of the commodities.

The invention provides a planar three-degrees-of-freedom space robot inverse kinematics solution method. First, a general planar three-degrees-of-freedom space robot is equivalent to a ground fixed base robot using a virtual mechanical arm method; then, the self-motion characteristics of the planar three-degrees-of-freedom space robot are analyzed, and the geometric relational expression between joint angles of the equivalent space robot is deduced under the condition that the initial base attitude angle and the end manipulator position are given; and finally, all combinations composed of three joint angles, namely, all kinematics inverse solutions, are obtained. Through the method, all kinematics inverse solutions can be calculated out simply and efficiently under the condition that the position of an end manipulator of a planar three-degrees-of-freedom redundant space robot is given, and thus, robot path planning problems related to obstacle avoidance, singular avoidance, joint limit avoidance and other optimization indexes can be studied.

The invention discloses a stadium marker line-based method for determining camera exterior parameters, which comprises the following steps of: selecting the marker line from the shot stadium image; acquiring the first coordinate of the mark line on the actual plane of the stadium and the second coordinate of the marker line on the shot stadium image; generating the projection mapping relationship between the marker line on the actual plane of the stadium and the marker line on the stadium image according to the first and second coordinates; calculating the projection mapping relationship between the point on the actual plane of the stadium and the point on the stadium image according to the projection mapping relationship of the marker lines; calculating the initial value of the current camera exterior parameter matrix according to the projection mapping relationship of the points and the camera interior parameters acquired in advance; and optimizing the initial value of the current camera exterior parameter matrix to acquire the final camera exterior parameter matrix. According to the method and system, the problem that a point cannot be selected actually can be solved and the accuracy of camera exterior parameter calibration can be guaranteed.

The invention discloses a flight test and ground simulation aerodynamic force data comprehensive modeling method, which is used for establishing a full-airspace and full-speed-domain aerodynamic forcemathematical model of an aircraft according to more than 10 flight test data samples and aerodynamic force ground simulation data in a flight envelope range of the aircraft. The method comprises thefollowing steps: firstly, performing aerodynamic force calculation ground simulation within a flight envelope range of an aircraft, obtaining aerodynamic force simulation data in a full airspace and afull speed domain, and converting the aerodynamic force simulation data into an aerodynamic force simulation sample; And performing a flight test, obtaining more than 10 times of flight test state measurement data, and obtaining an aerodynamic force identification data sample of the flight test according to an aerodynamic force identification method. And finally, taking the aerodynamic force identification data sample and the aerodynamic force simulation sample together as a total sample, and establishing an aircraft full-airspace and full-speed-domain aerodynamic mathematical model integrating flight test and ground simulation data through machine learning by utilizing an artificial intelligence-based BP neural network method.

The embodiment of the invention provides a crop canopy in-situ growth phenotype monitoring device and a three-dimensional reconstruction method. Through a data acquisition and collection module arranged on a canopy top, an image acquisition module arranged in a canopy, and a fisheye lens arranged on a canopy bottom, the morphological structure data of the top, the interior and the bottom of the canopy of the crop to be detected can be obtained, and the height and the angle of the image obtaining module in the canopy can be adjusted, so that the morphological structure data in the canopy are richer, and finally, the growth monitoring of the canopy of the crop to be monitored is more accurate and overall, and the universality is very good. The canopy growth monitoring device is used for obtaining the morphological structure data of the top, the interior and the bottom of the canopy, the morphological structure data are used for phenotypic analysis and three-dimensional reconstruction after certain pretreatment, the workload is small, the efficiency is high, and the obtained three-dimensional reconstruction model is high in precision.

The invention provides a method for identifying an action scale of land utilization fractal dimension based on a genetic algorithm. The method comprises the specific steps of firstly, interpreting and extracting land utilization classified information by using remote-sensing images, conducting statistics on land classified area information of buffer areas of different radiuses, and establishing logarithmic relation between the area and the radius of each class of land according to a fractal dimension calculation formula; and then analyzing a parametric coding mechanism of the genetic algorithm so as to generate an initial group; designing a fitness function on the basis of the initial group; and obtaining the optimal action scale of fractal dimension by using three basic genetic operators, namely a selection operator, a cross operator and a mutation operator, thus accurately illustrating an effective region of the land utilization fractal dimension. According to the method disclosed by the invention, the problem that a land utilization fractal method gives no consideration to scale effect caused by data source, such as topography and remote-sensing images, is solved, therefore, the optimization of solving of a non-scale region of fractal dimension can be implemented, and the obvious advantages of accuracy and efficiency in identification of the non-scale region are formed.

