104results about How to "Simplify the amount of computation" patented technology

The invention provides a user selecting method of a multi-user MIMO communication system based on a codebook. An e Node B groups and selects dispatch users according to a PMI and a CQI feedback of users in a cell. The dispatch users are grouped according to the quasi orthogonality among codebook vectors expressed by the PMI. When the system runs, the e Node B ensures a signal path area which belong to each user according to the PMI feedback by a UE, selects a user with the maximal CQI to be taken as an optimum user of the area in the signal path subarea according to the CQI reported by the UE, computes the combined spectrum efficiency of each paired user group according to a previously set quasi orthogonality relationship, and selects one group of users with the largest combined spectrum efficiency to be taken as the current paired user group. The method can greatly simplify the computing complexity selected by the user in the multi-user MIMO communication system and can obtain the throughput performance near the optimum.

The invention discloses a lab-space-based detection method based on image color cast. The method includes the steps of firstly, acquiring a current video frame; secondly, converting RGB (red, green, blue) into a lab space; thirdly, evaluating ratio K of image average chroma D to chroma center distance M according to a lab value; fourthly, evaluating histograms of a and b, calculating a factor according to the histograms, and multiplying the factor to the ratio K to acquire a result f; and fifthly, determining color cast occurs in an image to be detected when value of f is larger than a threshold value. Limitation of scene or prior knowledge is avoided, and compared with the prior art, the lab-space-based detection method based on image color cast is higher in detection precision and has excellent detection capacity in terms of scene color occupying large area also, universal adaptability and higher detection accuracy for color cast and reliability. In addition, conversion of color space and knowledge of the histograms are utilized flexibly, calculation is simplified, and real-time performance is high.

A self-integrity monitoring method suitable for a satellite-borne navigation receiver can reflect statistics detection amount of monitoring data quality by utilizing a pseudo-range information structure of the satellite-borne navigation receiver and detect faulted satellites according to the index requirements of a system and a distribution structure detection threshold obeyed by statistics amount. The self-integrity monitoring method suitable constructs the statistics detection amount of an odd-even space through QR decomposition of a primary observation array and based on the least square method, and carries out usability judgment, fault detection and fault identification in the odd-even space. The self-integrity monitoring method has higher response speed, saves more resources and is more suitable for satellite-borne application.

The invention discloses method and device for judging the confidence of speech recognition. The method comprises the following steps of: receiving an input speech; confirming a recognition result of the input speech corresponding to a recognition grammar; confirming a competition result of the input speech corresponding to an intrinsic phoneme-decoding space; and utilizing the recognition result and the competition result to calculate the confidence of the recognition result, wherein the construction process of the intrinsic phoneme-decoding space comprises the following steps of: clustering various phonemes contained in a preliminarily-trained acoustic module to acquire an intrinsic phoneme set; retraining an acoustic module and a speech module representing various intrinsic phonemes; and constructing the intrinsic phoneme-decoding space based on the intrinsic phoneme set and the acoustic module and the speech module acquired by retraining. By clustering the phonemes to construct the intrinsic phoneme-decoding space and decoding based on the space to acquire the competition result, the invention can rapidly and effectively acquire the competition result to be commonly suitable, thereby estimating the confidence more accurately and efficiently.

The invention belongs to the technical field of wireless communication, and particularly relates to a channel estimation algorithm for a multi-user massive MIMO (Multiple Input Multiple Output) system under a frequency division duplex (FDD) mode. In the multi-user massive MIMO system, channel estimation is realized by a sparse signal recovering technology through an inference method based on Bayes compressed sensing, and channel estimation overhead of the FDD massive MIMO system can be lowered greatly.

