37results about How to "Avoid calculation" patented technology

The invention discloses an anaphora disambiguation method and system based on big language model enhanced text and structured query language generation. The method comprises the following steps: receiving a natural language question of a user, and analyzing and generating structured table field information through mode information; identifying and rewriting fuzzy time expression in the problem; extracting ambiguous entities, and replacing the ambiguous entities with database standard values through semantic matching and fuzzy matching; synthesizing the information to generate a structured query language statement and executing the structured query language statement; and if the execution result is null, automatically triggering an anaphora disambiguation process, performing error correction and re-matching on field values in the statement through a dual matching mechanism, and generating and executing a corrected query statement. According to the method, the problems of fuzzy anaphora, indefinite time expression, low matching accuracy, resource waste and the like in a traditional SQL system are effectively solved through a strategy of combining pre-processing and post-processing, and the accuracy, robustness and execution efficiency of complex query are remarkably improved.

A converter station AC filter group circuit breaker abnormal arcing modeling analysis method comprises the following steps: firstly, carrying out time mark alignment, amplitude calibration and filtering de-noising on wave recording signals such as AC bus voltage and filter branch current to obtain high-quality input data; then, combining a primary wiring diagram, enabling an alternating current system to be equivalent to an ideal power supply and series impedance, enabling circuit breaker arcing to be equivalent to arc resistance changing along with time, and constructing an equivalent circuit model comprising the power supply, the impedance, a capacitor, an electric reactor and dynamic arc resistance; establishing a state space equation according to a voltage and current constraint relation, and performing time domain solution by adopting a two-step two-order or three-order diagonal implicit Runge-Kutta method to obtain voltage and current responses of the whole process of the fault; on this basis, the arc resistance change is inversed according to the ratio of arc voltage to current, the current zero passage is identified through the resistance symbol change at adjacent moments, and the zero passage section is smoothed; and finally, the inversion result is verified again by means of the simplified equivalent model.

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.

The application discloses a kind of based on benchmark magnetic flux state electromagnetic power network and electromagnetic transient load two-way power coupling method and device, it is related to pulse power network, electromagnetic transient load and plasma load coupling simulation technical field.The method includes: reading external electromagnetic power network state and magnetic flux state;Candidate external response is generated;Magnetic update entry arbitration is executed to candidate external response;Forming to be accepted magnetic flux state;From the projection of to be accepted magnetic flux state, a plurality of loop magnetic chains are obtained, and to be accepted counter electromotive force and network residual error are generated.Through the magnetic chain calculated by sharing magnetic flux state, external electromagnetic power network does not need to maintain second set of magnetic field state, and power closure, magnetic chain residual error and magnetic update constraint can be accepted simultaneously.

The application provides a converter valve area short-circuit fault positioning method, device, equipment and medium, the method comprises the following steps: dividing the converter valve fault into eight typical valve area short-circuit faults; RTDS simulation analysis, according to the different converter protection action results, dividing into three types of protection actions, giving the typical fault characteristic quantity of the valve difference protection action for positioning the valve area fault and the eight typical valve area short-circuit quick criterion; according to whether the actual converter protection action occurs in the rectifying station or the inverting station, confirming the valve area short-circuit fault occurring station; according to the converter protection action event of the monitoring background, confirming the converter protection action type and the valve area fault converter bridge; calling the built-in fault recording of the DC control protection, analyzing the converter protection action waveform, analyzing the fault characteristics of the current waveform of the valve side of the converter bridge and the outlet current waveform of the DC side, positioning the valve area short-circuit fault point according to the typical fault characteristic quantity of the valve difference protection action and the eight typical valve area short-circuit quick criterion; and has the advantages of fast positioning speed and simple analysis method.

The invention provides a surface clutter real-time generation method and device, and the method comprises the steps: determining a target time domain range and a target frequency domain range of a to-be-generated surface clutter signal according to the real-time motion state parameter of a carrier carrying a radar seeker and the beam pointing parameter of the radar seeker; generating a random clutter sequence covering the target time domain range and the target frequency domain range according to a preset amplitude statistical distribution model and a preset power spectrum model; a surface clutter signal is generated by modulating a reference radar signal, the modulation including convolving the reference radar signal using a random clutter sequence, and parameter adjusting the convolved signal to match dynamic characteristics of a carrier. According to the invention, the problem of insufficient fidelity of clutter simulation signals in the prior art is solved.

