362results about How to "Reduce computation" patented technology

A method of removing a motion vector from a group of motion vectors used in an encoding process includes providing a list of motion vectors, selecting an initial motion vector from the list of motion vectors, and providing an intermediate motion vector using a motion vector refinement process. The motion vector refinement process uses, in part, the initial motion vector. The method also includes forming a region defined by one or more parameters associated with the initial motion vector and one or more parameters associated with the intermediate motion vector, selecting an additional motion vector from the list of motion vectors, determining that the additional motion vector points into the region, and modifying a state of the additional motion vector.

A method and data structure is described that builds summary information using processor time that is usually proportional to the size of input data and a depth of hierarchies for a plurality of attributes. The output of computation is stored in a smaller area by eliminating redundant storage and computation. An index is generated which includes tuples or rows that include lower bound values for each of the attributes, values of children of the lower bound values based on the hierarchies of the attributes, and coverage class indicators of the coverage classes of the children.

The invention discloses a traffic flow statistic and violation detection method based on surveillance video processing. The method comprises the steps of obtaining a video source, reading a frame of image, judging if the current frame is a first frame, utilizing difference information of corresponding pixel points of the current frame and the previous frame to update foreground and background images if the current frame is not the first frame, utilizing the foreground and background images to extract an interest area MASK of the current frame, calculating integral images and integrated square images of the current frame, detecting in the interest area MASK of the current frame if a plate number target exists according to the obtained integral images and integrated square images of the current frame and utilizing an Adaboost algorithm, storing position information of the detected plate number target in a list TmpList, judging if the temporary target list TmpList is empty, judging if a tracking target sequence Track-List is empty if the temporary target list TmpList is empty, and adding the detected target in an ObjList into the Track-List if the tracking target sequence Track-List is empty. By means of the method, crossing traffic flow can be counted precisely, and if a vehicle target breaks the rule is judged combining indication of traffic signal lamps.

The invention provides a voice wake-up method and device based on artificial intelligence, and computer equipment. The method comprises steps that windowing enframing operation of online recorded voice is carried out to acquire at least one voice frame; characteristic extraction of the voice frame is carried out to acquire voice characteristics; a voice wake-up model based on the convolutional neural network is utilized to carry out calculation of the static voice characteristics contained in the voice frame, and posterior probabilities of the static voice characteristics contained in the voice frame corresponding to a non-wake-up word class and a wake-up word class are acquired; if the posterior probability of the static voice characteristics contained in the voice frame corresponding to the wake-up word class is greater than a pre-determined threshold, the online recorded voice is determined to contain wake-up words. The method is advantaged in that parameter quantities of the voice wake-up model based on the convolutional neural model can be effectively reduced, so the operand reduction effect is realized, and the voice wake-up effect can be improved.

The invention discloses a conformal array radar amplitude-phase error correction fast achieving method which mainly solves the problem in the prior art that the existing method is large in calculation amount and can not finish correction in real time fast. The achieving method includes arranging a virtual array, and calculating an ideal transformation matrix; obtaining conformal array radar receiving data; iteration initializing; correcting the transformation matrix; achieving direction of arrival (DOA) estimation on the virtual array; obtaining original conformal array radar amplitude-phase error parameters; judging whether iteration is converged; finishing conformal array radar amplitude-phase error correction. Iteration initial values are selected appropriately, DOA estimation is achieved fast by means of the virtual array, iteration is conducted between original conformal array radar amplitude-phase error parameters and the DOA estimation till iteration is converged, and the defects in the prior art that operation amount is large and correction can not be finished fast in real time due to the fact that a spectrum peak is used for searching and estimating information source orientation are overcome. The method is small in operation amount, easy to achieve, flexible and various in achievement mode and is applied to conformal array radar real-time amplitude-phase error correction.