The invention proposes a method for estimating SOC of a lithium battery based on the improved EKF algorithm. On the basis of the EKF algorithm, by means of the robust data correction idea, a residualek between the EKF observation variance mid-end voltage estimation value (shown in the description) and an actual measured value yk is tacked as the reference; an influence function is adopted, and athreshold Eta is set; the residual value between the end voltage estimation value and the actual measured value yk is compared with the threshold value Eta; a noise variance Qk of the filtering process is corrected in real time; the weight of the noise Qk estimation error is reduced, and the robust objective function reaches a minimum value; then the observed noise covariance matrix Rk is adjustedthrough dynamic intervals of different SOCs, and the influence of the model error on the SOC estimation accuracy can be reduced.

The invention discloses an orthogonal frequency division multiplexing (OFDM) received frame synchronizing method for co-training sequence mutual-correlation information and is applicable to all OFDM systems adding training sequences in front of data frames. The OFDM received frame synchronizing method specifically includes 1), designing a frame structure of transmitting end; 2), performing mutual-correlation calculation and energy calculation synchronously by means of delay mutual-correlation algorithm at a receiving end, and obtaining regular measuring results; and 3) selecting energy threshold value and regular measuring threshold value of an energy window, and judging synchronous time. By accumulating a plurality of segment correlation values and uniformizing, operation load is reduced. The problems about peak platform and a plurality of high and side peak values are solved, and synchronization accuracy is increased. Channel influence is eliminated by receiving segmental mutual correlation of the sequence, and high frequency offset can be resisted. Dynamic range of receiving data is enlarged after uniformization, and selection of regular judgment time threshold value and synchronous detection can be simply and accurately realized. Meanwhile, adaptability is improved, frame synchronization in various channels can be completed accurately.

The invention relates to a statistics downscaling method based on the SVM algorithm. The statistics downscaling method includes the following steps that firstly, large-scale weather forecast factors are screened and representative factors are obtained; secondly, a wet day state and a dry day state of existing day-by-day rainfall data are judged, the data are processed, an SVM is used for constructing a classification relation between a raining state and the corresponding forecast factor at first, and a statistics relation on raining in wet days and meteorological factors is constructed; thirdly, an existing classification relation is used for classifying rainfall states in the incoming days through the weather forecast factors; fourthly, an existing statistics relation is used for fitting rainfall data on incoming rainy days through the weather forecast factors; fifthly, the data are restored and a forecast rainfall sequence is obtained. The statistics downscaling method based on the SVM algorithm has the advantages that effects of the statistics downscaling method are superior to those of a traditional statistics downscaling method and advantages are prominent when downsizing is carried out on the temperature; the operation volume is little, the method is convenient to use and specialized in processing a large volume of data of multiple batches, rainy days and dry days can be distinguished and higher precision is obtained; random values are added into rainfall and the fitting precision of rainstorms is higher.

The invention provides a hierarchical software health management method and system for microsatellites. The hierarchical software health management system performs hierarchical health management on satellites according to different subsystems, and comprises a sensor layer, a state monitoring layer, a health evaluation layer, a prediction layer, a decision processing layer and the like. Through software sensors deployed at operation key nodes of all sub-components of satellite housekeeping, the hierarchical software health management method collects running state data of each sub-component andanalyzes the running state data to obtain fault information, hierarchically evaluates the health condition of the current satellite system in combination with satellite historical telemetry data, software system knowledge and the fault information, pre-judges the health condition of the system at the future moment, and finally makes a health hazard strategy.

The invention discloses an image enhancement method and an image processing device. The image enhancement method comprises the steps of acquiring at least two original RGB images which is acquired foraiming at the same scene at different exposure time points; transferring each original RGB image to an HSI space for obtaining a brightness component image, a saturation component image and a tone component image; performing weighing fusion on each brightness component image, thereby obtaining a fused brightness component image; and transferring the fused brightness component image, the tone component image and the saturation component image to an RGB space for obtaining an enhanced RGB image. Through the image enhancement method and the image processing device, only the brightness componentis calculated and weighted fusion is performed, thereby simplifying a computing process, reducing calculation amount, reducing computing complexity in manufacturing a high dynamic comparing image.

