31results about How to "Reduce computational cost" patented technology

The invention discloses a DNA binding residue prediction method based on a multi-modal protein language model in the technical field of bioinformatics. The DNA binding residue prediction method based on the multi-modal protein language model comprises the following steps: sequentially using tools to respectively obtain feature files of a protein sequence P to be subjected to DNA binding residue prediction; inputting all the feature files into the ESM3 to obtain an embedded matrix M; the method comprises the following steps: processing a protein sequence into a residue sample, constructing an isovariant graph neural network model according to a protein structure diagram, constructing a data set by utilizing the protein sequence of known DNA binding residues, and training the constructed network; and inputting a residue sample of the protein sequence into the trained model to obtain a DNA binding residue of the protein sequence. According to the DNA binding residue prediction method based on the multi-modal protein language model, the efficiency and accuracy of DNA binding residue prediction are improved.

The invention provides a sliding bearing oil film lubrication reliability evaluation method, and relates to the field of oil film lubrication reliability diagnosis, and the method comprises the steps: building a minimum film thickness limit state function according to a bearing minimum oil film thickness calculation equation and a preset minimum oil film thickness threshold value, and determining a reliability domain, iteratively solving the minimum oil film thickness through an elastohydrodynamic lubrication equation, and solving a non-probabilistic reliability index and designing a checking point through a numerical optimization algorithm; according to a non-probability reliability index, designing a checking point and a relation between an uncertain domain and a reliable domain to judge a joint focal element needing extreme value analysis, inheriting maximum and minimum values of the joint focal element by using an intelligent genetic method, and obtaining credibility and likelihood under a threshold value; and according to the above steps, obtaining the credibility and the likelihood under each minimum oil film thickness threshold value, and determining an accumulated credibility function curve and an accumulated likelihood function curve used for evaluating the reliability degree. In this way, focal elements needing extreme value analysis are reduced, and the efficiency of evaluating the lubricating reliability of the oil film of the sliding bearing is improved.

The present invention belongs to the technical field of gas field exploitation, and discloses a numerical simulation method for shale condensate gas reservoir development based on a universal pEDFM, comprising the following steps: S1, using the universal pEDFM to calculate the conductivity of the connection between various grids; S2, constructing a commercial numerical simulation software CMG equivalent model based on the universal pEDFM; S3, calling the nonlinear solver in CMG to solve the shale condensate gas reservoir component model. The present invention adopts the above-mentioned numerical simulation method for shale condensate gas reservoir development based on the universal pEDFM, which not only ensures the advantages of the universal pEDFM in calculation accuracy and processing of complex fracture networks, but also avoids the difficulty of developing additional high-performance nonlinear solvers, and provides a numerical simulation tool with the best theoretical comprehensive calculation performance for the development of fractured shale condensate gas reservoirs.

The present invention provides a method and system for constructing a knowledge graph for mathematics courses, and relates to the technical field of knowledge graphs. The present invention solves the technical problem that the current relationship extraction method cannot accurately mine the entity relationship of mathematics courses, and realizes a relationship classification method by integrating sparse biaffine attention and adaptive threshold loss, overcoming the difficulties of extracting complex dependencies and overlapping relationships. At the same time, a hyperbolic space word embedding sparse projection model is used to predict the upper and lower relationships. The hyperbolic space has the ability to express hierarchical structures, can better predict the upper and lower relationships, and reduce the computational cost.

The invention discloses a method and device for recognizing an abnormal region in a CT image, a medium and a product, and relates to the technical field of abnormal region recognition, and the method comprises the steps: decomposing a to-be-detected CT image into a plurality of to-be-detected super voxels through employing a linear iteration clustering algorithm; determining any to-be-tested super voxel as a current to-be-tested super voxel; determining an abnormal value of the current to-be-detected super voxel based on a spectral space function principle and a healthy super voxel library; the abnormal value is used for expressing the probability that the super voxel region in the CT image to be detected is an abnormal region. According to the method, the abnormal region in the CT image is identified based on the spectral-space function, so that the accuracy and efficiency of identifying the abnormal region of the CT image can be improved.

