62results about How to "Fast operation" patented technology

The application provides a frequency domain convolution operation acceleration system for an 8bit frequency domain convolution neural network, relates to the technical field of frequency domain convolution operation acceleration design, and solves the problems of slow convolution calculation of the current frequency domain convolution neural network, difficulty in deployment to a hardware platform with limited resources, influence on product market time and design cost, and the like, obtains a pair of 8bit multiplication data, packs the obtained pair of 8bit multiplication data, performs multiplication calculation on the packed pair of 8bit multiplication data, outputs a frequency domain convolution multiplication result, accumulates the frequency domain convolution multiplication result, outputs a frequency domain convolution accumulation result, restores the frequency domain convolution accumulation result, and obtains a final frequency domain convolution operation complex output result; the application improves the operation speed of the frequency domain convolution, facilitates deployment of the frequency domain convolution to a hardware platform, improves the efficiency of processing digital signals in an FPGA accelerator, accelerates product market time and reduces product design cost.

The invention provides a marine target detection method based on an unmanned aerial vehicle, and the method comprises the following steps: 1, carrying out the time-space synchronization of point cloud original data collected by a plurality of sensors on the unmanned aerial vehicle, and obtaining the point cloud original data after the time-space synchronization; step 2, preprocessing the point cloud original data after time-space synchronization acquired by each sensor, and obtaining preprocessed single-sensor point cloud data through voxel filtering and statistical filtering; step 3, based on a K-dimensional tree and a DBSCAN clustering method, performing clustering segmentation on the point cloud data of the single sensor to obtain a single sensor target detection result; and 4, performing comprehensive analysis on all single-sensor target detection results by adopting a multi-source point cloud information decision level information fusion method to obtain a final target detection result.

This invention provides an intelligent front-end data acquisition system for boilers, comprising: a detection module, a data acquisition module, and a monitoring module; the detection module is used to acquire unit temperature data; the data acquisition module is a YS-24D data acquisition box, which is used to acquire data from the detection module. This invention provides an intelligent front-end data acquisition system and device for boilers, wherein the data acquisition module uses the YS-24D data acquisition box to acquire data from the detection module at unit temperature measurement points. The YS-24D data acquisition box is high-performance and highly reliable, with a hardware design employing a 32-bit low-power ARM processor, high integration, few peripheral components, high reliability, fast processing speed, high resolution, and reliable and accurate data. It uses a high-performance, high-precision 24-bit A / D converter, with all 24 channels being independent fully differential input loops, providing high resolution, low drift, and high stability.

A broadband voltage transformer error compensation method is proposed. First, a broadband test platform is constructed using a signal generator, power amplifier, standard voltage transformer, and high-bandwidth data acquisition system. The ratio and phase errors of the transformer at multiple frequency points are acquired through primary-side excitation and secondary-side synchronous acquisition. Based on the slope, first derivative, and second derivative characteristics of the error curve, the entire frequency band is adaptively divided into segments. Within each segment, multiple candidate models are fitted using a polynomial or exponential-linear combination model. The optimal model is selected based on the coefficient of determination, mean absolute error, and root mean square error. Ratio and phase error compensators are constructed based on the fitting functions of each segment, and error compensation is achieved using a feedforward approach. This method can significantly reduce amplitude and phase errors over a wide frequency range, improving the dynamic measurement accuracy of the voltage transformer.

The application discloses a laser vision real-time weld seam tracking method based on a lightweight target detection network, and the method comprises the following steps: before welding, an industrial camera collects an initial weld seam image and transmits the initial weld seam image to an embedded industrial computer, pixel coordinate values of initial weld seam feature points are acquired, and a starting position of the weld seam is obtained; a lightweight target neural network is constructed; during welding, the industrial camera continuously collects weld seam images and sends the weld seam images to the embedded industrial computer, pixel coordinate values of weld seam feature points in the weld seam images are extracted by using the lightweight target neural network; a welding position to be reached by a robot is obtained through the pixel coordinate values of the weld seam feature points; a difference between the welding position of the robot and a current position of the robot is calculated, and a deviation value is sent to a control cabinet, so that the welding gun of the robot is controlled to automatically track and move along the weld seam. The weld seam image is detected by using the lightweight target neural network, and the problem that the real-time rate of the existing weld seam tracking method is not high on a resource-limited device is solved.