The invention discloses a method for identifying an event argument and an argument role based on maximum entropy. The method comprises the following steps of extracting the entity information of an event, so as to determine candidate options of the event argument and extract the argument information of the event; extracting the feature information of the candidate options of the event argument, and utilizing the feature information and the argument information to process data, so as to obtain an argument identifying corpus and an argument role identifying corpus; utilizing the argument identifying corpus to train a maximum entropy binary classifying model, so as to obtain the maximum entropy binary classifying model; utilizing the argument role identifying corpus to train a maximum entropy multi-element classifying model, so as to obtain the maximum entropy multi-element classifying model; utilizing the maximum entropy binary classifying model to identify the candidate options of the event argument in the event, so as to obtain the argument identifying result; utilizing the maximum entropy multi-element classifying model to identify the argument identifying result, so as to obtain the argument role identifying result. The method has the advantage that the identifying efficiency of the event argument and the argument role is improved. The invention also discloses a system for identifying the event argument and the argument role based on the maximum entropy.

The invention discloses a quantum-behaved particle swarm optimization (QPSO) recurrent predictor neural network (RPNN) method for financial time series prediction, which relates to analysis and prediction on a time sequence. Chaos and phase space reconstruction theories are firstly applied, through a saturation correlation dimension (G-P) method, a chaotic financial time series attractor dimension is calculated, and the structure of the RPNN is determined; then, the RPNN is trained by the QPSO algorithm; and finally, the dynamic optimal weight and the threshold of the network are determined, and thus, the simulation prediction value of the RPNN and the actual value can reach the minimum error precision. The problem that optimization of the RPNN based on a gradient algorithm is easy to fall into local minimum can be solved, and the built QPSO-RPNN optimization prediction method can be widely applied in financial investment and social economy, and has the advantages that the convergence rate is quick, the searching is global, the programming is simple and efficient, and the prediction precision is high.

The invention provides a decision-making assisting method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring business activity data; inputting the service activity data into a pre-established digital twinning model, performing analog analysis on the digital twinning model according to a time sequence relationship, and establishing the digitaltwinning model according to the corresponding service activity node; receiving an optimization direction selected by a user, and obtaining one or more optimization paths according to the optimizationdirection; receiving the selection of the user for the optimized path; according to the optimized path and a pre-established digital twinning model, obtaining a quantitative influence for executing the optimized path, wherein the quantitative influence is used for guiding a user to select a production and operation decision scheme; and when the production and operation decision scheme meeting theuser requirements is not obtained according to the current quantitative influence, repeatedly executing the step of executing the quantitative influence of the optimized path according to the optimized path and the pre-established digital twinning model until the obtained production and operation decision scheme meets the user requirements.

The invention relates to a sky image cloud cluster movement velocity computing method based on the phase correlation principle. The method comprises the following steps that 1, an initial image and a displacement image of a sky image are acquired; 2, gray matrixes of the initial image and the displacement image are generated respectively; 3, image frequency spectrums of the initial image and the displacement image are acquired through the two-dimensional Fourier transform; 4, cross-power spectrum of the initial image and the displacement image and an inverse Fourier transform response matrix of the cross-power spectrum are computed; 5, spike pulse coordinates of the response matrix are extracted to serve as cloud cluster displacement vectors; 6, the cloud cluster movement velocity is computed according to the cloud cluster displacement and the image time interval. According to the sky image cloud cluster movement velocity computing method based on the phase correlation principle, the operation flow is simple and direct, and the cloud cluster displacement prediction consuming time can be greatly shortened; the overall movement condition of a cloud cluster in the image can be recognized more effectively, and due to normalization processing in cross-power spectrum computing, the higher robustness is achieved for global image noise.

The embodiment of the invention discloses a face recognition method and device based on key point region feature fusion. The method comprises the following steps: carrying out data preprocessing on atraining set or a test set; constructing a convolutional neural network model; training the convolutional neural network model by using the preprocessed training set, extracting face key point regionfeatures, and fusing the face key point region features; and testing the convolutional neural network model by using the preprocessed test set. The extracted features of the face key point area are fused by using a convolutional neural network; the defect that the traditional face recognition method is sensitive to various uncertain interferences such as illumination difference, facial expression,occlusion and the like is overcome; the global features and the invariant features of the human face are comprehensively considered, the recognition model has higher robustness and accuracy, the relation between different features is enhanced, the extracted features are more perfect and complete, the human face information can be described more accurately, and the classification capacity of a classifier is improved.

The invention discloses an AUV (Autonomous Underwater Vehicle) three-dimensional airway planning method, which improves the planning speed and planning speed of an algorithm by improving a particle initialization method, setting a punishment value for entering an obstacle and setting the step length of an added particle moving position. The method mainly comprises the steps of three-dimensional environment abstract modeling, particle position initialization, punishment value setting of entering obstacles, cost function value calculation, particle movement step length setting, cost value comparison of a particle swarm, particle position updating and planning result output. Compared with a traditional quantum particle swarm airway planning algorithm, the three-dimensional airway planning method provided by the invention considers the influence of ocean current factors, so that the algorithm application is wider; meanwhile, compared with two-dimensional air route planning, three-dimensional air route planning is more practical, and the actual navigation requirement can be better met.

