Patents
Literature
Patsnap Eureka AI that helps you search prior art, draft patents, and assess FTO risks, powered by patent and scientific literature data.

38results about How to "Reduce calculation steps" patented technology

Trigonometric function curve planning method and system for continuous jerk, and chip placement machine

ActiveCN114879609BGuaranteed driving abilityReduce calculation stepsTotal factory controlNumerical controlMachine controlClassical mechanics
The present invention relates to the field of chip placement machine control technology, and specifically discloses a trigonometric function curve planning method and system with continuous jerk, and a chip placement machine. The method includes the following steps: dividing the curve planning into an acceleration section, a uniform speed section, and a deceleration section, and the acceleration section and the deceleration section are respectively divided into a jerk section, a uniform acceleration section, and a deceleration section; setting and initializing the motion parameters of the curve planning; defining the average acceleration ratio of the deceleration section to the acceleration section as a softening factor, and defining the maximum acceleration in the deceleration section or the acceleration section to the average acceleration ratio as an acceleration coefficient; calculating the time nodes of each stage of the curve planning according to the motion parameters, the softening factor, and the acceleration coefficient; and updating the state of each stage of the curve planning according to the time nodes. In combination with the continuous characteristics of the second-order derivative of the trigonometric function, an S-curve planning with continuous jerk is designed, which effectively reduces the motion impact, ensures the stable time in place, and is simple to operate.
Owner:HEFEI ANXIN PRECISION TECH CO LTD

Automatic driving target detection method in night low-light environment

The invention discloses an automatic driving target detection method in a night low-light environment. The method comprises the following steps: constructing an automatic driving target detection model AQ9; according to the AQ9, an LKNet9 network is adopted as a backbone network to extract basic features from an input image, and for local features, night target obstacles are identified; a self-adaptive pooling algorithm is adopted, and the pooling scale and type are automatically adjusted according to the content of the input feature map; a nearest neighbor interpolation up-sampling module in the Head layer is used for DySample sampling, and self-adaptive adjustment is carried out according to the content of the input features; the method comprises the following steps: extracting features in a larger range through convolution operation of enhancing a receptive field, applying an attention mechanism on a feature map after convolution operation, and respectively adjusting channel attention and space attention; a C2f module is adopted, the spatial dimension of the feature map is focused, and the distribution weight of each spatial position is enhanced; and training the AQ9 by using the enhanced KITTI data set, and performing target detection by using the trained AQ9. According to the invention, night target detection can be carried out more efficiently, stably and accurately.
Owner:HUAIYIN INSTITUTE OF TECHNOLOGY

Key parameter optimization method for ground well electromagnetic exploration simulation calculation

The invention belongs to the technical field of electromagnetic exploration, and particularly relates to a key parameter optimization method for ground well electromagnetic exploration simulation calculation, which is based on solved irregular layered background medium Green function response, utilizes the plane symmetry property of a union-vector Green function and adopts a fast Fourier transform algorithm to optimize the key parameters of the ground well electromagnetic exploration simulation calculation. Only calculating and researching half of the grid point and vector Green function of the horizontal section diagonal line of the area, and reconstructing the distribution of the complete horizontal section grid point and vector Green function through a spatial interpolation algorithm; and solving the three-dimensional ground well electromagnetic field integral equation set under the background of the three-dimensional irregular layered medium by adopting a stable biconjugate gradient-fast Fourier transform algorithm, so as to obtain the total electromagnetic field for researching the center point of the subdivision grid in the three-dimensional region. The primary field and the scattered field of the electromagnetic field at each observation point in the ground well electromagnetic shaft are efficiently calculated.
Owner:GUILIN UNIVERSITY OF TECHNOLOGY

Three-dimensional target detection method based on stereoscopic vision technology