The application provides an integrated circuit yield estimation method and a memory, and the method comprises the following steps: formulating a boundary of a success region and a failure region of an integrated circuit performance in a parameter space, and representing the boundary as a boundary surface formula equation; calculating a size of a space defined by the boundary surface formula equation and a boundary surface formed by each parameter axis; and calculating a ratio of the size of the space defined by the boundary surface formula equation and the boundary surface formed by each parameter axis to a size of a space formed by each parameter range, to obtain an estimated yield. The technical scheme of the application can improve the estimation efficiency of the integrated circuit yield.

This invention relates to the field of five-axis machining technology and discloses a method for measuring the Z-axis value of the workpiece coordinate system origin and the tool length. The method involves placing the Z-axis setter on the upper surface of the workpiece; with the spindle stopped, lifting the handwheel pulse generator and gently pressing the spindle end face against the elastic upper plane of the Z-axis setter; adjusting the movement ratio of the handwheel pulse generator to "X1"; moving the spindle away from the Z-axis setter; mounting the tool on the spindle; adjusting the movement ratio of the handwheel pulse generator to "X1" again, and the pointer of the Z-axis setter pointing to the "0" mark; due to the addition of the Z-axis setter, the Z-axis value h0 of point B in the workpiece coordinate system in the machine tool coordinate system A is -h - h1 - 50, where 50 is the fixed height of the Z-axis setter. This method ultimately yields the Z-axis value of the workpiece coordinate system origin. This method reduces the cost for companies to purchase on-machine measuring instruments and tool pre-setting devices, achieving high operation speed, high efficiency, high precision, and excellent cost-effectiveness.

This invention discloses a coal spill identification method based on information entropy, relating to the field of detection technology. The method includes the following steps: Step 1: For each frame image of the monitored area, convert it into a grayscale image and represent it using matrix A; Step 2: Perform edge detection on matrix A, and record the result as matrix C; Step 3: Calculate the information entropy of matrix C to obtain the frame entropy H of the grayscale image; Step 4: Process each frame of the real-time frame images generated sequentially in the monitored area through steps 1 to 3 to obtain a frame entropy sequence; Step 5: Calculate the frame entropy mean μ and variance σ of the frame entropy sequence. 2 Determine the threshold H for frame entropy alarm. gate Step 6: When the frame entropy of several consecutive frames exceeds the threshold H gate If a coal spill occurs, a shutdown command is sent. This invention utilizes a pure algorithm, avoiding the time-consuming and labor-intensive cleanup caused by coal blockage, and providing a reliable technical guarantee.

The application relates to a plane circle center point-based SolidWorks standard part automatic installation method for a parameter-free model, and belongs to the field of computer-aided design and three-dimensional mechanical design automation. The method first selects any part plane and any assembly hole on the part plane through a point selection mode; a sketch is created based on the geometric boundary of the part plane, and the coordinates of the sketch are determined according to the geometric boundary position information of the part plane; all assembly holes with the same hole feature on the part plane are automatically identified according to the extracted selected assembly hole feature, and corresponding sketch coordinates are established; sketch points are established for all sketch coordinates under the part plane; and the sketch-driven part array function is called, after manually assembling a standard part at any sketch point, the standard part assembly is automatically executed at all sketch points. The application can automatically batch install standard parts by using the sketch-driven part array function of an assembly body.

