34results about How to "Solve data redundancy" patented technology

The invention provides a shared parallel data reading-writing apparatus of multi memories among multi controllers, and a reading-writing method of the same. The shared parallel data reading-writing apparatus of multi memories among multi controllers includes a plurality of processor units and a plurality of storage units, wherein each storage unit is corresponding to each arbitration unit; each processor unit is corresponding to an address partition unit and a read data acquisition unit; aiming at the situation that the plurality of processor units access one storage unit at the same time, the arbitration unit can be used for carrying out arbitration management on a data operation request; and only the data operation request satisfying the preset conditions is selected; and the corresponding storage unit is used to execute data operation. As the processor units can directly access the storage units according to the addresses distributed by the address partition units, higher bandwidth can be provided; the operating performance of the processors can be satisfied; and the data access efficiency is greatly improved. Besides, the data exchange redundancy problem among the plurality of processors can be solved, so that the shared parallel data reading-writing apparatus of multi memories among multi controllers has wide market prospect in the integrated circuit field.

The invention discloses an AIS (Automatic Identification System) trajectory big data-based ship traffic environment reconstruction method. The method includes the following steps that: step 1-1, a time interval parameter is segmented so as to be divided into various time segments, the trajectory data of all ships in a specified geographical range within a time segment are only requested each time;step 1-2, a web front end temporarily stores the trajectory data in the step 1-1, preprocesses the trajectory data, and calculates and reconstructs the spatial relationship of the trajectory data; step 1-3, the web front end periodically requests new trajectory data from a server and performs dynamic update; and the step 1-1, 1-2, and 1-3 are repeated, so that the continuous reconstruction of a traffic environment can be realized at the web front-end. With the method adopted, an AIS data-based visualization technological means is provided for shipborne equipment and the equipment of a maritime supervision department, and inland navigation can be smoother, more orderly, safer and more efficient, and obvious economic and social benefits can be brought about.

The invention relates to an indoor positioning fingerprint database construction method based on crowd sensing. The method includes: acquiring fingerprint data in a target area within a period in themanner of crowd sensing, and establishing a fingerprint data model; initializing a prototype fingerprint quantity, quantifying an original fingerprint data set F by employing a learning vector to obtain a set of prototype fingerprints, and performing spatial clustering division on the fingerprints; and calculating an Euclidean distance between the prototype fingerprints, establishing a dissimilarity matrix, matching the prototype fingerprints and physical position tags by employing a multidimensional scale algorithm, and finally generating a fingerprint database data model. The method is advantaged by high construction efficiency.

The invention discloses a high-speed train axle temperature prediction method based on a data-driven support vector machine, and the method comprises the steps: firstly, collecting the temperature ofan axle and factors influencing the temperature change of the axle through a sensor, and carrying out the data preprocessing of the collected real-time data; secondly, correlation between the axle temperature and other factors influencing axle temperature change is obtained through mutual information and Pearson correlation coefficients, and factors with large correlation with axle temperature change and factors with small correlation are screened out; then, selecting a kernel function for the axle temperature prediction model by utilizing a support vector regression machine, and establishingthe axle temperature prediction model; and finally, substituting the to-be-measured data into the established axle temperature prediction model to obtain a predicted value of the axle temperature. Themethod disclosed by the invention can be used for analyzing factors with obvious influence on the temperature change of the axle, providing theoretical support for the work of the axle temperature alarm and simultaneously ensuring the safe operation of a train.

The invention provides a streaming media on-demand method and a streaming media on-demand system. The method includes that when any streaming media server in a streaming media server cluster receives a streaming media demand request from a terminal, the any streaming media server sends the streaming media demand request to a distributed storage device; the distributed storage device selects a target storage server from a distributed storage server cluster and sends the streaming media demand request to the target storage server; the target storage server acquires target data according to an index number in the streaming media demand request and sends the target data to the any streaming media server; the any streaming media server performs preset processing operation on the target data according to a preset streaming media protocol and distributes streamed data obtained after processing to the terminal. The streaming media on-demand method and the streaming media on-demand system have the advantages that data storage of the on-demand system is achieved, and video-on-demand localization is solved quite conveniently by combining distribution and deployment characteristics of distributed storage, so that user experience is improved.