The invention discloses a staggered base line angle measurement method based on a distributed sub-array. The problem that the angle is fuzzy in the distributed sub-array wave arrival direction estimation process is mainly solved. The method comprises the following implementation processes of: extracting a target signal from a radar echo; performing coarse angle measurement on a target signal by using a digital beam forming method; dividing the sub-array, equally dividing each sub-array into two novel sub-arrays, forming four novel sub-arrays, generating four base lines, and constructing a staggered relationship by two base lines selected from the four base lines; performing beam forming on the four sub-arrays, and calculating a phase difference observation value of the two base lines; calculating a value range of a fuzzy number by utilizing the constructed staggered relationship; and introducing a cost function, solving the fuzzy number by utilizing the minimum mean square error criterion, and solving the cost function value by utilizing the fuzzy number, so as to obtain a target wave arrival angle. The operation amount is reduced by adopting the processing method for constructing the difference of the two base lines, the calculation speed of the target arrival angle estimation is increased, and the method can be used for target identification.

The invention discloses an image identification-based intelligent transformation substation state monitoring system, which is characterized in that: a camera is used for monitoring the states of a circuit breaker indicator, a pressure plate, a knob switch and the like on a switch cabinet in an intelligent transformer substation, the image of an area where a monitoring device is arranged is acquired, the states of each indicating lamp, the pressure plate and the switch are obtained by using an image identifying technology, and relevant identified states are transmitted to a background monitoring system through a 61850 protocol, so that automatic management of the running situation of the intelligent transformer substation is realized.

The invention relates to the technical field of remote sensing surveying and mapping, in particular to a seasat remote sensing observation data matching method and a seasat remote sensing observation data matching system. The method comprises the steps of setting a time match threshold value, a space match threshold value and a match time length in advance, and dividing the sphere face of the earth into uniform virtual reseaus according to longitude and latitude; inputting time, searching the remote sensing observation data in the match time length by taking the time as the index; and computing the reseaus corresponding to the remote sensing observation data, traversing each reseau according to the row sequence of the reseaus, and computing time difference of the remote sensing observation data of a first satellite and a second satellite, corresponding to the same reseau, judging whether the time difference is smaller than or equal to the time match threshold value or not, and finishing the match according to the judging result. The seasat remote sensing observation data match system comprises a networking module, an index module and a network match module. The seasat remote sensing observation data match method and the seasat remote sensing observation data matching system can greatly reduce the data processing amount and are higher in operation efficiency.

The invention discloses a carrier synchronization method of an MPSK system based on maximum likelihood estimation, belonging to the technical field of the modulation and the demodulation of digital communication and comprising the following steps: firstly, eliminating the modulation information of a receiving signal to obtain a single carrier only polluted by noises; then, utilizing a fast search algorithm based on dichotomy to obtain maximum likelihood estimation values of frequency deviation and phase deviation; and then, compensating an original receiving signal by the maximum likelihood estimation values to complete carrier synchronization and demodulation, wherein the estimation and compensation method needs to be continuously used for tracking the carrier synchronization in the demodulation process. A simulation result shows that the carrier synchronization method has very good property approximated to the Cramer-Rao lower bound, very little calculation amount and very wide frequency deviation estimation range and is a practical carrier synchronization method.

The invention provides a predistortion device and a method. An input data is treated by a high speed predistortion and a digital signal is obtained by the feedback of a power amplifier. The result obtained from the parameter relation transformation of a digital signal is compared with the digital signal obtained from the process in which the input data is treated by the high speed predistortion and delayed. When the specification error requirement is satisfied, the related parameter in the parameter relation transformation is updated and saved in a lookup table as a predistortion parameter. The invention can use a uniform predistortion structure in an initial predistortion treatment and a normal predistortion treatment. Adopting the predistortion structure of an indirect discipline parameter, the invention collects data in a descending time slot and proceeds adaptive filtering treatment in an ascending time slot in a mobile communication system of a time division duplex, such as TD-SCDMA and OFDM_TDD system.

The invention provides a pedestrian crosswalk signal lamp green wave self-adaption control system and a method thereof; by fully utilizing the phenomenon that the dead time of road vehicles is more than pedestrian walking time when crossing the street, the pedestrian crosswalk signal lamp is controlled by the detection of the vehicle condition at the reserved area in front of the pedestrian crosswalk; the control method of the signal lamp is that: the control parameter of the signal lamp can be automatically adjusted according to the dynamic change of traffic flow, and the control system can be suitable for the change of the traffic flow automatically. When the vehicle condition in a certain distance in front of the pedestrian crosswalk is detected, the technology that a plurality of detected areas are arranged on the road is adopted to replace real physical induction coil; the number, size and position of the detected area can be set, or the relation of actual road distance and image pixel pitch is realized by adopting the camera calibration technology.