The application discloses an electric read-write memory device based on 180-degree flip of anti-ferromagnetic Neel vector, a preparation method and a read-write method. The electric read-write memory device comprises a substrate layer, an anti-ferromagnetic layer and a spin source layer which are sequentially stacked from bottom to top and constitute a heterostructure. The material of the substrate layer is a single crystal wafer capable of epitaxially growing the anti-ferromagnetic layer. The material of the anti-ferromagnetic layer is a collinear anti-ferromagnetic material with abnormal Hall effect or a staggered magnet with abnormal Hall effect. The material of the spin source layer has a spin current generation capability. The memory utilizes a current pulse to drive 180-degree flip of the Neel vector. The Neel vector before and after the 180-degree flip has two abnormal Hall resistance states of high and low, and can be used for storing information "0" and "1". The application not only enriches the connotation of the interaction between the spin current and the anti-ferromagnetic Neel vector, but also lays a foundation for constructing a high-density, high read-write speed and low-power nonvolatile magnetic random memory.

The embodiment of the specification discloses a high-imitation fake certificate detection method, comprising: acquiring multiple frames of certificate images under different environments collected by a client; performing block division on each frame of certificate image respectively by using the same division mode; performing feature extraction on each block in each frame of certificate image to obtain block features corresponding to each block respectively; performing spatial coding on the features of each block according to the position of each block in the certificate image to obtain spatial features corresponding to each block in the certificate image; for each position divided in any certificate image, performing time coding on the spatial features of the blocks located at the position in different frames of certificate image to obtain space-time features corresponding to the position, so as to represent the changes of the part in the certificate located at the position under different environments; and judging whether the certificate is a fake certificate according to the space-time features corresponding to each position respectively. Accordingly, the application discloses a high-imitation fake certificate detection device.

The application provides a remote sensing target matching method based on a convolutional neural network of fused local features, compared with a CNN feature based on an overall image, the CNN feature extracted at a region level has higher robustness and flexibility in identifying small similar regions between partial matching images, a matching region pair is accurately matched through a generating step, which makes the CNN feature extracted from the regions more robust for partial repeated image matching, because a region suggestion model in a rough division efficiently matches local features between images to select a small number of top candidate images, which can greatly reduce the target detection range of the partial repeated images, and the application can realize relatively accurate matching because the good robustness of SIFT features and the high resolution capability of CNN features are fully utilized.

The present application relates to a kind of computing method for realizing wide range dynamic noise map system fast update, comprising the following steps: using the data information of system database, based on existing day and night period and city road grade, simplify the structure of traditional road sound level mechanism model;Using big data technology, using hybrid modeling method, regression each grade road sound source (day and night) sound level contribution amount prediction model;Road noise fast calculation method based on sound level contribution amount prediction model is constructed, the fast update of dynamic noise map is realized.Compared with prior art, while ensuring the calculation precision, the operation and update efficiency of noise map system is greatly improved, and the technical constraints of foreign core computing software are broken, the localization of noise map core computing technology is realized, which provides technical support for wide range deep development of noise map.

The invention discloses an alternating current real-time parameter rapid correction method, which relates to the technical field of alternating current parameter correction, and comprises the following steps of: 1, establishing an alternating current real-time parameter coordinate system; 2, selecting a plurality of standard calibration points in the coordinate system; 3, selecting a plurality of to-be-corrected points in the coordinate system, and calculating a broken line slope K corresponding to each to-be-corrected point; and 4, directly updating the to-be-corrected point through the broken line slope K to obtain corrected data. According to the invention, a curing operation can be carried out before real-time measurement is started, and a large amount of resources occupied by division operation of a single-chip microcomputer are avoided, so that the operation speed is improved, and the accuracy of a measurement result is ensured. The method is suitable for measurement of real-time electrical parameters of alternating current and other basic measurement fields, and has a wide application prospect.

The application discloses a kind of self-adapting terminology normalization methods based on double-tower model. The original word of terminology is retrieved in standard terminology dictionary to retrieve multiple similar standard terminology of the original word of terminology;Sample pair is formed, and balanced processing obtains balanced sample pair set;Balanced sample pair set is input into Sentence-BERT double-tower model for training, Sentence-BERT double-tower model outputs label and prediction result, and the standard terminology dictionary is processed using the trained Sentence-BERT double-tower model, to obtain sentence vector and save in offline vector database;The original word of terminology to be predicted is processed using the trained Sentence-BERT double-tower model, and then combined with offline vector database processing to predict and obtain high similarity standard terminology, and then match additional to the original word of terminology to be predicted.The application can normalize and match a non-standard terminology text, optimize processing in industrial scene, has small amount of calculation, fast operation speed, and greatly improves the efficiency of matching normalization.

