51results about How to "Reduce collection costs" patented technology

The invention provides a satellite solar short-wave radiation inversion method based on knowledge distillation, which relates to the field of solar short-wave radiation inversion, and comprises the following steps: collecting polar orbit satellite data, stationary satellite data, geographic information data and ground observation data; preprocessing the data to obtain input features and labels for training the teacher model and the student model; sorting the importance of the input features by using a random forest model, selecting key features, and removing non-key features; constructing a deep learning model by using the key features, and generating a teacher model of solar short-wave radiation inversion; and taking an output result of the teacher model as a soft label, introducing ground observation data as a hard label, and carrying out knowledge distillation to obtain a student model of solar short-wave radiation inversion. According to the technical scheme of the invention, the problems of data scarcity and high data acquisition cost caused by the fact that a model is constructed only depending on ground observation data in the prior art are solved, and the problem that the inversion precision completely based on satellite data is limited is solved.

The invention belongs to the field of big data processing, and discloses a crowd travel mode identification method and system for multi-view subspace learning based on dynamic and static feature fusion, and the method comprises the steps: carrying out the preprocessing of original signaling data, and cleaning invalid data; extracting a travel track of each user, and processing possible abnormal conditions of ping-pong data and drift data; after track data of a single user in a sample time period is obtained, segmenting the track data, and extracting a single trip chain track; extracting static characteristics such as travel behaviors and travel traffic of a travel chain in combination with multi-source data such as mobile phone signaling data, an OSM road network and public traffic data; extracting time sequence features by using an LSTM model, and modeling dynamic features of a travel track; static and dynamic features are fused, and travel modes are captured from multiple views. According to the method, the complex travel mode can be more accurately captured, data redundancy and noise are reduced, and the accuracy of travel mode identification is remarkably improved.

This invention provides a method for modeling a cylinder motion system based on a physical information neural network, comprising: acquiring time-series data of the cylinder motion system, including time-series control quantities and system state quantities; constructing a physical information neural network model with a dynamic feature input layer and a physical parameter input layer, receiving the time-series control quantities and physical parameter vectors respectively; performing a nonlinear transformation on the physical parameter vectors through a physical parameter encoding network to obtain encoded physical feature vectors; fusing the time-series control quantities and physical feature vectors to form a joint input vector; inputting the joint input vector into the backbone neural network, synchronously predicting the system state quantities through multi-layer nonlinear transformations; constructing a composite loss function including a data loss term and a physical loss term; training the model using time-series data, updating the model parameters by minimizing the composite loss function until the model converges, thereby achieving high-precision modeling using only single motion data.

The application provides a vehicle operation data collection method and system, computer equipment and medium, and belongs to the field of vehicle data processing. The method comprises the following steps: obtaining vehicle model information and user demand of a vehicle to be collected; determining data analysis service project content options based on a data analysis service project content information table and vehicle model information corresponding information table, and screening data analysis service project content according to the user demand of the vehicle to be collected; generating a natural language text for the data analysis service project content to obtain a vehicle data collection strategy; collecting environmental data and judging based on data collection conditions; when the environmental data falls within the range of the data collection conditions, collecting operation data of the vehicle to be collected according to the data collection project, and obtaining operation data of the vehicle to be collected. The method can dynamically collect data based on the actual operation state, driving environment and user demand of the vehicle, and ensure the effectiveness of operation data collection.

The invention belongs to the technical field of new energy power generation and grid connection, discloses a grid-side converter fault ride-through control structure modeling method, system and equipment and a medium, and aims to solve the problems of difficult modeling, high cost, poor extrapolation and lack of universality caused by dependence on manufacturer internal logic or massive experimental data in the prior art. According to the method, disturbance input and current output data under a limited working condition are acquired, a dimension raising mapping function is constructed to map the input to a high-dimensional feature space, a Koopman operator is solved based on ridge regression to establish a global linear mapping model, active and reactive current reference components are predicted, and the current reference components are optimized. And finally, inputting the current inner loop time domain model to generate a current transient response curve during the fault period. According to the method, high-precision modeling and prediction can be realized only by a small number of samples, and the method has good extrapolation and universality and can be adapted to various fault ride-through strategies.