The invention discloses an unconventional natural gas fracturing effect evaluation and productivity prediction method based on principal component analysis. The method comprises the following steps: 1) establishing an unconventional natural gas fracturing effect evaluation and productivity prediction evaluation database; 2) standardizing the original data; 3) calculating a covariance matrix; 4) solving a characteristic value and a characteristic vector; 5) calculating a variance contribution rate and an accumulated variance contribution rate, and determining the number of principal components;6) determining m vectors of the principal component, and outputting a new data sample; 7) establishing a multiple linear regression model according to the new sample, and performing fitting simulation test on the established multiple linear regression model; 8) performing fitting simulation inspection; and 9) carrying out shale gas fracturing effect evaluation and productivity prediction by usingthe established multiple linear regression model. According to the method, the problems of low linear relevancy and data redundancy of original data, artificial subjective factors caused by the factthat an existing evaluation method needs a large amount of sample data and expert scoring and the like can be well solved, and the accuracy of unconventional natural gas fracturing effect evaluation and productivity prediction is improved.

This invention discloses a method for processing variation of proxy call control function of IP multimedia system, in which, after receiving registriation request sent by a user, S-CSCF tansmits a subscription notification message to P-CSCF1 labeling a change of the user registration state if it tests that P-CSCF registered by the user has been changed from the first proxy call session control function P-CSCF1 to a second P-CSCF2, which resolves the problem of data redundancy on the original P-CSCF due to the change of the registration state of the user.

The invention discloses an MR data storage method, a device, a server and a storage medium. The method comprises the steps of obtaining MR data of a base station; decompressing and analyzing the MR data to generate a plurality of storage tables; carrying out format unification on the plurality of storage tables to generate a plurality of unified tables, wherein the unified tables comprise at leastone key field; performing association integration on the plurality of unified tables based on the key field to generate a wide table, wherein the wide table comprises a preset label; collecting statistics based on a preset label is carried out on the wide table according to user requirements to generate a dimension table for a user to check; and summarizing and counting the dimension tables basedon a preset calculation rule to generate a service table, and storing the service table in a spark big data cluster. According to the technical scheme of the invention, the method achieves the reuseof calculation results, and greatly reduces the storage and calculation cost.

The invention discloses a sweep frequency device and a method. The sweep frequency device and the method are used for solving the problems of too much sweep frequency data and data redundancy to improve sweep frequency speed and processing efficiency of the sweep frequency data. The sweep frequency device comprises a radio frequency processing module, a screening module and a signal processing module. The radio frequency processing module is used for receiving wireless radio-frequency signals of each frequency point according to a configured frequency point list, and converting the received wireless radio-frequency signals to base band digital carrier signals and then outputting the base band digital carrier signals. The screening signals is used for measuring signal strength of each received base band digital carrier signal, and screening out a set number of base band digital carrier signals with the highest signal strength and storing the base band digital carrier signals in a storage module. The signal processing module is used for conducting corresponding signal processing on the base band digital carrier signal stored in the storage module according to a configured sweep frequency parameter, and reporting obtained sweep frequency data to a data collecting system through a port module.

The invention provides a historical and cultural heritage digital display method. The method comprises the steps of analyzing basic elements and a model file, and processing the model file into a data stream; sending the basic elements to a scene layering model rebuilding process, and performing code transmission and visualization by cooperative processing of a main processor and a graphic processor; organizing the elements in a scene into a scene tree; searching for a vertex index of a scene layering model, calculating vertex topology of the scene layering model, performing grid vertex combination according to error measurement, and rebuilding approximate representation of an original grid; performing priority sorting on rendering objects, performing batch rendering on the objects with consistent rendering states, and then performing smoothing processing on space-time continuity of the scene. According to the historical and cultural heritage digital display method provided by the invention, the problems of data redundancy and low storage and scheduling efficiency of a conventional multilevel model are solved.

The invention discloses an intelligent voice question-answering method based on deep learning and emotion recognition, and the method comprises the steps: obtaining voice from a user, processing the voice to extract text information corresponding to the voice, carrying out the feature parameter extraction of the voice of the user, and carrying out the emotion analysis of the voice according to the extracted feature parameters to form an emotion tag; and inputting the generated text information into the trained semantic representation and matching model, matching with a question and answer library in combination with the obtained emotion tag, obtaining an answer of the question, and outputting the answer. The technical problem that an existing intelligent question-answering system is still low in accuracy due to ambiguity and complexity of natural languages and not rich in application of a deep learning algorithm can be solved, and the problem that when Chinese sentence information is captured, data redundancy and single dimension still exist can be solved. And therefore, system matching deviation is easily caused.