The application proposes an underwater DOA and SAP estimation method based on fast multi-snapshot non-inverse sparse Bayesian. The method introduces the smooth function property and generalizes to a complex matrix form. Based on the generalized property, the method is introduced into multi-snapshot sparse Bayesian learning, and the matrix form of the relaxation evidence lower bound (relax-ELBO) is derived. Based on the relax-ELBO, the VEM method is used to solve the posterior distribution approximation of the hidden variable. According to the posterior distribution approximation, the DOA and SAP are estimated, the matrix inversion is avoided, the operation speed is greatly improved, and the limitation that the SBL is difficult to be applied in large-scale field and online is overcome.

The embodiment of the invention discloses a semiconductor test structure and a thickness measurement method. The semiconductor test structure comprises a substrate, wherein a groove is formed in the substrate; the epitaxial layer is formed based on the groove, and the epitaxial layer comprises a buffer layer, a seed crystal layer and a silicon cap layer which are stacked in sequence; the buffer layer covers the bottom surface of the groove; the seed crystal layer is in a convex shape and is positioned above the buffer layer; the silicon cap layer is positioned above the seed crystal layer and covers the convex part of the seed crystal layer; and the combination of the silicon cap layer and the seed crystal layer is in a halfpace shape protruding above the opening of the groove. According to the semiconductor test structure and the thickness measurement method provided by the embodiment of the invention, the size information of each layer in the epitaxial layer can be quickly and accurately obtained.

The present application relates to the field of agricultural machinery control technology, and discloses a tractor power reversing control method, controller, system and medium. The method comprises the following steps: when the gear change information of the tractor shows that the driver intends to control the tractor to enter the forward gear or the reverse gear, the neutral switch valve is controlled to be closed so that the clutch assembly oil circuit is correspondingly turned on, and a pressure control strategy based on twice pressure pre-charging and time-sharing pressure boosting is correspondingly executed for the forward proportional valve or the reverse proportional valve until a preset maximum pressure ensuring no-slip engagement of the corresponding clutch is reached, so that the corresponding proportional valve is opened and the pressure drives the corresponding clutch to engage, thereby driving the gearbox to switch gears. The present application controls the corresponding proportional valve based on the pressure control strategy of twice pre-charging and time-sharing pressure boosting to drive the clutch to engage quickly, and at the same time avoids that the impact of clutch engagement is too large.

The invention discloses a vehicle-mounted ISP image follow-up visual field processing module. The module comprises a camera, an FPGA (Field Programmable Gate Array) programmable logic device and an image data storage DDR (Double Data Rate), the FPGA programmable logic device comprises a CAN (Controller Area Network) bus interface, an MIPI (Mobile Industry Processor Interface) input interface and an MIPI output interface; an ISP visual field processing program is arranged in the FPGA programmable logic device, the ISP visual field processing program is burnt in the FPGA programmable logic device, the ISP visual field processing program receives image data from the camera through the MIPI input interface and stores the image data in the image data DDR memory, the ISP visual field processing program receives corner information through the CAN bus interface, and the corner information is stored in the image data DDR memory. The ISP visual field processing program processes the image data, and then the MIPI output interface outputs image visual field information. When the commercial vehicle turns, the condition of a wider area can be observed, and driving safety guarantee is provided for a driver.