The invention relates to the technical field of image recognition, and discloses a wind power fan blade defect recognition method and system based on image recognition, and the method comprises the steps: carrying out the blocking cutting, illumination normalization and image enhancement processing of an original image, constructing a defect-free sample and a defect sample, and dividing the samples into a training set and a test set; the method comprises the following steps: taking a GANopen network as a basic framework, fusing the lightweight design of Mamba-YOLO, constructing a joint loss function by adversarial loss, reconstruction loss and coding loss based on a training set, carrying out unsupervised training, optimizing network parameters until convergence, and obtaining a defect identification model; inputting the preprocessed to-be-detected fan blade image into the trained defect recognition model, performing defect recognition and positioning, and outputting the defect type and position; based on an output result of the defect identification model, dynamically adjusting an early warning level and response measures through a self-adaptive early warning mechanism; according to the invention, the efficiency and accuracy of wind power fan blade defect identification are improved.

This application discloses a method for training a recognition model, a recognition method, a computer device, and a computer-readable storage medium. The training method includes the following steps: controlling a single-photon lidar to emit photon pulses towards a target scene and receiving one-dimensional single-photon echoes reflected from the target scene, where the target scene includes at least one object to be recognized; constructing a training sample set based on the one-dimensional single-photon echoes; constructing an initial model and inputting the training sample set into the initial model for iterative training; validating the initial model after each iteration, marking the validated initial model as the recognition model, and outputting it. The application method involves inputting the echo signal to be recognized into the trained recognition model, and obtaining the recognition result based on the processing of the recognition model. The recognition result includes the pose and type of the object in the target scene. Therefore, this application can achieve high-precision recognition of long-distance targets.

A method for evaluating and sampling the value of a dataset is disclosed, which aims to reasonably assess the value of the dataset and guide the sampling of high-value datasets. The method first establishes an evaluation model for assessing the value of individual data points and a function to describe the degree of value redundancy among data points. Then, it constructs a value evaluation model for the dataset by comprehensively considering the value of individual data points and the degree of value redundancy among them. Based on this data set value evaluation model, high-value datasets can be sampled from the data sampling space according to user needs. The main application of this invention is to evaluate the value of datasets and sample high-value datasets. It can guide data sampling in data-driven analysis, modeling, and decision-making tasks, thereby improving the quality of datasets and effectively reducing data acquisition costs while ensuring the effectiveness of the target task.

The embodiment of the present invention discloses a method, device and electronic device for acquiring speech training data and training a model, including: inputting speech data in a speech data storage device to a target device in the form of an electrical signal; processing the speech data through a channel; and acquiring the through-channel speech data after the through-channel processing. The embodiment of the present invention can reduce the acquisition cost of speech training data, improve the quality and versatility of speech training data, and thus improve the training efficiency and recognition effect of a speech recognition model.

The invention discloses a method for detecting abnormal electricity consumption of an intelligent electric meter, which relates to the technical field of intelligent meter reading, and comprises the following steps of: 1, data acquisition and preprocessing, 2, data compression and dimension reduction, 3, lightweight data aggregation, 4, transfer learning and fine tuning, 5, self-supervised learning, and 6, model lightweight. According to the method, a lightweight aggregation algorithm is used, that is, data aggregation is carried out through a local weighted average method instead of a complex distributed data aggregation model, so that the calculation complexity is reduced, the aggregation efficiency is improved, and the dependence on hardware resources is reduced; before data aggregation, a data compression technology is used to carry out dimension reduction on power consumption data, so that the data volume is reduced, the aggregation efficiency is improved, and the information loss is reduced; transfer learning and fine tuning are adopted, a BERT pre-training model is utilized to accelerate training of the deep belief network, training data requirements are reduced, and model training efficiency is improved.

The application provides a microseismic real-time monitoring method and device based on compressed sensing and 5G node collection, solves the problem that microseismic monitoring data is easy to be submerged in noise interference, and microseismic data collection equipment cannot transmit data in real time, or the transmitted data is uneven and incomplete. It comprises the following steps: collecting data, establishing a geological and geophysical model, simulating fracturing based on the parameters of the well to be fractured to demonstrate the parameters of the microseismic monitoring observation system; establishing an adaptive optimal learning dictionary to form a compressed sensing collection scheme; conducting a due diligence investigation of field obstacles and interference sources to form a compressed sensing active obstacle avoidance scheme; using a 5G node collection system to collect field data; applying compressed sensing data reconstruction technology to reconstruct the collected field data to form three-dimensional microseismic monitoring data; carrying out microseismic data processing and interpretation, and feeding back the microseismic analysis results to the fracturing site to realize real-time monitoring of the fracturing construction process.