The invention relates to a RGB-D target recognition method based on quaternion generalized discriminant analysis, including quaternion-based RGB-D image representation method, define quaternion generalized discriminant analysis, two-way quaternion generalized discriminant analysis recognition method base on average row and average column; wherein the quaternion-based RGB-D image representation method solves that problems of data redundancy and extra computational overhead when the existing quaternion color image representation method adopts four-dimensional quaternion to represent three-dimensional color images, The kernel function is introduced into the field of quaternion subspace analysis, and the quaternion generalized discriminant analysis is defined, which solves the problem that theexisting quaternion subspace analysis algorithms are not effective in processing quaternion nonlinear signals. Finally, the bidirectional quaternion generalized discriminant analysis based on averagerows and average columns is used to eliminate the influence that aiming at the RGB-D recognition process, the computational complexity of quaternion kernel matrix eigendecomposition is too large, which improves the recognition effect of the target recognition method.

The invention discloses a representation learning method based on a superimposed convolution sparse auto-encoder. The method is characterized by comprising the following steps of 1) designing and realizing a reconstruction independent component analysis algorithm including whitening, taking an image data set as input, and iteratively optimizing and learning an output reconstruction matrix thereofto obtain a trained sparse auto-encoding model; 2) constructing a semi-supervised superimposed sparse auto-encoder model to train the feature representation; 3) constructing a convolution model to extract block features from the data, and generating convolution feature representation by applying convolution and pooling operations; 4) superposing the convolution sparse auto-encoder, and further optimizing the convolution feature representation; and 5), training a classifier on the image data set by using a logistic regression model and obtaining a classification result. According to the method,the characteristics of the auto-encoder model and the convolution pooling structure are combined, and a small part of labeled data in a large-scale data set is utilized so that feature representationvectors are optimized, and the classification accuracy of the image data set is improved.

The invention provides a cross-chain asset transfer method, computer equipment and a storage medium, and the method comprises the steps: executing a first asset cross-chain transaction, storing a first asset which is to be transferred to a second block chain and of which the asset type is a second asset type into a first asset cross-chain contract of a current block chain, and taking the first evidence hash of the first asset as a newly added leaf node of a first cross-chain single-class asset tree corresponding to the second block chain and the second asset type in a first cross-chain asset aggregation tree of the first asset cross-chain contract so as to update the first cross-chain asset aggregation tree. According to the method and the device, transfer of various different types of assets among multiple block chains can be realized only by maintaining one tree in the contract, so that the problems of data redundancy, inconvenience in tree root query and the like in a scene of transferring various types of assets in multiple block chains are solved.

The invention provides an adaptive scale DEM (digital elevation model) modeling method capable of keeping terrain features. A fine DEM database suitable for wide areas to local areas is built; an adaptive scale data structure and an analysis method are put forward; multi-scale terrain parameters and features are extracted from the fine-scale DEM database according to the application requirements;according to the scale information, different grades of terrain features are kept; and terrain skeletons are almost same under different scales. With the method, the adaptive scale technology is applied to the field of digital terrain analysis, the multi-scale DEM generation efficiency is improved, and the terrain characteristic consistency is guaranteed, and the requirements on the scale in the practical application are satisfied.

The invention discloses a plastic injection molding process parameter optimization method and device. The method comprises the following steps: constructing a fuzzy neural network; obtaining an original data set; setting optimization parameters; generating a training set according to the original data set and the optimization parameters; calculating membership layer node output of the fuzzy neural network; calculating the rule standard trigger strength of the fuzzy neural network according to the optimization parameters; training the fuzzy neural network according to the training set; and calculating the optimized output of the fuzzy neural network. According to the method, the fuzzy neural network is used for optimization and reasoning of the plastic injection molding process rule, end-to-end learning can be directly carried out in the fuzzy network, preprocessing such as data clustering or dimension reduction does not need to be carried out firstly, the problem of data redundancy in the plastic injection molding process rule optimization reasoning process is effectively solved, and the efficiency of the plastic injection molding process rule optimization reasoning is improved. The reasoning efficiency is greatly improved, and the rules in the obtained plastic injection molding rule base have high precision and good interpretability at the same time.

The present invention is a RGB-D target recognition method based on quaternion generalized discriminant analysis, including RGB-D image representation mode based on quaternion, definition of quaternion generalized discriminant analysis, two-way quadruple based on average row and average column A generalized discriminant analysis recognition method for quaternions; among them, the quaternion-based RGB‑D image representation method solves the data redundancy and additional calculations that exist in the existing quaternion color image representation method when using four-dimensional quaternions to represent three-dimensional color images For the problem of insufficient overhead, the kernel function is introduced into the field of quaternion subspace analysis, and the generalized discriminant analysis of quaternion is defined, which solves the problem that the existing quaternion subspace analysis algorithm has an unsatisfactory effect on quaternion nonlinear signal processing, and finally Through the bidirectional quaternion generalized discriminant analysis recognition method based on the average row and average column, the problem of excessive computational complexity of the quaternion kernel matrix eigendecomposition in the RGB-D recognition process is eliminated, and the target recognition method is improved. recognition effect.