The invention discloses a three-dimensional target detection method based on a stereoscopic vision technology, and the method comprises the following steps: obtaining a to-be-detected target image, and carrying out the preprocessing of the to-be-detected target image; acquiring depth information of the target image to be detected by adopting a binocular stereoscopic vision technology; extracting target features in the preprocessed to-be-detected target image by using the optimized CNN model; fusing the depth information of the to-be-detected target image with the target features; and identifying the fused target features by using an RPN algorithm based on a regional proposal network. The method solves the problems that in an existing three-dimensional target detection method, when depth information in an image is obtained, the image and point cloud information of the environment where a target object is located need to be obtained firstly, then coordinate information of the point cloud information in the image is obtained, and then the coordinate information is combined with 2D and 3D bounding box information in the image, so that the depth information in the image cannot be obtained. The process relates to a large amount of calculation, so that the calculation complexity is increased, and the efficiency of obtaining the depth information is reduced.
Owner:SUZHOU XIANYANG ROBOT CO LTD

Geographic information data acquisition vehicle

The invention discloses a geographic information data acquisition vehicle, relates to the technical field of geographic information data acquisition vehicles, and aims to solve the problem that data acquisition is incomplete when a data acquisition vehicle performs geographic information acquisition. According to the method, data from different sensors are fused by using a particle filtering method, advantages of various sensors can be comprehensively considered, uncertainty and errors of single sensor data are reduced, scene model construction is performed on fused feature information by using a neural network model algorithm, and the accuracy of scene modeling is improved. The standard geographic data model of the data acquisition area can be quickly and accurately generated. The system can reflect the change of an actual scene in real time, can adapt to various complex environments and scenes, and combines a satellite navigation sensor, a laser radar sensor, a camera, a millimeter wave radar and inertial navigation. Data including geographic positions, environment information, images and videos, dynamic states of target objects, body motion states and the like can be obtained.
Owner:INST OF GEOGRAPHY HENAN ACAD OF SCI

Intelligent driving front-view image sensing algorithm based on heterogeneous convolution

The invention discloses an intelligent driving front-view image perception algorithm based on heterogeneous convolution, and the algorithm comprises the steps: S1, carrying out the basic feature extraction of an original image through an initial convolution layer, and enabling the original image to enter a plurality of parallel convolution units which are integrated together; s2, performing feature fusion on the output of the plurality of parallel convolution units in a weighted fusion mode; s3, stacking a plurality of parallel convolution units into a backbone network, and performing multi-scale target detection through an FPN feature pyramid; and S4, after the detection of the targets of multiple scales is finished, completing the training of the intelligent driving front-view image perception algorithm model. According to the intelligent driving front-view image perception algorithm based on heterogeneous convolution, the reasoning efficiency and the feature extraction capability of the model can be remarkably improved, and the balance problem of the lightweight target detection model between the calculation complexity and the feature extraction capability is solved.
Owner:CHANGZHOU XINGYU AUTOMOTIVE LIGHTING SYST CO LTD

Method and system for processing coal mine disaster data based on principal component analysis method

ActiveCN119442103Aeasy to understandReduce calculation stepsFeature vectorPrincipal component analysis
The invention provides a method and system for processing coal mine disaster data based on a PCA (principal component analysis) method. The method comprises the following steps: preprocessing and standardizing working data to obtain effective data, performing dimension reduction analysis on the effective data by using PCA to convert the effective data into a plurality of feature vectors, and constructing a dimension reduction space of the effective data, and respectively calculating a reconstruction error and an interpretation variance ratio corresponding to each feature vector in the dimension reduction space, constructing dimension information corresponding to each feature vector to adjust the dimension reduction space, inputting effective data into the adjusted dimension reduction space for processing, generating dimension reduction data of the coal mine, and storing the dimension reduction data of the coal mine. And abnormal dimension reduction data segments which do not conform to the specified interpretation variance ratio are screened to obtain corresponding disaster data of the coal mine under different detection dimensions, so that the calculation steps can be effectively reduced, the analysis cost is reduced, whole-course visual analysis is realized, managers can conveniently understand the data, and the data processing efficiency is improved. And a powerful reference is provided for the safety of coal miners and the environment.
Owner:BEIJING LANTRACK INFO TECH CO LTD