The invention relates to a patent map visualization method and system. The patent map visualization method comprises the steps that multi-dimensional text vectors corresponding to patent literatures in a patent literature data set are obtained; clustering analysis is conducted on the multi-dimensional text vectors corresponding to the patent literatures to obtain multiple clustering results; multiple patent literatures corresponding to the clustering results are mapped to a two-dimensional plane so as to perform dimension reduction processing, and patent literature data points are obtained; visualized calculation is conducted on the multiple patent literature data points to generate visualized data, and the visualized data is displayed on the two-dimensional plane in a three-dimensional image mode. By the adoption of the patent map visualization method and system, the operation amount in the patent map generating process can be simplified, accordingly the occupation to a CPU and a memory is decreased, and real-time patent map generation and display are achieved.

The invention discloses a face image processing method and device. The method comprises the following steps of: receiving an input to-be-processed face image; normalizing the to-be-processed face image to obtain a standard face image with a preset standard size; inputting the standard face image into a deep learning-based full-connection neural network; and mapping and outputting a face template corresponding to the to-be-processed face image by adoption of the trained full-connection neural network. By adoption of the method and the device, the face feature extraction speed can be improved.

The invention discloses a method, a device, a processor and a terminal for processing a face model of a game character. The method comprises the following steps of: acquiring a feature point set of each of a plurality of presegmented parts contained in a facial model of a game character; associating feature points in a feature point set with weights of corresponding parts in a facial model; in response to a first touch operation acting on the facial model, a feature point of a corresponding portion of the facial model for the first touch operation is acquired and the feature point is transformed into a position. In response to a second touch operation acting on the facial model, a currently adjusted portion in the facial model is determined based on the positional relationship of the ray generated from the position information of the camera and the second touch operation with each feature point in the feature point set. The invention solves the technical problems of high complexity, poor scalability and low fault tolerance of the face pinching function designed for the game in the related art.

The invention discloses a capturing device with position compensation. The capturing device comprises a drive system, a control system and a bottom plate provided with at least one walking wheel. The capturing device further comprises a target mechanism used for intercepting a target, a moving mechanism arranged on the bottom plate and connected with the target mechanism, and a feedback mechanism. The moving mechanism is used for driving the target mechanism to move on the bottom plate. The feedback mechanism is arranged on the bottom plate and comprises a movement detection part and an orthogonal omnidirectional wheel part which is connected with the movement detection part, and the feedback mechanism is used for obtaining the deviation value between the bottom plate and the target and feeding back the deviation value to the control system. By adoption of the device, although the device slides in an inertial manner at the target drop point position, the feedback mechanism records and feeds back the sliding path to the control system, meanwhile, the control system makes the target mechanism move on the bottom plate, the target mechanism is static relative to the drop point, and the target is captured. The capturing device is simple and compact in structure and fast in feedback real-time responding, and the target mechanism can be accurately controlled at the target drop point.

The invention discloses a method for realization of matrix transformation of three dimensional models. The method comprises: arranging a scene root node corresponding to a three dimensional scene and a plurality of three dimensional model nodes respectively corresponding to a plurality of three dimensional models in the three dimensional scene; adding the three dimensional model nodes chosen by a user for editing and corresponding to the three dimensional model to the matrix transformation node, using the matrix transformation node as a sub-node of the scene root node, and using the three dimensional model nodes as sub-nodes of the matrix transformation node; mapping the three dimensional model matrix transformation operation carried by the user to the matrix transformation operation of the matrix transformation node corresponding to the three dimensional model, thereby the matrix vector corresponding to the matrix transformation node is obtained; carrying out matrix operation of the coordinate figures of all coordinate points of the three dimensional model chosen for editing and the matrix vector, and obtaining the coordinate figures of all coordinate points of the three dimensional model after the matrix transformation, thereby the matrix transformation of the three dimensional model is realized. The method simplifies the operand of the three dimensional scene real time calculation, [and] saves the occupancy rate of the hardware resource.