The invention discloses an intelligent fruit picking robot based on machine vision. The robot comprises a control system and a walking mechanism, a picking mechanical arm, a visual system and a powersystem which are connected with the control system respectively, the control system, the power system and the walking mechanism are fixedly connected, and the picking mechanical arm is a four degreesof freedom joint mechanical arm and is fixedly connected with the walking mechanism through a base; a visual sensor of the visual system is fixedly arranged above the wrist of the picking mechanical arm, a visual control module of the visual system is integrated with the control system and is connected with the visual sensor through signals, and the visual control module is used for conducting target positioning and identifying on images of the visual sensor based on a full-convolution network image processing technology of image partition; the visual control module transmits target positioning information to the control system, and the control system controls the walking mechanism to move and/or the mechanical arm to pick fruits. The intelligent fruit picking robot based on the machine vision has the advantages of compact structure, good stability, high intelligent degree, quick positioning, high positioning precision, and high environmental adaptability.

The embodiment of the invention provides an age estimation method and equipment. The method comprises the following steps: processing face images of users to extract initial characteristic data; selecting subsets of the initial characteristic data as low-dimension characteristic data with dimensionality less than that of the initial characteristic data; carrying out relevant component analysis on the low-dimension characteristic data to obtain training data; according to the obtained training data, training regression analysis parameters used for age estimation; adopting the regression analysis parameters to estimate the ages of the users. The embodiment of the invention can reduce the computation of the age estimation in the face recognition system.

The invention provides a method for overlaying image layers, which comprises: A, obtaining expected overlay areas of a current and a non-current on video and determining an overlay clipping area on video; B, calculating an overlay area and a non-overlay area between the expected overlay areas of the current layer and the incurrent layer; and C, overlaying the overlay area of the layers and placing in the relevant position in the overlay clipping region and directly placing the non-overlay area of the current layer in the relevant position in the overlay clipping region. The inventive method is easy to operate and realize.

The invention discloses a human body action real-time identification method based on an infrared image, and relates to the technical field of human body action recognition, and the method comprises the following steps: constructing an infrared image human body posture extraction network and an action recognition network based on skeleton features; acquiring an infrared pedestrian image data set and preprocessing the infrared pedestrian image data set, training an infrared image human body posture extraction network based on the preprocessed infrared pedestrian image data set, and obtaining a human body dynamic skeleton feature image; segmenting a region-of-interest sequence of the human body dynamic skeleton feature image, and training an action recognition network by utilizing a segmentation result; and acquiring a to-be-recognized infrared image, and performing action classification and prediction on the preprocessed to-be-recognized infrared image based on the infrared image human body posture extraction network and the action recognition network. According to the invention, the problems of poor real-time performance and low recognition rate of an existing behavior recognition method generally aiming at a visible light environment and performing human body behavior action recognition through an infrared image in a dark environment at night or in a severe weather environmentare solved.

The invention discloses a soft measurement method of free calcium of cement clinker. The method comprises the following steps: selecting 10 variables as auxiliary variables for clinker fCaO soft measurement according to a cement process, inputting the time sequence, serving as a model, of each variable, and performing normalization processing on the time sequence of each selected variable; according to the characteristic of the time sequence in the cement sintering process, establishing a clinker fCaO soft measurement model based on a multi-variable time sequence convolutional neural network,determining an initial parameter of an MT-CNN model and performing forward-direction training on the network; performing supervised training by utilizing error reverse fine adjustment, optimizing weight w and bias b in the MT-CNN by correcting the error, and predicting the cement clinker fCaO in real time by utilizing the trained MT-CNN model. Calculation of time delay between each variable and the clinker fCaO is avoided and the calculation quantity needed by time sequence matching is reduced; the convergence speed, the precision and the generalization ability of the model are improved; and the fCaO content of the cement clinker can be predicted well, the quality of the cement clinker can be improved and the production energy consumption can be reduced.