Prediction method for hydration expansibility of compacted bentonite material

PendingCN121787001AResolve technical shortcomings that predict uncertaintyRefined modelingGeometric CADDesign optimisation/simulationClassical mechanicsBentonitic clay
The invention provides a method for predicting hydration expansibility of a compacted bentonite material, which comprises the following steps of: generalizing soil particles into a spherical shape, and dividing a compacted bentonite structure into a characteristic structure element consisting of four particles based on a space accumulation configuration; according to constraint conditions of apparent density, elastic deformation energy minimization and structural entropy maximization of the soil body structure, the distribution probability of each characteristic structure element is solved; a Hertz contact model is adopted to calculate the expansive force between particles in the characteristic structure elements; calculating the equivalent expansive force of each characteristic structure element through a finite element method; based on the Hill lemma, the macroscopic hydration expansibility of the compacted bentonite structure is predicted. According to the method, the prediction process is simple, the calculated amount is moderate, the hydration expansibility of the compacted bentonite structures with different dry densities can be accurately and theoretically predicted, and theoretical guidance can be provided for design and prediction of the hydration expansibility of the compacted bentonite structures in the environments of refuse landfills, nuclear waste repositories and the like.
Owner:TONGJI UNIV

Prediction method, device, storage medium and product for rotational equivariant physical quantity

The present application discloses a prediction method, device, medium and product for rotational equivariant physical quantities, and relates to the technical field of electrical digital data processing. Among them, the prediction method for rotational equivariant physical quantities includes: constructing and training a nonlinear equivariant deep learning model, the nonlinear equivariant deep learning model includes a graph neural network and a decoder with linear computational complexity; when the training of the nonlinear equivariant deep learning model is completed, the physical system model of the material structure to be tested is input into the nonlinear equivariant deep learning model; the graph neural network is used to extract rotational equivariant features; the nonlinear rotational invariant features corresponding to the rotational equivariant features are converted according to the gradient induction mechanism to obtain nonlinear rotational equivariant features; the decoder is used to decode the nonlinear rotational equivariant features to obtain the rotational equivariant physical quantities of the material structure to be tested. The present application can solve the technical problem that the prior art is difficult to simultaneously ensure the prediction accuracy and computational efficiency of rotational equivariant physical quantities, which limits the application of material structures.
Owner:ARTIFICIAL INTELLIGENCE RES INST OF HEFEI COMPREHENSIVE NAT SCI CENT (ANHUI ARTIFICIAL INTELLIGENCE LAB)

Rectangular aperture curved surface shape error compensation method and system

The invention provides a rectangular aperture curved surface shape error compensation method and system, and belongs to the field of free-form surface optical element ultra-precision machining.The method comprises the steps that the surface of a workpiece is machined based on an optimal moving path, coordinates and actual height values of all points on a curved surface shape are collected, and machining errors are obtained; performing unit square normalization on the processing error curved surface by using a Zernike polynomial, and then fitting a prediction error curved surface; and the predicted error curved surface is negated to correct the optimal moving path in the machining process, and the surface of the workpiece is remachined till the machining error meets the machining precision. The method strictly controls the precision and range of error compensation, and ensures that the compensated curved surface can meet the design requirements.
Owner:HUAZHONG UNIV OF SCI & TECH

Checking calculation method, device and system of redundant braking system and medium

The invention provides a checking calculation method, device and system for a redundant braking system and a medium, and the checking calculation method comprises the steps that the whole vehicle mass, the wheel rolling radius and the deceleration of a target vehicle are substituted into a first mathematical model, and a target braking torque is obtained; the braking parameters and the target braking torque are substituted into the second mathematical model, and target hydraulic pressure is obtained; substituting the first performance parameter into the third mathematical model to obtain the displacement of the plunger pump; establishing a simulation model according to the displacement, the second performance parameter and the brake liquid demand, and performing simulation to obtain a P-t relation curve of the pressure buildup hydraulic pressure and the pressure buildup time; substituting the target hydraulic pressure into the P-t relation curve to obtain a first time value corresponding to the target hydraulic pressure; and when the first time value is smaller than the maximum time value required by the working condition to reach the corresponding deceleration, judging that the plunger pump meets the working condition requirement. The parts of the redundant braking system can be checked and calculated, so that whether the system meets the working condition requirement or not is judged.
Owner:CHINA FAW CO LTD