The present invention discloses a grid adaptive turbulence simulation method based on Vreman dynamic coefficient coupled SA series model, and the main process of its implementation includes: step 1, judging whether to apply shielding function; step 2, calculating Vreman dynamic coefficient; step 3, determining local dynamic grid length scale; step 4, constructing scale-related adjustment function; step 5, reconstructing turbulent viscosity of SA series model; step 6, using reconstructed turbulent viscosity to perform turbulence simulation. The present invention constructs Vreman dynamic coefficient based on local velocity gradient tensor, identifies local grid size, and determines local grid length scale Δ * , the turbulent viscosity is reconstructed by constructing scale-related functions through turbulent energy spectrum integration, and a new algebraic empirical formula is constructed to characterize the turbulent kinetic energy and turbulent length scale. This overcomes the problem of high dependence of existing turbulent mixing models on grid experience, broadens grid adaptability, improves calculation accuracy, reduces calculation cost, and significantly accelerates the turbulence simulation process.

The present invention discloses a hyperspectral image compression and reconstruction method, device and equipment based on an autoencoder, which comprises: obtaining a Houston image data set and performing preprocessing including random cropping to generate an image block of a preset size as an original hyperspectral image; inputting the original hyperspectral image into an autoencoder model including an encoder and a decoder and training based on a preset loss function to obtain a trained hyperspectral compression model, wherein the network structure of the encoder includes a thumbnail generation branch and a potential encoding branch, and the network structure of the decoder includes a thumbnail decoding branch and a potential representation decoding branch; inputting a hyperspectral image to be compressed into the hyperspectral compression model, encoding the hyperspectral image to be compressed by the encoder to obtain compressed data, and decoding the compressed data by the decoder to obtain a reconstructed hyperspectral image.

The present invention discloses a method for eliminating the influence of BLE direction finding carrier frequency deviation to obtain phase delay, so that the phase delay only related to the direction of the incident signal source can be obtained with extremely low computing resources without estimating the carrier frequency deviation, thereby improving the accuracy of the BLE direction finding system in the presence of carrier frequency deviation and improving the stability of the BLE direction finding system under various operating temperature conditions.

The invention relates to the technical field of train aerodynamics optimization, in particular to a train surface bionic microstructure partition arrangement and drag reduction optimization method and system.The method comprises the steps that the train surface is divided into design partitions which are geometrically continuous and relatively consistent in near-wall flow form according to train appearance geometric parameters and external flow field characteristics; resistance mechanisms in the same subarea tend to be the same; a turbulent flow correction model is adopted for the design partition where the bionic microstructure is arranged, geometric parameters of the bionic microstructure are mapped into near-wall dissipation correction, and the wall surface shear stress is equivalently changed; an initial turbulence model is adopted for the non-arranged partition, and physical quantity continuity is kept at the junction through data exchange of a virtual grid; and a stable and comparable aerodynamic resistance evaluation result is obtained, and a target partition arrangement scheme is screened from candidate arrangement schemes, so that the reliability of resistance reduction optimization is improved at relatively low calculation cost.

The present application relates to the field of underwater vehicle flow field analysis, and discloses a variable reconstruction underwater vehicle flow field mutual interference analysis method, comprising: modeling a single underwater vehicle; based on the single underwater vehicle, constructing a variable reconstruction underwater vehicle; converting the variable reconstruction process into a series of transient processes parameterized by the axial spacing between the head end face and the tail end face of adjacent single underwater vehicles; building a fluid simulation calculation framework, and batch calculating the fluid dynamic parameters of different parts of the variable reconstruction underwater vehicle under different axial spacings corresponding to each transient process; based on the variation law of the fluid dynamic parameters with the axial spacing, determining a characteristic spacing from the axial spacing; using the velocity nephogram and the pressure nephogram, and combining the characteristic spacing and the fluid dynamic parameters of different parts of the variable reconstruction underwater vehicle, the flow field mutual interference is analyzed. The present application effectively reduces the calculation cost and improves the efficiency of flow field analysis.