The invention provides a seepage analysis-based online social platform rumor spreading immune method. The method comprises the following steps of: 1, preprocessing practical microblog rumor data information; 2, aiming at preprocessed data information, establishing rumor spreading networks on basis of a relationship between transferring and transferred of users; 3, establishing a rumor coupling network according to a mutual relationship among co-participants of different rumor networks on the basis of existing rumor network information; and 4, carrying out seepage analysis on the rumor couplingnetwork from different angles, and comparing inhibition effect, for the rumor networks, of different seepage methods so as to provide an effective immune method. Through above steps, networks are established on the basis of rumor spreading data on practical online social platforms, the rumor coupling network is established according to the relationship among the networks, and seepage analysis iscarried out on the coupling network from different angles, so that assessment and quantification are carried out on fragility of the rumor network and powerful method support can be provided for further immune of rumor spreading of the online social platforms.

The invention provides a depth image tumbling recognition algorithm based on a wavelet moment. The algorithm comprises the steps of firstly normalizing the distances between all pixel points in an image and the centroid, then conducting coordinate and FFT transformation on the normalized image, finally conducting wavelet transformation on the image to obtain a characteristic vector and recognizinghuman body behaviors in combination with a minimum distance classification method. By testing a large quantity of human body behavior samples, the recognition rate of the algorithm reaches 90%. Through a depth image and translation, scaling and rotation invariance of the wavelet moment characteristics, the privacy problem which exists in traditional video image processing is solved, and the recognition performance is improved. The recognition capability of the human body tumbling behaviors is improved, and the algorithm has better robustness and a very good application prospect.

The invention relates to image feature optimization selection method and system based on improved ant colony optimization. The method comprises the steps of: establishing a feature attribute table, ants and an ant table of a feature sample; carrying out feature selection and coding on each ant, and calculating goodness of each ant; randomly selecting partial ants and carrying out variation on selected feature subsets; carrying out stop judgment on feature selection; after once feature selection is completed on all ants, calculating pheromone content and feature selection probability of the features and updating the data of the feature attribute table; and outputting an optimal feature set. The invention ensures that the improved ant colony optimization is applied to the image (especially a cotton foreign fiber image) target feature optimization, realizes the optimized selection of image high-dimensional features, effectively reduces the feature dimensions, and improves the accuracy of classification identification.

The invention discloses a low-carbon power supply grid planning model considering carbon emission flow. A target function is shown in specification as follows, wherein Cy, C<O>y, C<G>y and C<T>y respectively express total investment construction cost, fixed running cost, variable running cost and regional connection line onstruction cost of a system in the y(th) year, the fifth item in the target function expresses carbon transaction cost, the Cy, the C<O>y and the C<T>y are described by investment problem, the C<G>y is described by running sub-problem, Y expresses all years to be planned, and r expresses inflation rate.

The invention relates to the technical field of power plant maintenance task scheduling, in particular to a hill wind power plant maintenance task intelligent scheduling method, which comprises the following steps of S1, collecting maintenance task information and states; S2, collecting task resource information; S3, sorting the execution priorities of the maintenance tasks, and sorting the execution priorities of the maintenance tasks according to comprehensive calculation of the emergency degree of the tasks, the importance degree of maintenance equipment and the latest maintenance completion time factor; S4, performing a maintenance task resource allocation scheme; and S5, processing the tasks according to the task execution sequence and the resource allocation scheme. According to thehill wind power plant task scheduling method based on various resource constraint conditions, priority maintenance and execution of important equipment and important maintenance plans of a wind powerplant can be ensured, and through distribution of various task resources, the utilization rate of maintenance resources can be increased, the maintenance period of the equipment can be shortened, andthe effective utilization rate of the equipment for wind power generation can be increased.

The invention discloses a multi-unmanned aerial vehicle target allocation decision making method in an interval information condition, which comprises the following several steps: 1, multi-unmanned aerial vehicle target allocation decision making variables are defined; 2, ranking acceptability analysis indexes are calculated; 3, global acceptability analysis indexes are calculated; 4, a target allocation model is built; and 5, the obtained index result is substituted to the model, and the optimal allocation scheme is obtained. On the basis of stochastic multi-attribute acceptability analysis method, the target allocation model is brought forward, the multi-unmanned aerial vehicle target allocation decision making problem in an unclear interval information condition can be effectively solved, target allocation effects are improved, and needs of a real complex battlefield environment can be met.

The invention discloses a path planning method and device and electronic equipment. The method comprises the following steps: performing environmental modeling according to static road network information and dynamic road condition information of a road network to obtain an environmental model; determining a plurality of candidate paths according to the starting point and the terminal point; extracting environmental features corresponding to each candidate path from the environmental model by utilizing a feature extraction network of the path planning model; inputting the environmental features into a value estimation network of a path planning model to obtain estimated values of candidate paths output by the value estimation network; and determining an optimal path in the candidate pathsaccording to the estimated value. According to the technical scheme, the dynamic road condition information comprises motion states of a plurality of intelligent agents in a road network, and has spatial globality; when path planning is carried out, the environmental characteristics corresponding to the candidate paths determined according to the starting point and the terminal point are extractedaccording to the environmental model, so that full-path planning instead of planning in different time windows is achieved, and time globality is achieved.