Evaluation method for TNT equivalent of high-energy propellant

The invention discloses a method for evaluating the TNT equivalent of a high-energy propellant. The method comprises the following specific steps: step 1, obtaining enthalpy of formation and an assumed chemical formula of a solid propellant; 2, according to the formation enthalpy of the known solid propellant, the assumed chemical formula and the product composition, the mole number of all combustion products at the adiabatic combustion temperature of 1 kg of the solid propellant is calculated; 3, calculating the total enthalpy of combustion product generation of 1kg of solid propellant under 298K; 4, subtracting the total enthalpy of the combustion product under 298K from the formation enthalpy of the known solid propellant to obtain a theoretical constant pressure detonation heat value QP of the propellant; 5, obtaining theoretical constant-volume explosion heat Qv by utilizing the theoretical constant-pressure explosion heat value of the propellant and the expansion work of the generated gas-phase component; 6, the theoretical constant volume explosion heat of the propellant is divided by the explosion heat of the standard TNT, and the theoretical TNT equivalent coefficient CTNT of the propellant is obtained. The evaluation method has the advantages of few calculation steps, high efficiency, high calculation precision and wide application range.
Owner:INNER MONGOLIA INST OF POWER MASCH

A front wheel rotation angle sensor with adjustable range

This utility model relates to a front wheel steering angle sensor with adjustable range, including a sensor body, a magnetic block, and a PCB board. The sensor body is connected to the vehicle body through mounting holes, and its built-in center calibration structure is aligned with the steering axis. The magnetic block contains a radially magnetized bare magnet, which forms a magnetic field with the Hall chip on the PCB board. The PCB board is rigidly fixed to the body and is detachable, integrating a Hall sensor chip with an SPI interface for programmable operation, supporting software configuration of the detection range and zero-point parameters. The signal is converted from 3.3V to 5V via a level conversion chip, and a multi-stage power supply module provides dual-channel regulated power. After the microprocessor calculates the angle data, it is output by the CAN bus controller. This sensor establishes an initial zero point through a zero-point calibration component, featuring a compact structure and convenient installation. The software configuration is flexible and adaptable to different range requirements, making it particularly suitable for autonomous driving systems in construction machinery and agricultural machinery, effectively improving steering angle detection accuracy and multi-platform compatibility.
Owner:TIANJIN ENG MACHINERY INST

Radar real-time target detection method based on heterogeneous parallel computing

The invention belongs to the technical field of signal processing, and discloses a radar real-time target detection method based on heterogeneous parallel computing. According to the method, through the steps of channel correction, pulse compression, moving target display, digital beam forming, coherent accumulation or non-coherent accumulation on multi-pulse signals including sum beams and difference beams, and constant false alarm processing, the problems of an existing radar real-time signal processing technology in the aspects of processing speed, accuracy, efficiency, flexibility and the like are solved. By adopting a domestic heterogeneous computing technology and applying a universal software technology, efficient and real-time processing of radar signals is realized, and the overall performance of a radar system is improved.
Owner:NANJING RES INST OF ELECTRONICS TECH

Inverter grid-connected and off-grid detection method, device and equipment and storage medium

The invention relates to the technical field of photovoltaic energy storage inverter grid-connected and off-grid detection, and discloses an inverter grid-connected and off-grid detection method, device and equipment and a storage medium, and the method comprises the steps: judging whether an inverter is in a grid-connected operation state or not; when the inverter is in the grid-connected operation state, acquiring the fundamental voltage of the inverter and the sampling voltage of the power grid at the current moment; after virtual power grid voltage is obtained through calculation based on the fundamental wave voltage, phase voltages of the same phase between the virtual power grid voltage and the sampling voltage are compared; and in a preset time, if the difference value of the phase voltage of at least one phase is out of a preset range, determining that the inverter is in an off-grid state. According to the method, false triggering detection is not easy, the calculation process related to grid-connected and off-grid detection of the power grid voltage is simplified, the problem that the proportion of chip resources is too high during calculation is solved, and meanwhile the accuracy and rapidity of grid-connected and off-grid detection are improved.
Owner:GOODWE TECHNOLOGIES CO LTD