The invention discloses a tower model rapid construction method comprising the following steps: independently modeling tower parts required by a tower model to obtain tower part models corresponding to different voltage grades; according to the tower composition structures of different voltage grades and the hanging point positions of tower parts, tower assembly anchor point models of different tower types of different voltage grades are established respectively; classifying the established tower part models and tower assembly anchor point models according to voltage grades to obtain a plurality of classification files for the tower part models and tower assembly anchor point models of different voltage grades, and uploading the plurality of classification files to a server; when the tower model is constructed, the tower part model and the tower assembly anchor point model of the corresponding voltage grade are downloaded according to the voltage grade of the transmission, distribution and transformation line, and the tower part model and the tower assembly anchor point model are assembled together by using the corresponding tower assembly relation information read from the server to form the tower model.

The invention discloses a sweep frequency device and a method. The sweep frequency device and the method are used for solving the problems of too much sweep frequency data and data redundancy to improve sweep frequency speed and processing efficiency of the sweep frequency data. The sweep frequency device comprises a radio frequency processing module, a screening module and a signal processing module. The radio frequency processing module is used for receiving wireless radio-frequency signals of each frequency point according to a configured frequency point list, and converting the received wireless radio-frequency signals to base band digital carrier signals and then outputting the base band digital carrier signals. The screening signals is used for measuring signal strength of each received base band digital carrier signal, and screening out a set number of base band digital carrier signals with the highest signal strength and storing the base band digital carrier signals in a storage module. The signal processing module is used for conducting corresponding signal processing on the base band digital carrier signal stored in the storage module according to a configured sweep frequency parameter, and reporting obtained sweep frequency data to a data collecting system through a port module.

The invention discloses a block chain-based data asset transaction method, computer equipment and a readable storage medium. According to the method, the block chain network is established based on the Fabric block chain framework, and the data assets are used as private data to be chained, so that the specific data can be cleared within limited time while the chained data store the Hash abstract, the problem of data redundancy caused after the data assets are sold for multiple times is solved, and the storage pressure of the block chain is reduced. According to the method and the system, in combination with a DataX tool, localized storage of data assets from a chain database to a big data application database is realized, and the problem that a state database in a conventional block chain data transaction platform is not suitable for big data calculation is solved. According to the method, data streams of links such as data asset uplink, synchronization and storage are recorded, so that a data asset transaction process is perfected, and the problem of data asset transaction disputes is avoided to a certain extent.

The invention relates to the technical field of big data, is applied to the field of smart government affairs, and discloses a data traceability method, device and equipment and a storage medium, which are used for improving the data traceability efficiency. The data traceability method comprises the following steps: acquiring an original traceability object, and determining at least one traceability participation unit information corresponding to the original traceability object; obtaining a traceability link corresponding to the original traceability object to obtain an original traceability link; obtaining original evidence data uploaded by each traceability participation unit terminal based on the original traceability link; when a data traceability request is received, traceability object matching and traceability link matching are carried out on information to be traced through the data traceability request, and a target traceability object and a target traceability link are obtained; and determining a data processing result of the to-be-traced information in each traceability participation unit based on the target traceability object and the target traceability link. In addition, the invention also relates to a block chain technology, and a data processing result can be stored in a block chain.

The invention belongs to the technical field of data processing, and particularly relates to a data storage platform construction method compatible with a data warehouse and a data lake, which comprises the following steps of: 1, storing data of the data lake and the data warehouse by adopting column storage and row-column mixed storage; 2, integrating the storage layers of the data warehouse and the data lake, and independently partitioning the storage areas of the data lake and the data warehouse in a partitioning manner; 3, a unified metadata management layer is constructed, bottom storage implementation details are shielded for upper-layer application loads through the unified metadata management layer, and unified data services for the upper-layer application loads are provided. In addition, the problem of data redundancy between two products can be solved, a globally unified data management view is provided for the user, further, the technology stack can be reduced and simplified, the overall management and operation and maintenance cost is reduced, and the data redundancy and storage cost are reduced.

