11results about How to "Data augmentation" patented technology

The invention provides a blasting vibration peak velocity prediction method fusing parameter uncertainty and data-driven optimization, which comprises the following steps of: firstly, acquiring data such as blasting parameters, lithologic indexes, geological conditions and actually measured vibration peak velocity (PPV), and establishing a basic database; a prior probability model is constructed, a joint uncertainty model is established in combination with probability disturbance and a fuzzy triangular number, a multi-dimensional disturbance sample is generated through a joint central value and a joint standard deviation and is fused with original data, and an extended database is formed. Feature analysis is carried out on the fused data, and input variables which have obvious influence on the vibration peak velocity (PPV) are screened; and constructing a BP neural network model based on the screened features, carrying out global optimization on a network weight and a threshold by adopting a PSO algorithm, and then carrying out local fine tuning by utilizing Adam. Finally, through training and verification, prediction of the vibration peak velocity (PPV) is realized, and model precision is evaluated through RMSE, MAE, MAPE, Rand other indexes. The influence of rock and soil parameter uncertainty on prediction precision can be effectively processed, and the reliability and applicability of blasting vibration prediction are improved.

Embodiments of the present application provide a method, device, medium and electronic equipment for processing lead data, the method comprising: storing process data corresponding to a first type of lead in a temporary partition, wherein the first type of lead belongs to an uncompleted lead, and the uncompleted lead is a lead that needs business tracking; monitoring each lead stored in the temporary partition, and deleting process data corresponding to a second type of lead from the temporary partition and storing the process data in a fixed partition, wherein the second type of lead is a completed lead, and the completed lead is a lead that does not need continued business tracking. Through the embodiments of the present application, the data processing amount can be effectively reduced by monitoring the leads in the temporary partition, thereby saving the computing time and computing resources wasted by the join association operation of massive data.

The application discloses a user cold start recommendation system and method based on similarity matching, constructs a knowledge graph database according to users, products and the interaction relationship between the users and the products, and carries out standardization processing. The processed data is extracted to score information and mapped to a user like-dislike type space, so that the like-dislike product type of each user is determined, and a similarity matching algorithm is designed to find the old user with the highest similarity to the new user. The old user product type score vector and the new user product type vector score are processed by an association function and used as the recommendation basis of the system, the reasonable expansion of the new user data is completed, and the cold start problem of the recommendation system is effectively alleviated. The application solves the user cold start problem caused by the introduction of a new user into an industrial recommendation system. The application can be used for the scene of effectively recommending the product types that meet the preferences of new users in the industrial recommendation system.

This invention belongs to the field of high-precision position detection technology, and particularly relates to an absolute grating position calculation method based on an adaptive error-correcting sliding window. A photodetector array, in conjunction with an error-correcting sliding window, collects photoelectric signals of symbols, converting them into a symbol amplitude sequence. Abnormal erased symbols are identified through symbol confidence levels. If erased symbols are found, the window length is gradually expanded, and the process is repeated iteratively until a qualified target error-correcting sliding window is obtained. After removing erased symbols within the target error-correcting sliding window, grating position calculation is performed. After detecting a single acquisition position, the process is repeated for subsequent acquisition positions until the task is completed. This invention relies on adaptive iteration of the window to achieve signal error correction, reducing the interference of bad symbols on the calculation results and effectively improving the overall performance of grating position detection.

The utility model provides a LAN data centralized monitoring and calculating system relates to LAN dispatching monitoring and management field. Including data acquisition module selects node from LAN and gathers electric power instantaneous quantity data, data processing module carries out screening, analysis and operation to the instantaneous quantity of gathering, including: through the calculation of total output of self -generation of reverse method, according to the total output of self -generation with the main force self -generation plant output data that has gathered, calculates the secondary self -generation plant output, to chain type line structure, through two end substation data calculation line inner power generation and load balance value as approximate power output, to tree shape line structure, through main line data subtracts public power plant and public load data and obtains self -generation output, centralized display module shows instantaneous quantity data after data processing module processing in same monitoring interface, has promoted data acquisition efficiency, has strengthened data timeliness and accuracy, has optimized dispatching decision, has improved power grid operation reliability.

Provided in the present disclosure are a flash memory array, and a write method and erasure method therefor. The flash memory array according to the present disclosure comprises : a plurality of flash memory cells, which are arranged in a row direction and a column direction perpendicular to the row direction; a plurality of word line groups, which extend in the row direction; and a plurality of bit line groups, which extend in the column direction, wherein flash memory cell pairs are arranged at intersection points of the word line groups and the bit line groups, and the flash memory cell pair comprises a first flash memory cell and a second flash memory cell, which are adjacent in the row direction and share the same bit line group. By means of the flash memory array according to the present disclosure, the arrangement density of bit lines can be improved while the size of the array is not increased, and the bit line parasitic resistance can be reduced. In addition, compared with a flash memory array in the prior art, the flash memory array according to the present disclosure also has a better process compatibility and scaling characteristic. The write method for a flash memory array of the present disclosure has the advantages of a low operation power consumption and a fast programming speed, which is conducive to improving the number of concurrently written flash memory cells, thereby increasing the data writing throughput rate of a memory. The erasure method for a flash memory array of the present disclosure can improve a threshold voltage window and enhance the storage reliability, and also has the advantages of a low operation power consumption and a fast erasure speed.