Low-bit matrix computation device with floating-point scaling factor and computation method thereof

ActiveCN121350393Bsolve computational puzzlesSolve the complex calculation processAlgorithmFloating point
The application discloses a low-bit matrix calculation device with a floating-point scaling factor and a calculation method thereof. The device comprises a scaling factor processing module, an input module, a matrix calculation module, a floating-point accumulation module and an output module. The method first groups and quantizes weight and eigenvalue matrices to obtain quantization values and floating-point scaling factors; then, the scaling factor product is calculated in advance with a single accumulation length as a grouping unit; then, matrix multiplication and accumulation operation is performed, the result is multiplied by the scaling factor product after being converted into an fp16 or bfp16 format, and then accumulated into a final result; and finally, the result is output after being dequantized. The application simplifies the group-wise calculation process, saves hardware resources, guarantees calculation accuracy, meets different scene requirements and improves calculation efficiency by sharing the scaling factor product and using fp16 or bfp16 as an accumulator.
Owner:JINDIE SPACETIME (BEIJING) TECHNOLOGY CO LTD

Creeping formwork construction method for sea-crossing cable tower bridge main tower

The invention relates to a creeping formwork construction method for a sea-crossing cable tower bridge main tower, and belongs to the technical field of creeping formwork construction.The creeping formwork construction method comprises the steps that firstly, when a first layer of steel bars of the main tower are bound, needed embedded parts are embedded, first layer of concrete of the main tower is poured, and after the strength of the concrete of the main tower reaches the standard, formwork removal is conducted, and wall attaching devices are installed at the embedded parts of the first layer of the main tower; secondly, a truss is assembled, the assembled truss is hoisted and hung on a wall attaching device which is buried when a first layer of the main tower is poured, embedded parts are positioned and installed on formworks, the formworks are assembled, a second layer of concrete of the main tower is poured, and a third layer of steel bars is bound; thirdly, after the concrete strength of the main tower reaches the standard, the formwork is removed, the formwork is moved backwards, a wall attaching device and a hoisting guide rail are installed at the embedded part, a hydraulic system is established, and a synchronous climbing monitoring system is arranged on the hydraulic system; fourthly, the hydraulic system climbs, and meanwhile it is guaranteed that the truss and the formwork are synchronously lifted through a hydraulic synchronous climbing monitoring system; the positioning accuracy is high.
Owner:CHINA RAILWAY CONSTR BRIDGE ENG BUREAU GRP CO LTD +2

Quantum circuit simulation method based on circuit cutting

The invention belongs to the technical field of quantum computing, and particularly relates to a quantum circuit simulation method based on circuit cutting. By designing a cutting strategy, a complex quantum circuit is divided into a plurality of small sub-circuits, and each sub-circuit can be independently simulated; the method specifically comprises the steps of circuit preprocessing and dependency relationship analysis; sub-circuits are divided through circuit cutting, specifically, a heuristic cutting algorithm is adopted, and an optimal cutting point is determined according to the dependency relationship between the circuits; selecting a scheme capable of minimizing the overall simulation burden by evaluating the overhead of different cutting schemes to optimize the global calculation cost and resource allocation so as to ensure that the cut sub-circuits can be independently and efficiently simulated; and sub-circuit state vector multiplexing: storing the intermediate state vector of each sub-circuit, and searching similar state vectors for multiplexing when a new sub-circuit is calculated, so as to reduce the calculation amount and resource consumption. According to the method, the complexity of large-scale quantum circuit simulation can be reduced, and the calculation efficiency is improved.
Owner:FUDAN UNIVERSITY

Secure starting method and device of embedded system, equipment and storage medium

