345results about How to "Flexible handling" patented technology

The embodiment of the invention discloses a speculation decoding method and device for a large language model and a medium, and relates to the technical field of large language models, and the method comprises the steps: obtaining the current task information of a target large language model; the method comprises the steps of inputting original text prompt information into a target large language model, determining initial token information, constructing a plurality of patch data of a specified patch level according to the initial token information and the original text prompt information, generating a plurality of candidate patch sequences in parallel according to each patch data and a draft model, mapping each candidate patch sequence, determining initial draft information, and outputting the initial draft information. The preliminary draft information comprises a token sequence corresponding to each candidate patch sequence; and performing parallel verification on the token sequence corresponding to each candidate patch sequence through the target large language model, screening the token sequences corresponding to the candidate patch sequences, determining the next round of input of the draft model, and performing speculative decoding.

The invention provides a lightweight underwater image enhancement method based on adaptive feature fusion, and the method comprises the steps: decomposing an underwater image enhancement task into a strong prior enhancement task and a fine-grained enhancement task through employing a lightweight underwater image enhancement network based on adaptive feature fusion, so as to make full use of the multi-scale information in an image; the strong prior enhancement task is used for extracting global and low-layer features of the image; the fine-grained enhancement task is used for improving local details and color recovery of the image so as to improve the visual quality; the lightweight underwater image enhancement network comprises a multi-scale feature aggregation module and a global channel attention module, the multi-scale feature aggregation module is composed of an adaptive feature fusion module and a segmented convolution feed-forward network, and the lightweight underwater image enhancement network is used for combining global and local information of features and adaptively adjusting importance of each channel through information association between channels.

The invention discloses an image recommendation method based on a graph neural network, and relates to the technical field of recommendation systems, and the method comprises the following steps: constructing model input, constructing a graph neural network model, comparing a learning module, carrying out joint training and optimization, and carrying out recommendation generation and evaluation. According to the method, the relationship embedding matrix in the knowledge graph is considered into the graph neural network model, the splicing vector of the knowledge graph is used as a construction expansion node embedding vector input model, and the complex relationship between the user and the image can be deeply mined through the graph neural network and comparative learning in combination with the interaction graph, so that the user preference is accurately captured, and the user experience is improved. Therefore, the recommendation accuracy is remarkably improved, the multi-dimensional features of the user and the image nodes are expanded, the context sensing technology is combined, the method can provide highly personalized recommendation services for the user, the unique requirements of the user are met, and the user experience is improved.

The invention relates to the technical field of communication, and discloses a communication data transmission and temporary storage method and device and a storage medium, and the method comprises the following steps: S1, receiving and partitioning initial data, and storing the initial data in a cache region; s2, calculating weighting coefficients of the data blocks and performing priority ranking; s3, encrypting the data blocks according to the sequence; s4, transmitting the encrypted data to a target end through a security protocol; s5, the target end decrypts the data and verifies the integrity; s6, temporarily storing the verified data blocks in a temporary storage area until all the data blocks are transmitted; and S7, after the transmission is completed, recombining and storing the data block in the target equipment. According to the method, the data is subjected to blocking, priority ranking, encryption transmission and decryption recombination, so that the security, integrity and transmission efficiency of the data in the network transmission process are ensured, and the method is suitable for a communication system needing high security and high data integrity.

The invention discloses a traumatic orthopedic nursing data intelligent analysis method, and relates to the field of medical data analysis. A traumatic orthopedic nursing data intelligent analysis system constructed by the method comprises a scheme management module, a data acquisition module, a data processing module, a nursing stage analysis module, a nursing data analysis module and a nursing scheme optimization module. According to the method, a complete trauma orthopedic nursing scheme library is constructed, comprehensive nursing schemes of different trauma types are covered, a rich data basis is provided for subsequent analysis, and the comprehensiveness and applicability of the nursing schemes are improved. Through the nursing stage analysis model, the wound type and the nursing stage of the patient can be corrected and confirmed, and the analysis accuracy is remarkably improved.