The invention relates to the technical field of image processing, and discloses an ore sorting image processing method, device and equipment, a medium and a product, the method comprises the steps that an X-ray image of to-be-sorted ore is acquired, and the to-be-sorted ore comprises a plurality of ore particles; determining a target pixel region of at least one target ore particle in the X-ray image, wherein the target ore particle is an ore particle with a particle size value greater than a preset particle size threshold value in the plurality of ore particles; and performing logarithmic transformation on the pixel value of each pixel point in the target pixel region to obtain a target image of the target ore particles. According to the method, target ore particles with the particle size exceeding a preset threshold value in a plurality of ore particles are focused in a targeted manner, the pixel difference of an ore spot area originally covered by an overall deep shadow can be effectively amplified by means of the characteristic of compressing the dynamic range of a bright part and expanding the dynamic range of a dark part of logarithmic transformation, so that the pixel difference can be clearly highlighted from the overall shadow; and the distinguishing degree of the concentrate area is greatly improved.

The invention discloses an unmanned aerial vehicle cluster collaborative navigation method realized based on a factor graph, and relates to the technical field of unmanned aerial vehicle collaborative navigation. According to the method, a factor graph comprising variable nodes and factor nodes is constructed for each unmanned aerial vehicle, the variable nodes comprise the position, the speed, the attitude and the IMU zero offset of the unmanned aerial vehicle, and the factor nodes comprise an inertial measurement pre-integration factor, a visual measurement factor and a radio distance measurement and angle measurement factor. By initializing a factor graph, collecting sensor data in real time to update factors, and adopting an L-M algorithm based on Lie group and a sliding window to optimize the factor graph, the unmanned aerial vehicles communicate and share pose information to realize collaborative optimization. According to the method, the problems of high calculation complexity, low information fusion efficiency, poor expansibility and the like of cooperative navigation of the existing unmanned aerial vehicle cluster in a complex environment are effectively solved, cooperative navigation between the unmanned aerial vehicle clusters is realized through a factor graph mode, and the method has reliable navigation performance.

This invention provides a method, system, and related devices for adaptive state estimation of distribution network topology, belonging to the field of power system monitoring and control. It includes: obtaining measurement vectors and topology states based on acquired real-time operating data of the distribution network; using the measurement vectors and topology states as input to a physical information autoencoder, outputting estimated values ​​for the real and imaginary parts of the voltage of the distribution network; converting the obtained voltage real and imaginary part estimates into voltage amplitude and phase angle estimates to complete the adaptive state estimation of the distribution network topology. The physical information autoencoder sequentially includes an input projection layer, a first residual network module, a topology-aware latent module, a second residual network module, and an output projection layer. This invention integrates the advantages of physical constraints and data-driven approaches, possessing high estimation accuracy, excellent computational efficiency, strong physical consistency, and adaptive topology changes, effectively adapting to operating scenarios with frequent changes in distribution network topology.

The application discloses a long-range ground laser point cloud plane segmentation method, and specifically comprises the following steps: T1, determining the best neighborhood based on a dynamic search range, T2, designing a region growing algorithm based on dimensional features and the best neighborhood, and relates to the technical field of laser point cloud data processing application. The long-range ground laser point cloud plane segmentation method has density adaptability, the neighborhood search range is automatically generated based on a theoretical point spacing, and the needs of point cloud neighborhood searching at different scanning distances can be met; the construction of the best neighborhood guarantees that the calculation of point cloud local features is not affected by density changes; the estimation of the sampling interval conforms to the ground laser scanner acquisition principle, and the rationality of the theoretical point spacing is enhanced; the best neighborhood is added to the region growing, the plane segmentation algorithm is adapted to high-density change point clouds, has wide applicability of the dynamic search range, and the determination of the neighborhood search range starts from a single point, and the overall density change of the point cloud does not need to be analyzed.