A wideband acoustic echo cancellation process comprise the following steps filtering an input shifting register which comprises far-end signals and the data in the input shifting register which comprises far-end signals, conducting the IDFT arithmetic to obtain far-end multi-sub-band data and near-end multi-sub-band data, conducting the adaptive filtering and the parameter updating for the far-end multi-sub-band data and the near-end multi-sub-band data to gain echo estimation difference, carrying out the IDFT arithmetic for the echo estimation difference of the multi-sub-band, conducting the synthetic filtering for the echo estimation difference of the multi-sub-band, adding the filtering output data which is gained through the filtering with the value in an output shifting register whose length is K, updating the output shifting register through using the value which is gained through the addition, outputting the front r points in the output shifting register, shifting forwards the positions of r points by the sequent K-r points, zeroizing the positions of r points, carrying out the shifting operations for the input shifting register, and repeating all the steps.

The invention discloses a space-time anti-interference method based on the variable diagonal loading capacity. The method comprises the following steps: step one, collecting global positioning system (GPS) satellite signals, and inputting the signals into a filter structure after being sampled; step two, conducting forward decomposition of a relevant subtraction multi-stage wiener filter on the signals, and recording a received signal and a reference signal of each stage; step three, replacing a real covariance matrix with a covariance matrix of the sampling signals; step four, calculating the diagonal loading range through a matrix inversion theorem, and further obtaining the diagonal loading capacity needing to be loaded through a covariance diagonal matrix element relationship; and step five, adding the diagonal loading capacity, conducting backward composition of the relevant subtraction multi-stage wiener filter, obtaining an optimal weight vector of the filter, and conducting adaptive filtering on current data. Compared with a multi-stage wiener filter, the improved relevant subtraction multi-stage wiener filter is low in operation quantity and storage quantity, and is suitable for systems of large dimensionality.

A reduced complexity precoder provides an efficient method and structure to precode a vector-signal-point sequence for transmission through a band-limited channel. The precoder enables a block-oriented receiver to recover an underlying data stream in the presence of inter-symbol interference and noise. The precoder structure is applicable to multicarrier systems such as DMT (discrete multitone) or related transform domain and vector communication systems. The inventive precoder reduces the cost of precoding by an order of magnitude and eliminates the need for a cyclic prefix in DMT and related communication systems. Related multicarrier transmitter and receiver structures and methods which reduce computation, increase transmission bandwidth and reduce transmission power are also developed.

The invention provides a target positioning and recognition method. The positioning and recognition method includes the following steps that: magnetic anomaly signals are acquired, the distance and unit position vector of a target are calculated through using a magnetic gradient tensor angle and distance measurement algorithm, and a position vector is calculated; the distance and unit position vector are substituted into the magnetic gradient tensor matrix of the target, and a magnetic moment vector is calculated; a correction coefficient g and included angle cosine cosphi are calculated according to the position vector and the magnetic moment vector, and a coefficient Q1 and a coefficient Q2 are calculated; and a new unit position vector is outputted under a convergence condition. The method provided by the invention has the advantages of high positioning accuracy, accurate magnetic moment estimation and long detection distance.

The present invention belongs to the field of information processing and medical information processing technology, in particular, it relates to a pulse information processing method based on rise wavelet. Said method includes the following steps: using pulse sensor to extract pulse waveform of human body, making said pulse waveform be passed through conditioning circuit and AD converter, and converted into digital signal, then inputted into a microprocessor; in the microprocessor making the above-mentioned digital signal undergo the process of waveform pretreatment, namely adopting adaptive coherent template method to make filtration; then utilizing filtered pulse signal and adopting threshold search method to obtain continuous stable waveform, then using time domain cluster analysis method to obtain optimal waveform containing more than five waveform samples; making the optimal waveform undergo the processes of multilayer wavelet decomposition treatment and threshold quantization treatment so as to finally obtain the characteristic vector W of pulse waveform.