The present invention is applicable to the field of photoelectric detection technology, and provides a multi-channel lighting method and device based on optical fiber. The opening and closing states of shutters connected to optical fiber channels are determined according to the number of optical fiber channels undergoing lighting tests, and the collection interval of photon collection is determined according to the opening and closing states. According to the collection interval, the channel photon value collected within a preset lighting period is obtained. According to the collection interval and the channel photon value, the actual photon value collected within the preset lighting period is obtained, thereby realizing dynamic collection of multiple channels. Moreover, through the shutter structure, the interference when different channels are illuminated at the same time is reduced, the stability of collection and the accuracy of the results are improved, and the collection cost is reduced.

The present invention belongs to the technical field of mine transportation, and discloses a belt conveyor roller abnormality detection system based on optical fiber auscultation, comprising: a data acquisition unit, a preprocessing unit and a prediction model connected in sequence, and an abnormality detection unit and a linear layer connected to the prediction model. The data acquisition unit collects audio data during the operation of the roller, converts it into vibration data, and outputs it to the preprocessing unit. The preprocessing unit segments the input vibration data, and inputs the segmented vibration data into a trained prediction model. The prediction model outputs the features of the latent space corresponding to the current vibration data to the abnormality detection unit, and calculates the distance between the features of the current latent space and the feature center of the latent space when the roller is operating normally. If the distance is greater than the set distance threshold, the corresponding segment of the roller is currently operating abnormally. Not only is the automatic detection of roller abnormalities realized, but the latent space features output by the prediction model contain richer information, thereby improving the accuracy of roller abnormality detection.

The application relates to a power distribution facility anomaly detection method based on deep learning, and relates to the technical field of power distribution facility anomaly detection. The application collects electric characteristic signals of a power distribution facility at a set sampling rate, selects electric characteristic signals of a non-anomaly time period located on a slow spectrum submanifold as training data, minimizes a loss function, uses a full connection network to learn a mapping matrix from a nonlinear feature vector of the electric characteristic signals to a latent variable on a low-dimensional slow spectrum submanifold of the electric characteristic signals and a linear dynamics matrix in a latent space, in application, collects electric characteristic signals, uses the last corresponding latent variable of the electric characteristic signals and a predicted future latent variable obtained by iteration using the linear dynamics matrix, reconstructs a predicted non-anomaly electric characteristic signal through inverse transformation between the latent variable and the electric characteristic signal, and performs anomaly analysis by using the difference between the predicted non-anomaly electric characteristic signal and an actual electric characteristic signal.

The application discloses a kind of automobile body parts resistance spot welding quality evaluation method and system.The method comprises: using the existing equipment of automobile production workshop and process parameter to carry out welding experiment to sheet part same as the material of key position of automobile body, open data record function in the host computer of welding equipment, obtain the data file of different quality types of welding spot;According to the result of spot welding, the data file is classified, and the voltage and current data corresponding to each quality type of welding spot are obtained by analyzing all data files, and the dynamic resistance data is calculated;Lightweight one-dimensional convolutional neural network model is built as spot welding quality evaluation model;The test data is input into the spot welding quality evaluation model after training and optimization, and the corresponding spot welding quality evaluation result is obtained.The method realizes the automatic extraction of the deep characterization features reflecting the spot welding quality result in dynamic resistance data, avoids the complex process of data feature extraction and analysis, and obtains higher detection accuracy.

An online monitoring method for the grinding state based on the body information of an industrial robot, which collects the robot joint current during the pre-grinding experiment and the surface quality texture after the grinding is completed; then the joint current time series signal is converted into an image, thereby generating a grinding state label and a grinding state classification model data set, and the data set is used to train the grinding state classification model; finally, the trained grinding state classification model is used in the online stage to realize the closed-loop feedback control of the robot. The present invention uses a convolutional neural network to capture the inherent time-frequency domain characteristics of the robot joint current, establishes a feature joint selection model based on the principal component analysis of the joint current, and realizes the online monitoring of the robot grinding state.