The application provides a road surface state recognition method and device, and solves the problems of single road surface state recognition category, coarse result granularity and inability to distinguish specific road surface abnormal types caused by dependence on simple threshold judgment and shallow features in the prior art. The method comprises the following steps: acquiring a vertical vibration acceleration data segment of a vehicle in operation; extracting a characteristic value according to a preset optimal feature subset, wherein the subset comprises a frequency domain peak value and a spectrum energy; inputting the characteristic value into a classification model to obtain a recognition result, wherein the model is obtained by training based on data comprising at least three road surface state labels. Through fixed feature optimization, the application realizes fine recognition of multiple states such as well frame difference, expansion joint and deceleration strip, and significantly improves the calculation efficiency while ensuring the accuracy.

This invention relates to a method for estimating the radiant energy emitted from the top of the atmosphere based on high-orbit satellite sensors. The method calculates the radiance of the radiant energy emitted from the top of the atmosphere using the following formula: where F is the radiance of the radiant energy emitted from the top of the atmosphere, and E... m Ω represents the energy emitted from the top of the atmosphere received by high-orbit satellite sensors. e To obtain the solid angle of observation of the top of the atmosphere by a high-orbit satellite sensor, the solid angle can be solved using a complete elliptic integral of the third kind. A rectangular coordinate system based on the solid angle is then constructed to calculate the boundary curve of the observation field of view. The corresponding geocentric and geofixed coordinates are obtained through a coordinate transformation matrix. This invention can accurately estimate the emitted energy from the top of the atmosphere, providing support for related research.

The application discloses a charging remaining time confirmation method, device and vehicle, relates to the battery charging technical field, and discloses a charging remaining time confirmation method, which comprises the following steps: obtaining current charging data of a battery pack, wherein the current charging data comprises a charging rate and an SOC value; obtaining a temperature rising charging time and a temperature rising charging end SOC value based on the current charging data and a first charging model; obtaining a constant temperature charging time based on the temperature rising charging end SOC value and a second charging model; and confirming a charging remaining time based on the temperature rising charging time and the constant temperature charging time.According to the basic characteristics of the temperature in the low-temperature fast charging process, the charging remaining time is calculated in sections, the fast and accurate calculation of the fast charging remaining time is realized in combination with a fitting method and given data, and the overcomplicated calculation and the model building and calculation of heat generation, heat conduction and heat dissipation are avoided.

The application discloses a guided wave tomography method based on full waveform inversion of a recurrent neural network, comprising the following steps: step 1, time series forward modeling based on the acoustic formula of the finite difference method is realized by using sequence calculation of the recurrent neural network, and an initial model is obtained; step 2, the wave field value of each time step at each excitation position is calculated in turn, and a simulation value is generated; step 3, after actual observation values are processed, a least square target equation of the simulation value and the observation value is generated; step 4, the least square target equation is iteratively minimized by an inversion part, including an automatic differentiation method and an Adam algorithm, which are respectively responsible for gradient calculation of the target equation and iterative updating of a velocity model and a UNet weight; step 5, the updated velocity model is subjected to depth image prior UNet regularization; and step 6, steps 1 to 5 are repeated, and the next iteration calculation is performed until the target equation is lower than a specified threshold value, so that a final velocity model is obtained, and the final velocity model is converted into a thickness model by using a dispersion relation.

This invention provides a trajectory recognition and grouping method based on low-precision random sampling data, comprising the following steps: Step 1, acquiring low-precision random sampling trajectory point data; Step 2, preprocessing the low-precision random sampling trajectory point data to obtain trajectory point data with the same time and trajectory point data with different times; Step 3, constructing a distance matrix M using the trajectory point data with different times. D and time matrix M T Step 4: Based on the distance matrix M D and time matrix M T Construct the spatiotemporal reachability matrix M ST Step 5: Based on the spatiotemporal reachability matrix M ST Step six: Based on the spatiotemporally reachable key points, group the trajectory points by acquiring the set of spatiotemporally reachable adjacent points; Step seven: Aggregate the trajectory point data with the same time frame according to the trajectory point grouping in step six. This invention has higher timeliness and wider applicability.