The invention discloses a transplantation control system and a transplantation control method for a robot for column cultivation. The transplantation control system for the robot for column cultivation comprises a control interface, a bottom-layer program and a track library. The track library contains seedling cultivating travel tracks from optional plug seedling fetching positions to optional seedling cultivating positions on cultivation columns, return travel tracks from the optional seedling cultivating positions on the cultivation columns to the optional plug seedling fetching positions and corresponding movement laws of various joints and an end actuator of a manipulator. During continuous transplantation operation, a serial number of an initial seedling fetching position and a serial number of an initial seedling cultivating position are inputted from the control interface, the bottom-layer program automatically and continuously invokes the corresponding seedling cultivating travel track, the corresponding return travel track and the corresponding movement laws of the various joints and the end actuator of the manipulator from the track library according to a preset seedling fetching and cultivating sequence, and plug seedlings are continuously transplanted on the cultivation columns. The transplantation control system and the transplantation control method have the advantages that calculation in a control procedure is greatly simplified, the flexibility, the instantaneity and the reliability of the control system are improved, and the plug seedlings can be automatically, quickly and reliably transplanted on the cultivation columns.

The embodiments of the invention provide a steady voice source direction finding method and apparatus. The method comprises the following steps: based on sound signals acquired by a microphone array, obtaining a beam pattern of each time-frequency slice and performing clustering on extreme points of the beam patterns so as to obtain a direction of a voice source, wherein one beam pattern is obtained for each time frequency slice through assuming that only one voice source signal plays a major role at each time frequency slice. The method and apparatus can convert the problem of multiple sound sources into a problem of single-source direction estimation, and also improves the resolution while simplifying the operation amount.

The invention belongs to the technical field of robots, and particularly relates to a mass center track generation method and device, a computer readable storage medium and a robot. The method comprises the steps of determining the mass center position, the mass center speed and the mass center acceleration of the robot at a first planning moment; calculating the mass center track increment of the robot according to the mass center speed and the mass center acceleration of the robot at the first planning moment; and calculating the centroid position of the robot at the second planning moment according to the centroid position of the robot at the first planning moment and the centroid trajectory increment. According to the method, an incremental centroid trajectory generation mode is introduced, that is, the centroid trajectory increment is calculated based on the previous centroid velocity and acceleration, and a new centroid position can be obtained by superposing the centroid trajectory increment on the basis of the previous centroid position, so that the calculation amount required in the centroid trajectory generation process is greatly reduced, the calculation efficiency is improved, and the real-time performance is good.

The invention relates to the technical field of computers. The invention provides a three-dimensional graph processing method, a three-dimensional graph processing device, a computer storage medium and electronic equipment, and the three-dimensional graph processing method is applied to a display terminal for carrying out display processing on an image sequence obtained after rendering processingis carried out on a three-dimensional graph, and comprises the steps: generating a man-machine interaction interface suitable for local equipment; obtaining a human-computer interaction instruction input by a user through the human-computer interaction interface; sending the human-computer interaction instruction to a cloud server; receiving a transmission video stream sent by the cloud server; decoding the transmission video stream to convert the transmission video stream into an image sequence; and playing and displaying the image sequence. According to the three-dimensional graph processingmethod, the technical problem that in the prior art, the requirement of a method for equipment configuration of a display terminal is high can be solved, three-dimensional simulation software can runon more low-configuration equipment, and the application range of related large three-dimensional simulation application software is expanded.

The invention discloses a three-dimensional data fault enhancing method based on curve Gabor filtering. The three-dimensional data fault enhancing method based on curve Gabor filtering is applied to the seismic data processing field, and includes the steps: performing structural directing filtering noise reduction and coherence cube calculation based on a gradient structure tensor azimuth on the original three-dimensional seismic amplitude data volume, and obtaining the coherence attribute volume which can fully reflect the self structural characteristics of the geologic body; using the improved curve Gabor filtering algorithm to process the coherence to enable the large fault structure in the original coherence cube to become more clear; and restraining the noise in the background, and significantly enhancing the tiny fault details.