The invention relates to the technical field of power equipment, in particular to an intelligent substation which comprises an intelligent comprehensive monitoring unit, the intelligent comprehensive monitoring unit comprises a central controller, an environment monitoring module, a security module, a video monitoring module and an alarm module, and the central controller comprises a processor, a touch screen and a communication module. The touch screen and the communication module are respectively and electrically connected with the processor, the environment monitoring module, the security module, the video monitoring module and the alarm module are respectively and electrically connected with the communication module, and the environment monitoring module, the security module, the video monitoring module and the alarm module which are originally independent and scattered are fully integrated through the intelligent comprehensive monitoring unit. The problems of unreasonable resource utilization, data redundancy, data inconsistency and the like in a distributed system are solved, real-time monitoring of in-station electrical equipment and environment conditions is truly realized, the operation environment optimization is realized, operation faults of power distribution equipment caused by out-of-control of the operation environment are avoided, and the service life of the equipment is prolonged.

The invention discloses a video human body behavior recognition method based on a CNN and an accumulated hidden layer state ConvLSTM, and the method comprises the steps: employing a 2D CNN which is pre-trained on an Image Net as a backbone model, and inserting an AH-ConvLSTM module behind a plurality of fixed network layer positions in the CNN to construct an overall network; acquiring a data sample set containing video data and labels, and dividing the data sample set into a training sample set and a test sample set; sampling a video in a fragment sampling mode, and using frames obtained through sampling as input to be sent to the whole network; setting a learning rate to train the whole network, and using the training sample set as training data; using a classifier for generating recognition scores of all categories, updating overall network parameters through back propagation, and storing the parameters as weight files; and initializing the whole network by using the weight file with the highest verification accuracy, performing fragment sampling on video frames of the test sample set, and inputting the frames obtained by sampling into the whole network to learn spatio-temporal information in the video so as to obtain an identification result.

The invention provides a streaming media on-demand method and a streaming media on-demand system. The method includes that when any streaming media server in a streaming media server cluster receives a streaming media demand request from a terminal, the any streaming media server sends the streaming media demand request to a distributed storage device; the distributed storage device selects a target storage server from a distributed storage server cluster and sends the streaming media demand request to the target storage server; the target storage server acquires target data according to an index number in the streaming media demand request and sends the target data to the any streaming media server; the any streaming media server performs preset processing operation on the target data according to a preset streaming media protocol and distributes streamed data obtained after processing to the terminal. The streaming media on-demand method and the streaming media on-demand system have the advantages that data storage of the on-demand system is achieved, and video-on-demand localization is solved quite conveniently by combining distribution and deployment characteristics of distributed storage, so that user experience is improved.

The invention provides a multistage data summarization method, a distributed data management system and a summarized data management system. The multistage data summarization method comprises the steps of obtaining M data interaction records; storing the M data interaction records in a distributed database, the distributed database comprising N unit databases, and multiple data interaction recordsin the M data interaction records being stored in each unit database; for each unit database in the N unit databases, summarizing the multiple data interaction records based on the target account togenerate J data interaction intermediate summarized records; and sending the plurality of data interaction intermediate summarized records in the N unit databases to a summary database.

The invention belongs to a technology of radio frequency identification and particularly relates to a multi-microcontroller radio frequency identification read-write device and a data transmission method thereof. The multi-microcontroller radio frequency identification read-write device comprises a radio frequency transceiving unit group, a microcontroller control unit group, a master micro controller control unit, a stabilized voltage power supply unit, a storage unit, a display unit, a reset unit, a programming unit and an upper computer interface unit. The data transmission method comprises the following steps: the micro controllers are initialized; radio frequency transceiving units send request transmission identifiers after receiving non-repetitive tag data, a passive mode of serial interface data transmission is started if effective chip selection exists; and the master micro controller unit is used for setting data transmission priority according to the request transmission identifiers, setting the chip selection of units corresponding to the highest priority to be low, starting an active mode of the serial interface data transmission, and setting corresponding chip selection to be high and regulating the data transmission priority after the data are uploaded according to an uploading mode. According to the multi-microcontroller radio frequency identification read-write device and the data transmission method, the reading efficiency of the read-write device is greatly improved, and the problem of a great amount of data redundancy in read-write systems is solved.

The embodiment of the invention provides an application subset division method and device, equipment, which can be used in the financial field or other fields. The method comprises the following steps: acquiring element information of a target application; under the condition that the target application is determined to be a product type application or a data type application according to application positioning information in the element information, determining a data object included in the target application, wherein each data object at least comprises a physical table of the application; and performing aggregation division on the data object according to an intimacy analysis model to obtain a subset division result of the target application, wherein the subset division result comprises a data object and a service. By utilizing the embodiment of the invention, the manual division cost can be reduced, the division efficiency and accuracy are improved, and the problems of application data coupling and data redundancy can be effectively solved.