The utility model provides a kind of wafer temperature uniform control and automatic transfer integrated device in hot plate air extraction process, including disc-shaped cavity main body and automatic wafer transfer assembly, cavity main body top is plane structure, hot plate upper surface is perpendicular to the central axis of cavity main body, the sealing disc is coaxially arranged with cavity main body;It further includes air inlet and air outlet, air inlet is circularly opened in the sidewall top of the cavity main body;Air outlet is opened in the top surface center of the cavity main body.The utility model determines suitable cavity wall surface material by optimizing cavity structure, reasonably selecting air inlet and air outlet position, solves the problem of wafer temperature uneven in prior art, and realizes wafer automatic taking through intelligent automatic wafer transfer assembly structure under the form of not changing cavity structure, improves the quality and efficiency of semiconductor wafer manufacturing.

The invention provides a submarine cable burial depth monitoring method, a computer program product and electronic equipment. The method comprises the following steps: acquiring basic temperature data of a to-be-measured space point in a first submarine cable in each unit time of a preset time period; the basic temperature data comprises soil temperature data; obtaining a feature set of the spatial point to be measured according to the basic temperature data; and inputting the feature set into a burial depth monitoring model to obtain burial depth data of the to-be-measured spatial point. And reasoning the feature set by using the burial depth monitoring model to obtain the burial depth data of the to-be-measured spatial point without calculating the heat dissipation condition, thereby reducing the error introduced by parameter estimation in the calculation, avoiding a large amount of grid calculation, saving the calculation resource, reducing the complexity of submarine cable burial depth measurement and calculation time consumption, and improving the calculation efficiency. And the burial depth measuring and calculating efficiency is improved.

The embodiment of the invention provides a content generation method and related equipment, and the method comprises the steps: indicating the execution of a calling reasoning model through business indication information related to a content generation business, when de-noising processing is carried out on corresponding input content coding characteristics of the noisy data content under a time step tj in the T time steps under the constraint of the first data content, obtaining a cache type for cache multiplexing according to content analysis parameters indicated by a cache multiplexing strategy related to the time step tj; and if the cache type is an inter-step cache type, multiplexing a predicted noise feature corresponding to a time step ti having an inter-step multiplexing relationship with the time step tj, performing denoising processing on an input content coding feature corresponding to the time step tj, and performing content generation based on an output content coding feature corresponding to the time step tj obtained by the denoising processing, the second data content is obtained. Therefore, the model reasoning speed can be effectively improved, and waste of computing resources is avoided.

This application provides a fast query method, apparatus, medium, program product, and terminal for mask rule checking based on directional grouping indexing. The method includes: acquiring mask image data after optical proximity correction; extracting features from the mask image data to obtain a feature line segment set; grouping all line segments in the feature line segment set according to their directions to obtain subsets of line segments in different directions; constructing a corresponding bucketed index structure for each line segment subset using a bucketing strategy; and performing fast parallel line segment querying for mask rule checking based on the bucketed index structure constructed for each line segment subset. This application provides an indexing method optimized for parallel line segment queries through directional grouping and bucketed indexing techniques. This not only significantly improves the efficiency of parallel line segment rule verification in MRC checking but also avoids the use of complex R-Tree structures, reducing memory consumption and making it suitable for large-scale mask data in advanced processes.

The application relates to the technical field of artificial intelligence, and provides a data processing method and device, electronic equipment and a storage medium, which comprise the following steps: obtaining a rearranged tensor, loading a target discrete index sequence required for current calculation; according to the target discrete index sequence, determining a plurality of corresponding discrete physical block addresses and continuously writing the discrete physical block addresses into an on-chip cache to obtain a continuously stored physical addressing list; and based on the physical addressing list, calling a hardware aggregation reading instruction to perform discrete addressing on a target addressing dimension, so that corresponding data in the rearranged tensor is batched from a global memory to a shared memory. Through the dimension rearrangement of the input tensor and the continuous caching of the discrete physical addresses, and in combination with the discrete addressing of the hardware aggregation reading instruction, the discrete discontinuous data is efficiently batched from the global memory to the shared memory, so that the advantages of breaking through the discrete access performance bottleneck, filling the specific hardware ecological blank and reducing the long text calculation complexity are achieved.