The invention discloses a safe starting method, device and equipment for an embedded system and a storage medium, and the method comprises the steps: obtaining to-be-transmitted system mirror image data through a host, and carrying out the fragmentation of the system mirror image data according to a specified data length, so as to generate data fragments; hash fragmentation processing is conducted on the data fragments through the host to generate an abstract chain, and the abstract chain comprises fragment serial numbers, abstract intermediate values and signatures of the data fragments; when the signature verification result of the abstract chain is that the signature verification is successful, reading each data fragment through the embedded device in a multi-thread manner, and determining an intermediate value verification result of each data fragment according to the abstract chain; system mirror image repairing is carried out through the embedded device according to the intermediate value verification results to generate repairing mirror image data, and partition safety starting is carried out according to the repairing mirror image data. By fragmenting the abstract process, the calculation steps are greatly reduced, the performance is higher, the occupied hardware resources are small, and the data transmission speed is increased while the data security is ensured.
Owner:HONGHU WANLIAN (JIANGSU) TECH DEV CO LTD

Game matching method and device, electronic equipment and readable storage medium

The invention provides a game matching method and device, electronic equipment and a readable storage medium. A matching mode is determined based on the number of participated players in a current game; when the matching mode is a first matching mode, distributing a preset number of to-be-matched game accounts to a target matching room based on the game information of each to-be-matched game account, and dividing a plurality of to-be-matched game accounts contained in the target matching room into a first preset number of target matching teams to obtain a matching result; and when the matching mode is a second matching mode, dividing the to-be-matched game accounts into a second preset number of target matching teams based on the game information of each to-be-matched game account, and dividing target matching rooms according to the matching teams to obtain a matching result. Therefore, matching requirements of players can be met, different matching modes are adopted for different matching modes, matching calculation steps are reduced, and matching efficiency is improved.
Owner:NETEASE (HANGZHOU) NETWORK CO LTD

A method for calculating the maximum power of feeder lines in distribution network based on flexibility resources

A method for calculating the maximum power of distribution network feeder on-line based on flexibility resources, the method includes S1, obtaining grid parameters; S2, calculating grid demand response load; S3, calculating the confidence output value of distributed photovoltaic; S4, calculating the confidence output coefficient of distributed photovoltaic; S5, calculating the maximum power of grid area on-line; S6, calculating the boundary interface value of the maximum power of each unit on-line in each area; S7, calculating the maximum power of 10kV feeder on-line. This design can not only calculate the maximum power supply capacity of the feeder by using multiple flexibility resources of source, grid, load and storage, effectively improving the accuracy of the maximum power supply capacity of the feeder, but also does not need to establish a hierarchical partition model of the power grid, effectively reducing the calculation steps and amount.
Owner:ECONOMIC & TECH RES INST OF HUBEI ELECTRIC POWER COMPANY SGCC

Adaptive question answering method based on large language model and intelligent large language model

The invention discloses a self-adaptive question answering method based on a large language model and an intelligent large language model. The method comprises the following steps: receiving a query text input by a user; analyzing the query text by using a preset query analysis model to determine the complexity of the query text; if the complexity is a first category, the query text is directly input into a large language model to generate an answer, and the first category corresponds to a preset low-complexity query; if the complexity is a second category, the retrieval process is executed, a reference document related to the query text is obtained from a preset data source, the query text and the reference document are jointly input into the large language model to generate answers, and the second category corresponds to preset high-complexity query. According to the method, query analysis is introduced, so that shunting processing, simple direct answering and answering after complex retrieval on queries with different complexities are realized. The self-adaptive mechanism maximizes the operation efficiency of the system while ensuring the answer quality.
Owner:CHINA MOBILE (SUZHOU) SOFTWARE TECH CO LTD +1

Quantum HNSW graph KNN method based on pellets

The invention relates to a particle-ball-based quantum HNSW graph KNN method, and belongs to the field of quantum calculation and machine learning. According to the method, firstly, a data set D is converted into a granular ball data set M through classical calculation, and the amount of data participating in calculation is effectively reduced; afterwards, by means of quantum computing parallelism, a QRAM algorithm and angle coding, the pellet data set is coded into a quantum state, Swap-test and iteration are used for comparing quantum circuits to calculate similarity, and a multilayer quantum HNSW graph is constructed; and finally, K nearest neighbors of the test data points are determined through quantum search and a priority queue, and the test data points are classified according to classification results of the neighbors. According to the method, the advantages of a classic computer and a quantum computer are combined, the calculation data volume and the time complexity are reduced, meanwhile, high classification precision is kept, and a new thought and method are provided for the quantum machine learning field.
Owner:SICHUAN KAIXIANGYUAN SOFTWARE TECH CO LTD