The invention discloses a ground target detection method based on an adaptive hierarchical gradient feature fusion strategy, and the method comprises the steps: obtaining a ground obstacle target data set, constructing a ground target detection network, introducing a C2f-EIRM module into a backbone network of the constructed ground target detection network, employing a hierarchical feature fusion strategy for the C2f-EIRM module, and carrying out the feature fusion of the C2f-EIRM module, introducing a plurality of branches to perform global feature extraction; in the multi-branch structure, an EIRM module is adopted to perform local feature extraction, gradient operator convolution is introduced into the EIRM module to construct an edge information module, the edge information module is combined with common convolution in a self-adaptive adjustment mechanism mode, the receptive field and edge information of the feature extraction network are enhanced, and the model detection performance is improved. According to the method, the target detection accuracy is improved, the efficient identification of the shielded small target is realized, and the real-time performance and the accuracy of ground target detection of the low-altitude flying car are balanced at the same time.

The invention discloses a vehicle-mounted message middleware processing method and system and an intelligent cabin thereof. The method comprises the steps that an abstract interface layer used for achieving data type abstraction and decoupling is defined, a communication interface is created based on a hardware abstract layer, and cross-device communication and cross-process communication are conducted through the abstract interface layer; based on the communication interface, a Fast DDS is adopted in a communication protocol transmission layer to serve as message middleware, an abstract interface layer is packaged in an application layer, communication nodes are created in a communication interface layer according to a communication protocol and a QoS strategy, and the communication nodes are used for managing objects of a communication mode in a cross-process and cross-device mode and in-process communication. According to the vehicle-mounted message middleware provided by the invention, a user interface is simplified, and two main communication modes are supported, so that a system can flexibly select a communication mode according to an application scene.

The invention relates to the technical field of freight scheduling, in particular to a freight scheduling information management method and system and a storage medium. The method comprises the following steps that uncompleted order collection is carried out on a freight platform to obtain an uncompleted order set, and the uncompleted order set comprises a plurality of uncompleted orders; each uncompleted order at least comprises an order ID, order release time, target cargo attribute data, publisher identity data, a departure place, a destination, an order state identifier, an order transportation task schedule, cargo stacking attributes, predicted delivery time and cargo expected arrival time; and performing order state classification and marking on the uncompleted order set to obtain a classified order set. By optimizing the loading space utilization rate of the vehicle, the unloaded and half-loaded conditions of the vehicle can be reduced, so that the fuel consumption and the transportation cost are reduced.

The invention discloses a data processing method and device based on a visual database and electronic equipment, and relates to the field of data processing. The method comprises the following steps: acquiring a data processing request of a user; analyzing the data processing request to obtain a corresponding key field and a logic conjunction; according to the key field, performing semantic matching in a preset data model library to obtain a corresponding target data table and a target field; synthesizing the logic conjunction, the target data table and the target field to generate a corresponding database processing statement; executing the database processing statement, and obtaining a result data set generated after the database processing statement is executed; and transmitting the result data set to a visual database terminal, and displaying the result data set on the visual database terminal according to a preset visual template. By implementing the technical scheme provided by the invention, the data query efficiency is improved.

The invention discloses a tomato maturity detection system and method based on deep learning, and relates to the technical field of fruit and vegetable detection models. Comprising the following steps: acquiring data from a robot perspective based on a natural scene of a tomato plantation, and making a tomato maturity detection data set; on the basis of a YOLO v5s model, an improved model YOLOv5s BCD model fusing a CA attention mechanism, a BiFPN neck feature fusion network and a Decup led Head head decoupling network is provided and serves as a tomato maturity detection model of the invention; experimental research of model training and effect comparison is introduced, the conditions of model evaluation and optimization are performed through a model comparison test, the tomato maturity grade is judged on the basis of the fruit red surface coloring proportion and in combination with fruit appearance characteristics, tomatoes with the required maturity grade are collected, the light weight and accuracy of edge side deployment are adapted, and the tomato maturity grade evaluation method is suitable for the tomato maturity grade evaluation and the tomato maturity grade evaluation and the tomato maturity grade evaluation and the tomato maturity grade evaluation and the tomato maturity grade evaluation. The system can be used as a visual system of a tomato harvesting robot, and technical support is provided for automatic picking.