The present application relates to the technical field of silicone key conductivity test, and discloses a silicone key conductivity reinforcement learning test system and method, wherein the silicone key conductivity reinforcement learning test method comprises the following steps: constructing a silicone key conductivity test environment; establishing a test knowledge base, extracting general feature representation of key conductivity performance through feature mapping and domain adaptation algorithm; constructing a test parameter optimization model, modeling the test process as a Markov decision process, and learning the optimal test strategy based on a reward function; adjusting the test parameters in real time during the test process, performing reliability analysis after obtaining the test results, and feeding back the new test experience to the knowledge base to realize the accumulation and reuse of test knowledge; the present application realizes the self-adaptive optimization of test parameters and the cross-key knowledge transfer by fusing reinforcement learning, transfer learning and meta-learning technology.

The invention provides a standing tree cutting degree parameter automatic inversion method based on a ground-based laser radar and deep learning, and relates to the technical field of forest tree measurement. The method comprises the following steps: acquiring an original three-dimensional laser point cloud based on multi-station scanning, and acquiring a clean trunk point cloud; and generating a high-density diameter-height observation data set through axial skeletonization extraction and slicing processing in combination with circle fitting calculation. A variable parameter cutting degree equation model introducing a relative tree height ratio and a height-diameter ratio factor is constructed, and a residual sum of squares objective function is established; and iteratively solving the optimal shape parameter vector by using a nonlinear least square algorithm, and outputting a model parameter of the target stumpage when a precision evaluation index is met. According to the method, the problem of measurement errors caused by branch and leaf shielding in a complex under-forest environment is effectively solved, a traditional destructive survey mode of cutting down and analyzing trees is replaced, non-destructive, high-precision and automatic inversion of standing tree cutting degree parameters is achieved, and the efficiency and the digital level of forest resource monitoring are improved.

The application provides a self-supervised speech denoising method and device, and relates to the field of acoustic signal processing, and comprises the following steps: S1, obtaining a DCUnet network, constructing a speech denoising model by adding a TCM module in the DCUnet network; S2, obtaining a noise speech set, training the speech denoising model through the noise speech set, and obtaining a trained speech denoising model; S3, denoising speech through the trained speech denoising model. The application constructs a speech denoising model based on a deep complex domain DCUnet network, combines a complex domain overall processing strategy, improves the reconstruction quality and denoising performance of the speech signal, dynamically captures the change characteristics of the noise scene through the TCM module in the speech denoising model, enhances the adaptability of the model to dynamic signals in a complex environment, trains the speech denoising model by using an ONT strategy, uses noise speech as training data, does not need clear target speech data, reduces the training data collection cost, and improves the generalization ability of the model.

This invention belongs to the field of intelligent seismic exploration technology, specifically relating to an intelligent reconstruction method for passive source seismic wavefields under sparse active source constraints and non-ideal illumination, to address the problems of false phase axes and coherent noise under non-ideal distribution of underground passive sources. To improve the reconstruction effect of passive source seismic wavefields under non-ideal illumination conditions, this method constructs an improved U-Net network, enabling it to intelligently extract data features and derive attention mechanisms along two independent dimensions—channel and space—within the network. The attention mechanism is then multiplied by the input feature map for adaptive feature refinement, achieving intelligent reconstruction of passive source data. Furthermore, a deep learning network, trained through learning, constrains the passive source seismic data prediction network with sparse active source seismic records, enabling the reconstruction of passive source data with co-located active source lines under sparse active source conditions, thus improving the multi-dimensional deconvolution reconstruction effect and computational stability.

The invention discloses a traffic signal control method based on machine visual perception and space-time sequence prediction, and relates to the technical field of intelligent traffic signal control. The traffic signal control method based on machine vision perception and time-space sequence prediction comprises the following steps: S1, machine vision perception: acquiring real-time image data of a traffic intersection through image acquisition equipment, preprocessing the image data, and extracting traffic flow key parameters; and S2, space-time sequence prediction: constructing a PCA-LSTM traffic flow prediction model based on principal component analysis PCA and a long short-term memory neural network LSTM, inputting the preprocessed traffic flow historical data, and outputting a short-term traffic flow prediction result. The traffic signal control method based on machine visual perception and time-space sequence prediction is comprehensive in data coverage and low in acquisition cost; and the prediction performance is better: through abnormal data processing and dynamic principal component selection of the PCA-LSTM model, the prediction precision is improved by 2%-5% compared with the traditional LSTM, and the calculation efficiency is improved by 15.01%.