Object identification method based on edge vectorization coding

ActiveCN119027748AImprove relationshipPowerful self-checking capability for coding errorsCharacter and pattern recognitionNeural architecturesPattern recognitionThumbnail
The invention discloses an object identification method based on edge vectorization coding, and the method comprises the steps: firstly carrying out the edge detection of an image, and carrying out the pruning of a detected edge; coding is carried out according to line segments based on an edge vectorization coding mode; performing primary decoding on each line segment code, converting the line segment codes into images, performing rotation, thumbnail and refitting processing on the images in sequence, and performing secondary coding; performing bitwise exclusive-or operation on each line segment code obtained by secondary coding and each model line segment code of the standard edge model, and judging whether line segments corresponding to the two codes of the operation are matched or not according to the bit proportion of 0 in the operation result; and when most of all the line segments have matched model line segments, considering that the edge detected in the image is an object corresponding to the standard edge model. According to the method, through the form of edge vectorization coding, the calculation steps are reduced, the operand is reduced, the lightweight requirement of image recognition is met, and the recognition precision is ensured.
Owner:JIANGSU PEREGRINE MICROELECTRONICS CO LTD +1

Adaptive efficient CAN and Ethernet translation system, method and device and storage medium

The invention discloses a self-adaptive efficient CAN and Ethernet translation system and device and a storage medium, and the system comprises a data packet detection module, a data pre-shaping module, a data path analysis module, a data composition analysis module, a data load analysis module and a data transmission module. According to the system, feature identification code insertion, data field segmentation scribing and data flow direction path definition are carried out on data through the data pre-shaping module, data signal parameterization can replace addressing, data can be grabbed and taken away at the same time, the calculation step of the data forwarding routing process is greatly shortened, the forwarding communication overhead of non-demand data is reduced, and the data forwarding efficiency is improved. The routing translation efficiency is effectively improved; meanwhile, an upper computer interface is provided, effective data packet extraction is carried out during gateway forwarding, rapid and convenient dynamic configuration is provided, address information of bottom layer signals does not need to be concerned during dynamic configuration, decoupling of an application layer and a bottom layer is achieved, the dynamic configuration efficiency is high, technical operability is simple and practical, and actual requirements are better met.
Owner:DONGFENG OFF ROAD VEHICLE CO LTD

Design method of high-speed dual-channel parallel 16b18b encoder based on HDMI2.1 protocol

The invention provides a design method of a high-speed two-channel parallel 16b18b encoder based on an HDMI2.1 protocol, and the method comprises the steps: carrying out the pre-analysis and calculation of a 16b18b encoding table, obtaining an optimized code table, designing a control module based on a BMG, replacing a conventional 9b / 7b encoding module and an RD calculation module, and carrying out the design of a high-speed two-channel parallel 16b18b encoder based on the HDMI2.1 protocol. Therefore, the 16b18b coding module can output a result in one clock delay. Besides, a dual-channel parallel redundancy encoder architecture is designed based on the novel 16b18b encoding module, internal RD dependence is eliminated, real parallel work is realized, and internal work clock frequency multiplication operation is not needed. And after 32-bit data passes through the improved double-channel 16b18b encoder, 36-bit coded data can be output. Compared with the prior art, the method has the advantages that 1) the clock frequency required by the interior of the encoder is reduced, so that the encoder works more stably; 2) under the clock frequency of 400 MHz, the coding rate can reach 14.4 Gbps; and 3) the coding module saves the intermediate calculation process and simplifies the development difficulty.
Owner:GUILIN UNIV OF ELECTRONIC TECH