The invention discloses a variable cycle combination method for multiple oil displacement systems based on a gradient boosting tree algorithm. The method comprises the following steps: S1, constructing a variable cycle combination large-scale learning sample of the oil displacement system; and S2, establishing a gradient boosting tree prediction model. According to the gradient boosting tree algorithm, a plurality of weak decision trees are integrated to gradually optimize model errors, flexibility and interpretability are kept, meanwhile, high robustness and over-fitting resistance are achieved, a plurality of weak learners can be weighted and combined in each step, and high prediction capacity and precision are achieved. According to the invention, on the basis of the gradient boosting tree algorithm, mining learning is carried out on large-scale learning samples by designing an appropriate deep neural network, the internal law of the numerical calculation process can be captured, and then time-consuming large-scale matrix iteration in numerical calculation is replaced; accurate prediction solving of the variable circulation oil displacement process of multiple oil displacement systems is achieved while calculation consumption is remarkably reduced.

The invention relates to a method for producing an electrode from an electrode-continuous web (9), comprising: a coating process in which the electrode-continuous web (9) is coated with an electrode coating (5); and a cutting process in which the uncoated region (11) is cut, in particular while forming the arrester flag (7). The method has a marking process in which an uncoated region (11) of the electrode-continuous web (9) is marked for product tracking in the case of origin certification or quality complaints. According to the invention, the marking process is an imprint process in which a continuous imprint pattern (19), which is void-free and varied in the production direction (F), is imprint as a mark into the uncoated region (11) of the electrode-continuous web (9).

The application discloses an optimization method of a rolling wing blade connecting rod mechanism, and belongs to the field of rolling wing aircrafts.The blade swing angle is represented by structural parameters of the blade connecting rod mechanism, and the structural parameters of the blade connecting rod mechanism, including the lengths of the connecting rods, the coordinates of key hinge points and configuration bias, are defined as a variable vector; multiple target functions are constructed for multiple key working conditions such as hovering, forward flight and transition, the multiple target functions not only quantify the performance of the connecting rod mechanism itself, such as kinematic fitting error with a target motion law, dynamic stability index and force transmission efficiency index, but also cover aerodynamic efficiency index and aerodynamic acoustic index which are directly related to the performance of the whole machine; and the optimal structural parameters of the blade connecting rod mechanism are obtained by solving the multiple target functions under constraints, so that the global optimization of the structural parameters of the blade connecting rod mechanism under multiple working conditions, multiple targets and multiple constraints can be realized.

The embodiment of the present invention discloses a vascular imaging method, device, equipment and storage medium, wherein the method comprises: obtaining an original filling image and at least two original mask images of a target acquisition part of a target object; determining an original mask image in the same heart beating state as the original filling image from the at least two original mask images as a mask image to be selected; determining a target mask image corresponding to the original filling image according to the mask image to be selected; and determining a vascular subtraction image of the target acquisition part according to the original filling image and the target mask image. The technical solution of the embodiment of the present invention improves the imaging quality of the vascular subtraction image by determining a target mask image that highly matches the original filling image from multiple original mask images.

The present application discloses a log template extraction method and device, which belongs to the field of computer technology. The method comprises: deleting specified characters in each line of log records in multiple lines of log records of a log to obtain each line of log records after preprocessing; determining at least one first log record group based on each line of log records after preprocessing, and different first log record groups include different lines of log records in the log; extracting log templates in each of the at least one first log record group, and determining the log template of the log based on the extracted log template. The present application solves the problem of low accuracy of the currently determined log templates, and the present application is applied to log template extraction of logs.