The invention relates to an OpenGL (open graphics library)-based inverted image display processing device and method. The method comprises the following steps: loading and calculating parameters of an object model, acquiring texture mapping pixel data and/or default vertex colour values of the object model; carrying out vertex position spatial transformation and cutting operation on the parameters of the object model; carrying out rasterization on the parameters of the object model after the vertex position spatial transformation and cutting operation are carried out; according to the texture mapping pixel data, carrying out segment colouring treatment on the parameters of the object model after the rasterization is carried out, so as to generate an inverted image of the object model; and outputting the inverted image to display equipment for display. The device and method in the invention has the advantages that inverted image effect of the object model is realized by parameter control in an OpenGL drawing process, the simulated reality is more vivid, the storage space is saved, and the computation load of a CPU (central processing unit) and the internal memory usage are reduced; the more efficient flexibility and adaptability are provided for the CPU, the preferable frame effect and the fluency are realized, and the burden of a UI (user interface) designer is alleviated.

The invention discloses a method of mining and analyzing information security vulnerabilities. The method comprises the following steps: establishing a configuration file for an object program, wherein a series of stain analysis strategies are described in the configuration file; starting object program by a dynamic binary pile insertion frame Pin, and initializing the dynamic pile insertion analysis tool PinTool; analyzing the conditional transfer instructions of the target program through the PinTool, and generating the control flow chart and control dependency information data for the target program. The staked object accesses PinTool and user-defined library files to complete the marking of the stain flow, propagation and vulnerability detection. The invention realizes the propagationof stain marks from two aspects of data flow and control flow. The invention can detect and prevent external untrusted data from being used for unsafe data operation, does not need to analyze the source code of the target program, has low computational load, is suitable for commercial software without open source, can defend the secondary attack based on the same vulnerability by extracting the signature, and has lower rate of omission and false positives than the existing vulnerability detection technology.

The invention discloses a deterministic cloud data deleting method based on sampling and fragmenting of a cryptograph. The method mainly solves the problems that the fine granularity of access control is low and the safety is low in the prior art. The method comprises implementation steps that (1) a data owner carries out sampling and fragmenting on the cryptograph to generate a sampling cryptograph and a residual cryptograph, and the sampling cryptograph and the residual cryptograph are sent to a credible third-party mechanism and a could service provider respectively; (2) a file secret key is shared by a plurality of users in a cloud storage environment; (3) when an authorized user accesses the cloud data, the credible third-party mechanism merges the cryptograph of a file, and the authorized user decrypts the cryptograph of the file; (4) the cryptograph sampling and fragmenting technology is adopted to prevent the cloud service provider from acquiring the integrated cryptograph of the file, the credible third-party mechanism destroys the sampling cryptograph, and then deterministic data deleting is achieved. By means of the method, real-time deterministic deleting of data in cloud storage services can be achieved, and the privacy of user data can be improved. The method can be used for guaranteeing safe access of the user data in the cloud storage mode.

The invention discloses a saturated sliding mode variable structure control method for rapidly maneuvering attitudes of satellites. The saturated sliding mode variable structure control method includes steps of 1, acquiring current attitudes of the satellites and acquiring attitude errors of the current attitudes of the satellites relative to target attitudes; 2, determining bandwidths, damping coefficients, the maximum torque of actuating mechanisms, torque regulation coefficients and fundamental frequencies of flexible accessories, and computing control parameters for rapidly maneuvering the attitudes of the satellites; 3, computing saturated sliding mode surfaces of the satellites in rolling directions and computing linear sliding mode surfaces of the satellites in pitching and yawing directions; 4, acquiring control torque by means of computation; 5, acquiring corresponding control instructions for the actuating mechanisms by the aid of the control torque acquired in the step 4, transmitting the instructions to the corresponding actuating mechanisms and controlling the actuating mechanisms to rapidly maneuver the attitudes of the satellites. The saturated sliding mode variable structure control method has the advantages that the method is simple and reliable and is low in computational complexity, the satellites can be rapidly maneuvered and stabilized by the aid of the saturated sliding mode variable structure control method, accordingly, the attitude maneuvering time can be shortened, and the saturated sliding mode variable structure control method is obviously superior to other methods.

A method of reducing computation of palm rejection by projecting touch data is disclosed, targeting the handheld devices. The method targets first at obtaining a difference array, followed by extracting maximum values of rows and columns of the difference array to obtain a row projection list and a column projection list respectively. By repeated implementing of mutual capacitance detection, ghost palm blocks can be wiped out from the multiple palm blocks. Once integrating with a local spatial boundary detection algorithm, the sensed signals of intended input located within a rectangular palm block yet beyond a real palm block are consequently detected. The computational algorithm of the palm rejection of this invention is successfully built into the touch panel controller due to its substantially reduced computation.