The application relates to a bearing damage positioning method based on Lamb wave dispersion mode matching of acoustic emission signals, wherein different frequency band components of original acoustic emission signals with the same frequency interval are obtained; a time of arrival curve is obtained; a theoretical dispersion curve of Lamb wave propagation in a bearing is solved, then a coordinate system is unified with the signal time of arrival curve; the dispersion curve and the time curve are matched through a Pearson correlation coefficient index; a formula is established according to a speed difference and a time difference, and a bearing damage distance is calculated; the sensor is moved by a certain distance to be measured again, and the damage position is determined through the two results. The method of matching the Lamb wave propagation mode can determine the propagation speed of the acoustic emission signals with different frequencies, the accuracy of the time difference positioning method is greatly improved, the bearing damage position can be directly found through a single sensor, the bearing fault diagnosis can be realized without depending on the bearing rotating speed, the installation difficulty can be effectively reduced, and the detection cost can be reduced.

The utility model discloses a construction waste recycling and classifying device, which belongs to the field of construction waste recycling and comprises an accommodating box, a control cylinder, a control panel, an electromagnet, a power mechanism and a control mechanism, the control cylinder is rotatably mounted in the accommodating box, the rotation of the control cylinder is controlled by the power mechanism, the electromagnet is arranged on the outer wall of the accommodating box, and the control panel is arranged on the control panel. The multiple electromagnets are evenly distributed along the outer wall of the containing box, the control mechanism comprises a mounting base, a control shaft and a connecting plate, the mounting base is slidably mounted in an inner cavity of the control cylinder, one end of the connecting plate is rotationally connected with the outer wall of the mounting base, and the other end of the connecting plate is rotationally connected with the outer wall of the control plate; the control shaft penetrates through the mounting seat and is connected with the mounting seat through internal and external thread fit; the power mechanism comprises a bearing seat, a driving motor, a driving wheel and a driven wheel; through cooperative use of all the devices, the collecting efficiency can be remarkably improved, and then the collecting cost is reduced.

The invention relates to the field of computer geographic data processing and GIS (geographic information system) data automatic acquisition, and discloses a method for automatically generating a high-precision map according to a vehicle trajectory, which comprises the following steps of: acquiring road data by using a vehicle carrying Beidou terminal equipment, cleaning the acquired road data, and storing the cleaned data into a database; and then reading the data from the database, carrying out parallel copying on the data and generating a buffer area to obtain vector data of the traffic facilities, and adjusting and rendering different data types by utilizing a map rendering tool to obtain a high-precision map. Compared with an existing method, on the premise that the data accuracy and integrity are guaranteed, the number of hardware equipment used for road data collection is reduced, the collection cost is reduced, centimeter-level real-time dynamic motion positioning Beidou terminal equipment is utilized, the data collection error is reduced, and the data collection efficiency is improved. And the influence of factors such as weather, terrains, shelters and the like on the generation of the high-precision map is avoided to a certain extent.

The invention provides a method and device for collecting substances on the surface of a wafer, and the method comprises the steps: enabling the wafer to be inverted, and enabling the surface of the wafer to face downwards; a scanning device is arranged below the wafer, the top end of the scanning device carries scanning liquid, and the surface of the wafer is in contact with the scanning liquid; the wafer and the scanning device move relatively, and the scanning device scans at least part of the surface of the wafer, so that the scanning liquid dissolves substances on the surface of the wafer and collects the substances into the scanning liquid. According to the method, the step of atomizing or vaporizing the HF etching wafer surface is omitted, and the collection difficulty and cost are reduced; the wafer is inverted, so that particle pollution caused by the fact that particles on a fan filter unit (FFU) fall onto the surface of the wafer is avoided; meanwhile, the method is not only suitable for the hydrophilic wafer surface, but also suitable for the hydrophobic wafer surface, and has the advantages of being wide in application range and high in scanning and collecting accuracy.