The present application provides a kind of training method, device, electronic equipment and storage medium of pedestrian search model based on twin network, wherein the method comprises: determining training image;According to the lower branch of twin network, the feature of training image is extracted, and the feature vector of each pedestrian in training image is obtained;The training image is processed to obtain multi-scale pedestrian image, and the feature of multi-scale pedestrian image is extracted according to the upper branch of twin network, and the multi-scale feature vector of each pedestrian in training image is obtained;According to the feature vector and multi-scale feature vector, training is carried out, and the trained pedestrian search model is obtained when training is completed. When pedestrian search is carried out, the pedestrian at different scales is accurately searched.

The application relates to a water conservancy project embankment hidden danger precise detection and reinforcement device and relates to the field of embankment hidden danger detection devices. The device comprises a moving mechanism, the upper end surface of the moving mechanism is fixedly provided with a rotating seat, the rotating seat is rotationally connected with the moving mechanism, a rotating oil cylinder for driving the rotating seat to rotate is arranged on the moving mechanism, a positioning vehicle frame is fixedly arranged on the rotating seat, an equipment box and a water storage tank are arranged on the upper end surface of the positioning vehicle frame, the equipment box and the water storage tank are fixedly connected with the positioning vehicle frame, and a detection box is further fixedly arranged at the front end of the positioning vehicle frame. The device comprehensively uses geological radar and a multispectral imager to detect the embankment from the aspects of internal structure and surface features, can comprehensively and accurately find various hidden dangers of the embankment, and greatly improves the detection precision. Meanwhile, the moving mechanism is designed in a track type.

The invention relates to the technical field of intelligent medical treatment, and discloses a comprehensive obesity evaluation model based on user health data and application thereof, and the model comprises a data acquisition module, a data preprocessing module, a feature engineering module, a model training module and an evaluation output module. According to the method, multi-source heterogeneous health data are integrated, the limitation of traditional single data assessment is broken through, multi-dimensional factors influencing obesity can be comprehensively captured, data preprocessing and feature engineering optimization are combined, the data and feature quality is effectively improved, and a foundation is laid for accurate assessment; secondly, by adopting a deep learning and causal reasoning hybrid architecture, not only is feature extraction and evaluation precision guaranteed, but also the function logic of the features on obesity evaluation is clarified through interpretability analysis, and the evaluation credibility is improved; besides, the dynamic obesity risk score updated along with the real-time health data of the user is generated, the risk change can be reflected in real time, and accurate and efficient technical support is provided for personalized health management and obesity intervention.

The invention discloses an unmanned aerial vehicle-based precise control method for diseases and pests of litchi, longan and garden plants, and belongs to the technical field of agricultural plant protection. The method comprises the following steps: collecting a multispectral image and geographic information of an orchard through an unmanned aerial vehicle; analyzing the image by using a deep learning algorithm and an image recognition technology, recognizing a disease and pest region and severity, and generating a disease and pest prescription map containing pesticide application parameters by combining a spectrum database for comparison; and the unmanned aerial vehicle flies to the target area according to a path planned by the prescription map, controls the on-off and flow of the nozzle according to the prescription map parameters and the real-time position, and carries out precise pesticide spraying on the target area. The problems of pesticide abuse, low efficiency and uneven coverage in traditional plant plantation pest control are solved, early recognition, accurate diagnosis and variable pesticide application of plant canopy pests in a garden, especially lychee and longan fruit trees, are achieved, and therefore the pesticide utilization rate, the operation efficiency and the control effect are remarkably improved, and the economic benefit is increased. And the labor cost and the environmental risk are greatly reduced.

This invention discloses a method and system for four-way linkage management of engineering costs based on a hybrid large model, belonging to the interdisciplinary field of engineering digitization and artificial intelligence. The method uses a multi-constraint fusion scientific computing hybrid large model as its technical foundation, constructing the model through a four-in-one mechanism of data pre-setting, boundary setting, threshold pre-setting, and dynamic data import. It adopts a hybrid architecture of decision trees and recurrent neural networks, integrating five types of constraints: construction period, resources, cost, changes, and engineering specifications. Based on this, it achieves three-dimensional synchronization, three-value quantification, three-calculation comparison, and intelligent risk control linkage, with the bill of quantities item as the smallest accounting unit. When the cost deviation exceeds the threshold, an early warning is automatically triggered, and the root cause is traced back along the data correlation link, realizing a closed-loop management system of "early warning, location, and processing." This invention solves the pain points of traditional methods, such as inaccurate calculations, delayed early warnings, and untraceable data, significantly improving the accuracy, efficiency, and reliability of cost management. This technical solution has been evaluated and certified by the Ministry of Housing and Urban-Rural Development (Jiankeping [2023] No. 066).