The present invention relates to a method for calculating the height of road facilities based on monocular camera images, including: calibrating a monocular camera to obtain preset parameters; using the calibrated monocular camera to collect the original image of the road facilities, and obtaining the ground point coordinates and vertex coordinates; using the mapping change matrix, the ground point coordinates and the vertex coordinates to calculate the pixel distance in the plan view image, and using the projection mapping matrix and the ground point coordinates to calculate the distance from the road facilities to the monocular camera; bringing the pixel distance of the road facilities in the plan view image, the distance from the road facilities to the monocular camera, and the distance from the imaging plane to the monocular camera position into a preset height model to calculate the height of the road facilities. The present invention can quickly calculate the height information of road facilities through monocular camera images, and fill in the gaps in the height of facilities currently realized in the intelligent detection of road facilities based on monocular camera images; and has the characteristics of low implementation cost, low calculation cost, high operation efficiency, etc.

The invention discloses a low-carbon economic dispatching method and device for a mixed hydrogen-electricity combined system, and belongs to the technical field of electrical control engineering.According to the method, the lowest system operation cost under the condition of considering carbon tax is taken as an optimization target, a discretization linear dynamic energy flow equation is taken as a mixed hydrogen net operation constraint, and a low-carbon economic dispatching result is obtained; a mixed hydrogen-electricity combined system low-carbon economic dispatching model is constructed, the dynamic characteristics in the mixed hydrogen natural gas transportation process can be described, the optimization problem solving requirement is met, and the change of gas components in the transportation process can be tracked to accurately reflect the running state of the mixed hydrogen natural gas system; the calculation cost for solving the nonlinear dynamic energy flow model in real time is reduced while the model characteristics are reserved; the low-carbon economic dispatching model of the mixed hydrogen-electricity combined system is solved for dispatching, it is guaranteed that the combined system always runs within the safety boundary, the renewable energy consumption capacity of the system is improved, and the running cost and carbon emission of the system are reduced.

The invention discloses a commutation branch tower wind disaster vulnerability evaluation method considering material parameter variation, which comprises the following steps of: during calculation of a wind-induced response of a commutation branch tower, considering a variation rule of various structural material parameters by using sampling operation, and learning a relationship between a mean value and a standard deviation of a wind speed and a structural demand index by using an intelligent prediction model; and finally, predicting the surpassing probability of the structure under a series of wind speed values, and forming a comprehensive wind disaster vulnerability result considering multi-structure material parameter variation. According to the method, more accurate beyond probability assessment is provided on the basis of the existing power transmission tower wind damage vulnerability assessment, and more uncertainty risks are considered. The uncertainty of various variation parameters is considered in a random sampling mode, the calculation amount is reduced, the quantitative relation between the wind speed and the override probability and the relation between the wind speed and the various material parameters and the override probability are obtained through the intelligent prediction model, and under the specified wind speed, the calculation amount is reduced, and the calculation efficiency is improved. The override probability can be calculated more quickly through the intelligent prediction model, a vulnerability curve or curved surface under multiple wind speeds is obtained, and the analysis efficiency is greatly improved.

The present invention discloses a multi-UAV self-organizing collaborative system and method, which realizes the communication networking and high-precision positioning of multiple UAVs through a differential satellite positioning system and Wi-Fi wireless communication technology, and designs a self-organizing collaborative control method with multiple functions such as autonomous interactive object selection, individual spacing maintenance, obstacle avoidance, target tracking, and viewpoint synthesis, thereby realizing autonomous communication and distributed autonomous control of multiple UAVs, and being able to provide a system solution for large-scale UAV self-organizing collaboration.

This invention discloses a low-carbon economic dispatch method and apparatus for a hydrogen-electricity combined system, belonging to the field of electrical control engineering technology. The method aims to minimize system operating costs under carbon tax considerations, using discretized linear dynamic energy flow equations as constraints for the operation of the hydrogen-electricity network. It constructs a low-carbon economic dispatch model for the hydrogen-electricity combined system, which can both characterize the dynamic characteristics of the hydrogen-natural gas transportation process and meet the requirements of optimization problem solving. It can accurately reflect the operating status of the hydrogen-natural gas system by tracking changes in gas composition during transportation. While preserving the model's characteristics, it reduces the computational cost of real-time solving of the nonlinear dynamic energy flow model. The method uses the low-carbon economic dispatch model of the hydrogen-electricity combined system for dispatching, ensuring that the combined system always operates within safe boundaries, improving the system's renewable energy absorption capacity, and reducing system operating costs and carbon emissions.