The invention provides an outer radiation source weak signal detection method based on PCA (principal component analysis). The method comprises the following steps of solving an autocorrelation matrix R=E{x(t)x<H>(t)} of a sampling sequence x(t); carrying out SVD (singular value decomposition) on the autocorrelation matrix R, to obtain R=Q Lambda Q<H>, utilizing the decomposed diagonal matrix Lambda and unitary matrix Q to carry out linear conversion on the sampling sequence x(t) to obtain a strength-decreased sequence y(t) of received signals; carrying out the two-dimensional correlation of time frequency distribution on the third path to the Mth path of signals in the received signal y(t) to obtain the corresponding correlation value, namely the absolute value of psi i (tao, f); when the signal-to-noise ratio of one certain path is greater than or equal to a threshold value, determining the time delay and Doppler frequency corresponding to a peak point as the time delay and Doppler frequency displacement of one target; and when the signal-to-noise ratio of one path is smaller than the threshold value, judging that the signal of the path has no target. The method has the advantage that a weak target signal can be extracted under the conditions of strong same-frequency interference and low signal-to-noise ratio.

The invention discloses a wheel torque multi-objective optimization method for an electric wheel drive vehicle based on an optimal vehicle comprehensive performance. The method comprises the followingsteps: step one, calculating a vehicle requirement yaw moment of couple; step two, estimating each drive wheel slippage rate; step three, fitting an electric drive system loss characteristic curve; step four, establishing an online optimized objective function, and according to the requirement yaw moment of couple, the wheel slippage rate and the electric drive system loss, solving the objectivefunction to obtain each wheel torque; and step five, sending the requirement moment of force decided by each wheel to a controller of each wheel hub motor. According to the wheel torque multi-objective optimization method for the electric wheel drive vehicle based on the optimal vehicle comprehensive performance provided by the invention, in an online optimization process, by comprehensively considering the motor energy consumption, a yaw velocity tracking error and the slippage loss, the online optimization on each wheel driving torque of the electric wheel drive vehicle is performed, therebyimproving the vehicle economic efficiency as possible under the precondition of guaranteeing the vehicle stability, and realizing the optimal vehicle comprehensive performance.

The invention relates to an intelligent scene image filtering system. The system comprises a height-adjustable closure element arranged at a drain hole of a wash basin, wherein the drain hole gap width is adjusted based on the height of the closure element, the higher the adjusted height of the closure element is, the larger the drain hole gap width of the water basin adjusted based on the heightof the closure element is; scene shooting equipment arranged in a bathroom and used for shooting the front of the water basin, thereby acquiring a front scene image and outputting the same; recursiveprocessing equipment connected with the scene shooting equipment and used for receiving the front scene image, wherein the self-adaptive recursive filtering processing is executed on the front scene image to acquire a first filtering image. Through the system disclosed by the invention, the drain hole gap width of the water basin can be adjusted to the preference height of the user, thereby catering to the use habit of each user.

The invention discloses a routing rule configuration method and a communication device, and the method comprises the steps that a session management function network element receives a protocol data unit (PDU) session creation request transmitted by a first terminal, and transmits the information of a data packet through an N6 interface according to the first information for indicating the forwarding of the information of the data packet through the N6 interface; and a routing rule is generated for indicating to send the data packet of the 5G local area network LAN group to which the first terminal belongs to the N6 interface or discard the data packet. The session management function network element can determine the routing rule of the forwarding mode based on the N6 interface accordingto the first information, so that the session management function network element can determine to use the local forwarding mode or the N19 forwarding mode under the condition that the session management function network element can use the local forwarding mode or the N19 forwarding mode; by adopting the scheme provided by the invention, any one of a local forwarding mode, an N19 forwarding modeand an N6 interface-based forwarding mode can be used in a 5G LAN group, and the problem that the three forwarding modes cannot be used in the 5G LAN group in the prior art can be solved.