The application provides a method, system, device and medium for screening a region suitable for low-altitude logistics layout, relates to the fields of low-altitude economy and geographic information, and comprises the following steps: acquiring navigation electronic map data, DEM elevation data and administrative division data, calculating the average altitude of each township based on the DEM elevation data and the administrative division data of the whole country to generate a preliminary screening region; calculating the elevation characteristics and slope characteristics of each road in the preliminary screening region by using the DEM elevation data to form a road elevation-slope characteristic data set; calculating the road bypassing coefficient, the number of sharp bends and the convenient accessibility index in the preliminary screening region based on the road elevation-slope characteristic data set and the navigation electronic map data, and constructing a road traffic obstacle index set; and evaluating the road traffic obstacle index set by using the analytic hierarchy process, and screening a region suitable for low-altitude logistics layout according to the evaluation result.

This utility model discloses a prefabricated building wall structure, which includes a T-shaped strip plate located between two wall panel bodies. The wall panel body has a fitting groove adapted to the T-shaped strip plate on the side closest to the T-shaped strip plate. Two hooks are embedded in one side of the cavity of the fitting groove. The T-shaped strip plate has two cavities inside, and two symmetrical pins are fixedly connected to the cavities. This utility model installs the two wall panel bodies by bringing them close to the T-shaped strip plate, causing the fitting groove to fit against the surface of the T-shaped strip plate. Then, by rotating a threaded rod, a rectangular frame descends, and the sliding pin slides within the rectangular frame, causing the hook to rotate. The hooks of the hooks engage with the cavities of the hooks, thus installing the two wall panels. This achieves the goal of conveniently connecting adjacent wall panels using concealed hooks, avoiding calculations and measurements, and improving installation efficiency.

The application discloses a high real-time license plate recognition system and method based on FPGA. The system comprises a camera driving module, an SDRAM control module, an image processing module and a VGA display module. The image processing module accurately locates the license plate boundary through RGB to YCbCr conversion, binarization, erosion and expansion, Sobel edge detection and row and column projection technology; then, character binarization and projection are carried out in the license plate area, character feature values are extracted, and a character recognition is realized by using a Xor template matching algorithm. The application utilizes the parallel processing capability and the pipeline design of FPGA, realizes low-delay (at most three frames) real-time license plate recognition and result visual display, and has high real-time performance, strong stability and good hardware compatibility.

The invention provides a submarine cable burial depth monitoring method, a program product and electronic equipment. The submarine cable burial depth monitoring method comprises the following steps: acquiring load data and temperature data of a to-be-measured space point in a first submarine cable in at least one load change process; obtaining a load characteristic value based on the load data; obtaining a temperature response characteristic value based on the temperature data; obtaining a feature set of the spatial point to be measured according to the load feature value and the temperature response feature value; and inputting the feature set into a burial depth monitoring model to obtain burial depth data of the to-be-measured spatial point. The burial depth monitoring model is used for reasoning the feature set to obtain the burial depth data of the to-be-measured space point, and the heat dissipation condition does not need to be calculated, so that errors caused by parameter estimation in the calculation are reduced, a large amount of grid calculation is avoided, calculation resources are saved, the burial depth measurement complexity of the submarine cable is reduced, and the calculation time consumption is reduced; and the burial depth measuring and calculating efficiency is improved. And the mode is lower in use threshold and convenient to operate.

The application discloses a coarse-grained soil natural and underwater repose angle measuring device based on a tilt positioning technology, which comprises a slope accumulation system, a water storage and drainage system, a pile releasing system and a repose angle measuring system, wherein the water storage and drainage system comprises a water storage tank, the slope accumulation system, the pile releasing system and the repose angle measuring system are fixed in the water storage tank, the pile releasing system is arranged at the tail end of the slope accumulation system, the slope accumulation system is provided with a vertical lifting surface, the pile releasing system is provided with a horizontal baffle, the height of the horizontal baffle is not lower than the height of the vertical lifting surface, the repose angle measuring system is arranged above the slope accumulation system, the repose angle measuring system is provided with a vertical buckle type sliding groove and a protractor, the left side of the vertical buckle type sliding groove is flush with the front end surface of the vertical lifting surface, and the lower end of the protractor is higher than the upper surface of the vertical lifting surface.