The present invention discloses a hyperspectral image band selection method based on group sparsity, and the method comprises the following steps: step 1, constructing a group sparse prior regularization; step 2, constructing a labeled indicator matrix; step 3, band selection of the hyperspectral image. The band selection method proposed in the present invention is a method that can be used to select important bands in a hyperspectral image. The method can effectively remove the bands that are severely affected by noise in the hyperspectral image, and at the same time can completely retain the important information in the original hyperspectral image. After the band selection method proposed in the present invention is used to select the bands of the hyperspectral image, the performance of the downstream classification task can be further improved. Therefore, the present invention has the characteristics of strong practical application value and good reliability, and is suitable for popularization and use. The present invention can provide a certain theoretical analysis basis for the further analysis and application of hyperspectral images by relevant professionals.

The embodiment of the invention provides a depth image acquisition method and device, electronic equipment and a storage medium. The method comprises the following steps: carrying out image matching on an obtained first image and an obtained second image shot by a binocular camera to obtain a parallax image; performing pixel remapping on the second image based on the parallax image to obtain a pseudo image corresponding to the second image; calling a pre-trained depth image prediction model to process the first image, the second image and the pseudo image to obtain correlation features between the first image and the second image and between the first image and the pseudo image, and processing the correlation features and the image features of the first image to obtain a depth image, the depth image prediction model is deployed in a deep learning processing unit of a programmable logic system of an SOC architecture. According to the embodiment of the invention, the real-time performance of depth image acquisition can be improved, the computing power demand is reduced, and the system energy consumption is reduced.

The present invention discloses a non-line-of-sight positioning error correction method and device. The present invention first uses the distance measurement value between a mobile node and multiple base stations to make a difference with the predicted value to obtain a residual; then assigns a weight to each distance measurement value according to the residual, and then solves the positioning equation group by weighted least squares method to obtain the coordinates of the mobile node; finally, Kalman filtering is performed to obtain the filtered mobile node coordinates and the next round of mobile node coordinate prediction values. The method only uses the distance measurement result, does not require the receiver to record the waveform of the signal, can avoid the problem that the signal characteristics change with the environment, has the characteristics of low calculation cost and low hardware requirements, and can be applied to indoor wireless positioning and other occasions.

The present application discloses a turning control method for a D-type robot based on an obstacle contour, the turning control method for the D-type robot comprising: step S1, the D-type robot uses a line laser sensor to extract the straight contour of the obstacle in front, and then according to the straight contour, the D-type robot calculates the rotation angle required for the side of the head of the robot to turn to be parallel to the straight contour; step S2, based on the rotation angle and the body size parameters of the D-type robot, the D-type robot calculates the head margin required to be adjusted before turning; wherein the head margin is expressed as the difference between the straight-line distance from the rotation center to the straight contour and the straight-line distance from the rotation center to the front end of the head in the vertical direction from the rotation center to the straight contour required for the D-type robot to turn; step S3, based on the head margin, the D-type robot adjusts its position, and then turns according to the rotation angle until the side of the head of the D-type robot is parallel to the straight contour.

This invention provides a method and system for intelligent prediction of hypersonic wall heat flux based on point-to-point modeling, relating to the field of high-speed aircraft design technology. The method includes: acquiring hypersonic flow field information to be predicted and generating grid node coordinates and corresponding unsigned distance functions covering the entire flow field computational domain; based on point-to-point modeling, treating each grid node as an independent sample point, inputting the coordinates and condition parameters of each node into the trained prediction model, and predicting the velocity component, Mach number, density, translational-rotational temperature, vibrational-electron temperature, and five-component mass fraction at each node; and calculating the wall heat flux based on the prediction data of the entire flow field wall and the outer layer of grid nodes. This invention achieves accurate reconstruction of a hypersonic cylindrical flow field considering thermochemical non-equilibrium effects under the condition of finite numerical simulation flow field data, thereby meeting the practical engineering needs for aerodynamic characteristic prediction.