The invention discloses a data expansion and exoskeleton joint end-to-end torque estimation method based on a diffusion model, and the method comprises the steps: carrying out the normalization processing of time series data from a multi-source sensor, carrying out the fragmentation according to a fixed length, and constructing a training sample with a motion class label; then, training a classifier-free conditional diffusion model by adopting a sample, and simultaneously learning conditional and unconditional denoising mapping relationships in a manner of randomly inactivating category conditions; in the generation stage, based on a classifier-free condition guidance mechanism, multi-modal time series data with specified motion category features are gradually generated from random noise; and finally, fusing the generated synthetic data with real acquired data to train a joint torque end-to-end prediction network, thereby realizing joint torque estimation of input sensor time sequence data. The method improves the prediction precision, generalization ability and stability of the end-to-end torque estimation model in a multi-action and few-sample scene, and has a good engineering application value.

This invention discloses a method, system, storage medium, and vehicle for collecting automotive fault data, relating to the field of automotive fault data collection. The method includes the following steps: providing a data reading interface for the device to be diagnosed to an onboard control device with data transmission capabilities; collecting diagnostic data from the device to be diagnosed corresponding to the diagnostic task through the onboard control device; and storing the diagnostic data as automotive fault data. This invention directly utilizes existing onboard control devices with data transmission capabilities to collect the operating information of the device to be diagnosed; it eliminates the need for external diagnostic instruments as in existing technologies, thereby improving collection efficiency and reducing collection costs. Simultaneously, the automotive fault data collected by this invention is automatically stored, eliminating the need for manual scanning and uploading operations as in existing technologies, thus improving the user experience.

The invention relates to the technical field of target recognition, discloses an unmanned aerial vehicle recognition model training method and an unmanned aerial vehicle recognition method, and aims to solve the problems of high cost and low efficiency of an existing unmanned aerial vehicle recognition scheme, and the scheme mainly comprises the following steps: segmenting radar data of an unmanned aerial vehicle track to obtain a plurality of first unmanned aerial vehicle samples; correspondingly adding the feature data of each trace point on any two first unmanned aerial vehicle samples according to the same random weight, and then adding random perturbation to obtain a plurality of second unmanned aerial vehicle samples; segmenting the radar data of the non-unmanned aerial vehicle flight path to obtain a plurality of non-unmanned aerial vehicle samples; randomly selecting a second unmanned aerial vehicle sample as a third unmanned aerial vehicle sample; and constructing an identification sample set based on the first unmanned aerial vehicle sample, the third unmanned aerial vehicle sample and the non-unmanned aerial vehicle sample, and training an unmanned aerial vehicle identification model based on the identification sample set. The training efficiency and accuracy of the unmanned aerial vehicle identification model are improved, the cost is reduced, and the method is particularly suitable for radar detection.

The disclosed embodiments disclose an image generation method and device, an electronic device and a storage medium, the method comprising: extracting content features of a first image; respectively extracting full-image style features of a second image and object style features of a local image block including an object in the second image; determining a target style feature based on at least the full-image style feature and the object style feature; and generating a third image based on the content features and the target style feature.

The invention discloses a copper-bismuth composite electrode material with a nanowire array structure, which comprises a foamy copper substrate and a copper-bismuth nanowire array layer deposited with a bismuth component, and the copper-bismuth nanowire array layer deposited with the bismuth component grows on the foamy copper substrate. The invention also discloses a preparation method of the material. The preparation method specifically comprises the following steps: step 1, preparing a sample; step 2, carrying out electroreduction treatment on the sample in an electrolyte to obtain a substrate loaded with a copper nanowire array; and step 3, preparing a bismuth-containing electro-deposition solution at room temperature, placing the copper nanowire array substrate in the step 2 in the bismuth-containing electro-deposition solution, loading a bismuth component on the copper nanowire array substrate in a constant-current electrochemical deposition mode, and after deposition is finished, performing immersion cleaning and drying to obtain the copper-bismuth composite electrode material with the nanowire array structure. According to the copper-bismuth composite electrode material with the nanowire array structure, provided by the invention, carbon dioxide can be reduced into formate with high selectivity, and the stability of an electrode is remarkably enhanced.