Image recognition method and device based on deep learning and storage medium

The invention relates to the technical field of image processing, in particular to an image recognition method and device based on deep learning, computer equipment, a storage medium and a computer program product. The method comprises the following steps: acquiring an initial feature recognition network model, and screening out a target processing unit meeting a preset removal index from processing units of the initial feature recognition network model; removing the target processing unit from the initial feature recognition network model to obtain a target feature recognition network model; obtaining a to-be-processed image, and processing the to-be-processed image based on a preset preprocessing process to obtain a target image; inputting the target image into a target feature recognition network model for processing to obtain a recognition result; and determining feature information of a feature unit in the to-be-processed image according to the identification result. By adopting the method, the image recognition efficiency can be improved.
Owner:CHANGZHOU MANCHEN TECH CO LTD

A method for quickly decomposing and typesetting mixed content of text, emoji expressions, and emoticons

ActiveCN118211560BReduce calculation stepsprocessing speedNatural language data processing
The present invention discloses a method for quickly decomposing and typeset mixed contents of text, emoji expressions and emoticons, comprising the following steps: S1, decomposing input text character by character to form a plurality of independent and inseparable character groups; S2, traversing the character groups, and performing Unicode category calculation on the first character of each character group to determine its type; S3, marking each character group with a label attribute according to the contextual relationship and characteristics of the character group; S4, recombining into a plurality of inseparable character groups according to the label attribute according to the maximum width of a display text area; S5, placing the inseparable character groups into a display area. The method is a highly customized text typeset method based on Unicode character classification, and can process different character types to adapt to various situations and requirements. Because it is based on Unicode, the method supports the vast majority of languages, symbols and characters in the world, including special characters such as emoji expressions and emoticons, and is more extensive and flexible.
Owner:SUZHOU YOUBISI EDUCATION TECH CO LTD

Visual drilling simulation test platform

The invention relates to a visual drilling simulation test platform, and belongs to the technical field of drilling simulation, the visual drilling simulation test platform comprises a main platform, four main supports are fixedly mounted at the bottom of the main platform, four auxiliary supports are fixedly mounted at the top of the main platform, and an auxiliary platform is fixedly mounted among the tops of the four auxiliary supports; a drilling mechanism is fixedly installed at the top of the auxiliary platform, a display screen is fixedly installed at the top of the main platform, a camera is fixedly installed at the bottom of the auxiliary platform, a drill bit assembly is installed at the output end of the drilling mechanism, and a fixing frame is placed between the opposite sides of the main supports on the left side and the right side. A land simulation assembly is installed at the top of the fixing frame, a wellhead protection assembly is installed at the top of the land simulation assembly, and an emptying assembly is installed at the bottom of the fixing frame. According to the visual drilling simulation test platform, the compression condition and the appearance condition of a drill bit in the drilling process can be displayed more visually.
Owner:CHINESE PEOPLES LIBERATION ARMY UNIT 31666

Multi-prediction-frame real-time display method for target detection algorithm acceleration based on FPGA

The invention provides a multi-prediction-frame real-time display method for target detection algorithm acceleration based on an FPGA, and relates to the technical field of target detection deep learning algorithms, and the method comprises the steps: carrying out the post-processing of each block feature map of a regression branch, and generating N prediction frames; the coordinate [(x11, y11), (x21, y21)] of the prediction frame bbox1 is analyzed, the original y11 row mask is read from the BRAM to the one-dimensional register set to serve as an initial value, and the y11 row mask is updated on the one-dimensional register set from the x11 column to the x21 column and written back to the BRAM; updating masks of all the remaining rows from y11 + 1 to y21 in the prediction frame bbox1 row by row and writing the masks back to the BRAM; according to the method, mask updating of the remaining bbox < 2 >-bbox < N > is completed one by one and written back to the BRAM; and carrying out picture drawing on the video stream according to all the updated masks, if the mask is 0, passing through the original picture data, and if the mask is 1, replacing the original picture data with the color of the prediction frame. According to the method, the mask of the picture can be generated in real time, the mask does not need to be generated after all prediction frames are generated, and the reasoning speed can be effectively improved.
Owner:CHENGDU SINO MICROELECTRONICS TECH CO LTD