The present application relates to a kind of improved polyphenyl ether production process state fast estimation method, including the linear high-dimensional system dynamic model generated by phenol and methanol alkylation polyphenyl ether;The state vector of high-dimensional system is divided into multiple state blocks of low dimension, while introducing an auxiliary variable to correct the system error caused by block.This method can meet the low calculation cost of linear high-dimensional system state estimation while achieving higher accuracy requirements.

The present invention discloses a grid adaptive turbulence simulation method based on Vreman dynamic coefficient coupled k-ε series model, and the main process of its implementation includes: step 1, judging whether to apply shielding function; step 2, calculating Vreman dynamic coefficient; step 3, determining local dynamic grid length scale; step 4, constructing scale-related adjustment function; step 5, reconstructing turbulent viscosity of k-ε series model; step 6, using reconstructed turbulent viscosity to simulate turbulence. The present invention constructs Vreman dynamic coefficient based on local velocity gradient tensor, identifies local grid size, and determines local grid length scale Δ * , and then reconstruct the turbulent viscosity by constructing scale-related functions through turbulent energy spectrum integration to achieve grid adaptive simulation, effectively overcoming the problem of high empirical dependence of existing turbulent mixing models on grids, further broadening the grid adaptability, effectively improving the calculation accuracy, greatly reducing the calculation cost, and significantly accelerating the turbulence simulation process.

This invention discloses a multimodal image feature matching method combining global affine transformation and local deformable convolution, comprising the following steps: constructing a multimodal target detection network architecture to obtain original visible light feature maps and original infrared feature maps; using a global affine transformation module to globally regulate large-scale rigid misalignment between images to obtain a coarsely aligned visible light feature map; extracting modality-invariant features to provide matching references for downstream networks, and performing deformable convolution on the two modality-invariant features; combining feature modulation masks to obtain the final finely aligned visible light feature map; performing feature fusion to output the final detection result. This invention, through global affine transformation, a priori corrects rigid geometric distortion between images, compressing spatial errors to the effective capture range of the fine alignment module, thereby ensuring the convergence and robustness of subsequent fine alignment and resolving the contradiction between large-scale spatial mismatch and limited local receptive field.

The present invention discloses a data anomaly detection method based on a distillation mechanism, comprising the following steps: 1: collecting and preprocessing original data; 2: compressing and encoding the preprocessed data, and projecting the input data into a latent space by an encoder to obtain an initial latent vector; 3: inputting the initial latent vector into an aggregation module in the latent space to obtain a landmark vector and a weight vector of each initial latent vector, and obtaining a constrained latent vector corresponding to the initial latent vector one by one through the landmark vector and the weight vector; 4: performing distillation and filtering on the constrained latent vector to obtain a clean constrained latent vector containing only normal mode features; 5: decoding and reconstructing the clean constrained latent vector to obtain reconstructed data corresponding to the input data; 6: calculating an anomaly score using the input data and the corresponding reconstructed data, setting an anomaly threshold, and a sample whose anomaly score exceeds the anomaly threshold is an abnormal sample; the present invention can adaptively update normal features during training, and has high accuracy in anomaly detection.

The present invention discloses a topology-preserving deformation method for a three-dimensional model based on a multi-body harmonic field, and belongs to the fields of computer graphics, computational mathematics, topology, and differential geometry. First, a tetrahedral mesh is constructed between the original three-dimensional model and the target three-dimensional model, and then the region where the topological transformation occurs during the deformation process is calculated based on the traditional one-body harmonic field, and a special multi-boundary condition is set, and then the multi-body harmonic field is calculated based on this, and finally a topology-preserving surface deformation is induced. The present invention can find the region where the topological transformation occurs during the deformation process based on the saddle point in the traditional body harmonic field, and adaptively construct a multi-body harmonic field that can induce topology-preserving deformation, thereby generating a topology-preserving deformation surface under the guidance of the multi-body harmonic field. The method is universal and efficient for the deformation of three-dimensional models with the same topology, requires less computational cost than the traditional large deformation differential homeomorphism metric mapping, and can be widely used.