This invention belongs to the field of neural network technology, specifically relating to an iterative random pruning method and system based on meta-transfer learning. The pruning method of this invention includes the following steps: S1, a transfer learning step, pre-training the convolutional neural network using a large-scale dataset to initialize the network parameters; S2, a meta-learning step, meta-training the network using a small sample dataset; S3, a pruning step, pruning the network weights after meta-training; S4, a testing step, testing the network using a small sample dataset; and S5, a fine-tuning stage, retraining the network to restore its accuracy. Compared to existing technologies, this invention can reduce the number of parameters in a convolutional neural network with small sample datasets while maintaining the accuracy of the network model, achieving the goal of model compression.

The application discloses an object visibility information acquisition method and device, a storage medium and an electronic device. The method comprises the following steps: obtaining a baking grid, wherein the baking grid is a unit space grid divided in a virtual scene; obtaining at least one rendering camera configured in the baking grid, wherein the rendering camera is used for rendering an object in the virtual scene; rendering the object in the virtual scene by using the at least one rendering camera to obtain rendering information, wherein the rendering information is used for representing visual performance of the object under a visual angle of the at least one rendering camera; and obtaining visibility information of the object in the baking grid based on the rendering information. The application solves the technical problem of low efficiency of obtaining object visibility information.

The invention discloses a point cloud noise filtering method and device and related equipment, and the method comprises the steps: carrying out the downsampling of first point cloud data, obtaining second point cloud data, and enabling a laser radar to detect a target region to obtain the first point cloud data; screening out a candidate noise point cloud set from the second point cloud data, the distance of each point cloud in the candidate noise point cloud set being within a preset distance range, and the reflection intensity of each point cloud being within a preset first intensity range; based on the number of point clouds in each voxel of a point cloud grid, the position of each point cloud and the reflection intensity, whether each point cloud in the voxel is noise is judged, and the point cloud grid is generated based on each point cloud in the candidate noise point cloud set; and rejecting point clouds which are judged to be noise from the candidate noise point cloud set to obtain target point cloud data. According to the invention, miscellaneous points received by the eVTOL during vertical take-off and landing and low-altitude flight can be effectively filtered, and the stability of obstacle recognition is improved.

This invention relates to a sequential filtering method for data fusion in a rotorcraft unmanned aerial vehicle (UAV) attitude reference system, belonging to the field of UAV attitude estimation. In this invention, a low-pass filter is used to process the acquired sensor data to obtain the initialized horizontal attitude angle and heading angle, thus initializing the attitude reference system. During the calculation process, Euler angle differential equations are used to construct the system's state transition equations, defining relevant state variables and obtaining the state transition matrix. Measurement equations are constructed based on the acceleration output and magnetometer output of the carrier's observations, yielding the corresponding observation matrix. Following the steps required for Kalman filtering, the state variables required during the Kalman filter's operation are mapped to the system, and calculations are performed using the Kalman filter data update method to complete the calculation derivation of the corresponding quantities, thus realizing the sequential processing method of Kalman filtering. Compared to traditional Kalman filtering algorithms, this algorithm implements Kalman filtering sequentially, requiring only partial data calculations during attitude updates, and does not involve matrix operations throughout the process, thereby improving the algorithm's computational efficiency.

The application discloses a data processing method and device, a medium and equipment, relates to the technical field of databases, and comprises the following steps: determining at least one logical block in a to-be-executed transaction; creating a target cache object when a current traversal operation statement in the to-be-executed transaction is parsed, if the current traversal operation statement comprises a target logical block in the at least one logical block and the target logical block has not been pre-executed; pre-executing the target logical block to obtain a pre-execution result and storing the pre-execution result in the target cache object; updating the current traversal operation statement based on the target cache object to obtain a current target operation statement; generating a target execution plan of a physical operator associated with the target cache object according to the current target operation statement; and executing the target execution plan to obtain a plan execution result, wherein the pre-execution result is read from the target cache object when a physical operator is executed, and the plan execution result is obtained based on the pre-execution result. The application can effectively improve the execution efficiency of a transaction.