The invention provides an automatic tool magazine, and is applied to the technical field of automatic tool magazines. The automatic tool magazine comprises a shell, an operation window is formed in a front plate of the shell, and the bottom end of a shell top plate of the shell is fixedly connected with the top end of the main magazine frame. The main garage frame comprises a first garage frame, a second garage frame, a third garage frame and a fourth garage frame; the first garage frame is provided with a plurality of first garage frame columns, and the first garage frame columns are fixedly connected with the first garage frame plate; the first garage frame and the second garage frame are adjacent and aligned, the third garage frame and the fourth garage frame are adjacent and aligned, and the length direction of the first garage frame is parallel to the length direction of the fourth garage frame; a sliding rail is arranged between the first garage frame and the fourth garage frame, the length direction of the sliding rail is parallel to the length direction of the first garage frame, the sliding rail is slidably connected with the bottom end of a sliding base, and the top end of the sliding base is fixedly connected with the bottom end of a six-axis robot. A clamp frame is arranged between the sliding rail and the fourth library frame; the inner end face of the shell front plate is fixedly connected with the transfer table. In this way, the tool storage efficiency can be improved.

The utility model discloses a kind of stacking mechanical arm and five-axis stacking robot, including turnover mechanism, elastic buffer mechanism and the suction cup assembly for adsorbing and grabbing, turnover mechanism includes turnover driver, first mounting plate, second mounting plate and the pivot between first mounting plate and second mounting plate, turnover driver is assembled in second mounting plate, and its driving end is rotationally connected in first mounting plate, to drive first mounting plate overturn relative to second mounting plate, suction cup assembly is connected in first mounting plate, and elastic buffer mechanism is used to provide compression buffer force for suction cup assembly, so more degrees of freedom can be increased to grasp posture, to realize complex multi-pile mode such as plane stacking and vertical stacking.

The utility model discloses an unwinding device of label processing equipment. The unwinding device of the label processing equipment comprises a winding machine and a winding core which is detachably installed on the winding machine and used for winding label paper. The winding machine comprises a machine base, a motor is arranged on the machine base, and the motor can drive a rotating shaft connected with the motor to rotate. The roll core comprises a fixed section and a movable section, one ends of the two parts are detachably connected, the two parts can be combined into a complete cylinder after being connected, and the roll core is detachably mounted on the rotating shaft; the label paper is clamped at the detachable connecting end of the fixed section and the movable section to be fixed. According to the unwinding device of the label processing equipment, the effect that label paper is wound on the winding core is achieved by arranging related components such as the rotating shaft, meanwhile, the label paper can be rapidly fixed to the winding core through the fixed section and the movable section of the winding core, and compared with an existing through groove fixing mode, operation is simpler and more convenient, and the production efficiency is improved. And the operation speed of workers can be improved.

This invention relates to a high-precision pose measurement method based on docking surface features, to solve the technical problems of high equipment cost, complex calibration process, and susceptibility to industrial environment in existing large component assembly methods. The method includes: 1. building an assembly platform; 2. calibrating the intrinsic parameters of a monocular vision sensor; 3. calibrating the monocular vision sensor and motion execution components using a hand-eye calibration method; 4. calibrating the corresponding target and pin hole; 5. measuring the pose of the first component relative to the second component.

This utility model discloses a combined synchronous display and controller utilizing AC sampling, including a current sampling circuit and a first voltage sampling circuit connected to the machine side, a second voltage sampling circuit connected to the system side, and a microprocessor. The microprocessor includes a first set of synchronous sampling AD ports and a second set of synchronous sampling AD ports. The current sampling circuit and the first voltage sampling circuit are connected to the first set of synchronous sampling AD ports, the first voltage sampling circuit is connected to the second set of synchronous sampling AD ports, and the second voltage sampling circuit is connected to the second set of synchronous sampling AD ports. By simplifying the number of sampling channels, replacing some hardware circuits with digital calculations performed by the microprocessor, and combining RC filtering with a unified bias reference, the system achieves simplified circuit structure, high sampling accuracy, strong anti-interference capability, accurate synchronization judgment, and effective electrical parameter measurement.

The application discloses a boundary inside and outside identification method of an electromagnetic shielding intelligent mower based on a DTW algorithm. First, members of a normal effective signal library are established under the condition of no background noise, and inside and outside classification and marking are carried out; an inductance probe of the intelligent mower periodically collects an induced current signal, and time domain characteristic values are extracted from the induced current signal. Second, the induced current signal is subjected to background noise suppression, and approximate effective signal waveform segments are intercepted and sorted and combined. Finally, matching value calculation of the approximate effective signal waveform segments and the normal effective signal library is realized based on the DTW algorithm, and the optimal reference signal is found by combining the time domain characteristic values of the segments and the matching values, so that the determination of the inside and outside states of the intelligent mower boundary is realized. The application has strong anti-interference ability, is suitable for a wide range of scenes, and can add effective signal templates under specific environments to the normal effective signal library, so that the accuracy of the inside and outside determination of the system boundary is improved.

The application discloses a kind of large-scale MIMO multi-star positioning and direction estimation method and system based on angle information.The application first obtains the channel matrix estimation between each satellite and user by sending pilot signal;Then the estimated value of channel parameter is obtained using tensor-ESPRIT algorithm, including angle of departure and angle of arrival;According to the estimated angle of departure, the direction vector between user and each satellite is obtained, the optimization problem on spherical surface is established, considering the altitude constraint, the user position estimation value is obtained by using Riemann conjugate gradient method;Finally, according to the estimated user position and angle of arrival, the optimization problem on SO (3) manifold is established, and the estimated value of user array rotation matrix is obtained by using Riemann gradient descent method.The application combines the channel information between multiple satellites and users for positioning and direction estimation, and considers the altitude constraint during positioning, compared with single satellite and unconstrained case, can significantly improve the positioning performance, while having good direction estimation performance.

The application discloses a kind of quick cross path scheduling methods of inter-city area customized public transport system, the shortest distance method between regions is used to build start point area end point set and end point area start point set, to generate start point area travel sequence Yo by insertion method, to generate end point area travel sequence Yd by insertion method, then the Yo and Yd are optimized by quick exchange method, finally the minimum value of start point × Yo × Yd is solved, which is the final inter-city area customized public transport system target path, with the shortest bus travel distance as the target;The method does not use probability distribution, effectively avoids the fluctuation of the final path result, according to the characteristics that the distance between nodes in different regions is much larger than the distance between nodes in the same region, the inter-regional traffic node set is established, and it is used as the end point of the start point region and the start point of the end point region, respectively, to maximize the range of optimal solution, effectively reducing the overall calculation scale.

The application belongs to the field of analog integrated circuit design, and particularly relates to a redundant current range current steering DAC based on digital pre-distortion. The application carries out rotation correction through an improved DEM decoding circuit, measures and quantizes error codes corresponding to the static mismatch error of each current unit, and reduces the circuit area and the influence of additional circuit parasitic on the DAC speed without additional control circuit. A redundant current unit is introduced at the highest bit to expand the quantization range, and the error code and clipping problem introduced by DPD correction are solved. Only a simple algorithm of a 14-bit adder and a 14-bit subtractor is needed to directly operate on binary codes, and a secondary decoding is used to realize thermometer code addition operation compatible with DEM, so that the timing delay and circuit complexity of the thermometer code addition circuit are greatly reduced. The circuit timing delay of the application meets the speed requirement of high-speed and high-precision current steering DAC, and compared with DEM correction only, the application significantly improves the SFDR without raising the noise floor.

The application discloses a kind of three-level n-dimensional digital twin systems supporting coupling operation of coal mine robot group, including terminal layer, edge layer and cloud end layer three levels.The robot group in terminal physical layer generates single-machine behavior level twin in terminal virtual layer based on real-time data through terminal AI model, realizes virtual and real two-way control, and field operator obtains augmented information through AR equipment, feeds back decision-making opinion, and multi-modal control node;Edge virtual layer generates collaborative task level twin based on single-machine behavior level twin, and constructs n-dimensional edge twin system through edge AI model.The computer host of cloud end physical layer provides computing power support for n-dimensional cloud end twin system, and centralized control operator accesses cloud end twin system through VR equipment, to realize remote monitoring and control;Cloud end virtual layer generates global planning level twin based on edge side data, and parallel analysis of n-dimensional cloud end twin system is realized through cloud end AI model, and the optimal global planning is comprehensively obtained.