The invention relates to the field of postoperative rehabilitation management, in particular to a postoperative rehabilitation management system for hand and foot surgical patients. The postoperative rehabilitation management system for the patient in the hand and foot surgery department comprises a data monitoring module, a rehabilitation plan prediction module and a rehabilitation plan dynamic adjustment module. By comprehensively analyzing the postoperative data of the patient, including the wound healing condition and the function recovery progress, the recovery trend of the patient in the postoperative training period can be predicted, a scientific initial function recovery training plan is generated, and clear guidance is provided for rehabilitation training of the patient; in the actual training process, the system further performs real-time dynamic adjustment on the rehabilitation training plan by comparing the difference between the actual recovery condition and the prediction trend; the method ensures that the rehabilitation training plan is always consistent with the actual recovery state of the patient, so that the rehabilitation effect is maximized, the rehabilitation time is shortened, and the risk of complications is reduced.

The invention belongs to the technical field of point cloud data processing, and discloses a point cloud smooth fitting method based on local density parameter adjustment and intensity optimization, and the method comprises the steps: obtaining original point cloud data, carrying out the preprocessing, carrying out the local density calculation, carrying out the dynamic adjustment of a search radius strategy, and carrying out the dynamic extension of a neighborhood, and after boundary judgment, performing boundary protection mechanism or joint weight calculation and updating smooth fitting on boundary points to obtain a point cloud visualization result graph after fitting reconstruction. The method effectively overcomes the defects of sparse point cloud smoothing, noise robustness, intensity feature retention, boundary feature fuzziness and the like in the prior art, can dynamically adapt to sparsity and non-uniform distribution of the point cloud, and remarkably improves the precision of smooth fitting of the point cloud and the optimization capability of intensity information. A more accurate and efficient technical scheme is provided for topographic surveying and mapping, underwater target detection and environment modeling of the forward-looking three-dimensional sonar.

The invention discloses a terahertz vortex phase plate based on a dumbbell-shaped metasurface. A plurality of metasurface units arranged on the same plane; the metasurface unit comprises a top basic unit metal structure layer, a middle dielectric layer and a substrate metal reflecting surface layer; the top basic unit metal structure layer is of a dumbbell-shaped structure, and the area is smaller than that of the middle dielectric layer and the base metal reflecting surface layer; the positions of all layers of the multiple metasurface units are correspondingly arranged, the phase distribution of the arranged multiple top layer basic unit metal structure layers on the plane meets the vortex beam phase distribution, and a metasurface array with the target topological charge is formed; the metasurface array is a terahertz wave incident plane. On the basis of improving the working efficiency, the metasurface technology and the geometric phase principle are combined, the whole-column arrangement of the terahertz vortex phase plate and the generation of terahertz vortex light beams with different topological charges are completed by designing a reflection type basic unit structure, and then the phase coverage corresponding to the specified topological charges is realized.

The invention provides a position prediction method and device based on an ultra wide band (UWB) digital key, and a position prediction system applied to the UWB digital key, the position prediction system comprises a UWB tag and a plurality of UWB anchor points arranged on a vehicle, and the position prediction method comprises the following steps: determining whether UWB positioning information of the UWB tag is invalid; when the UWB positioning information is invalid, acquiring a velocity vector and a historical track set of the UWB tag and orientation information of the vehicle; and inputting the velocity vector, the historical trajectory set and the orientation information into a position prediction calculation model to calculate a predicted position of the UWB tag. The technical problem that when an existing automobile digital key is blocked by an obstacle and the digital key cannot be connected to an automobile UWB anchor point, the function of the digital key is in a failure or invalid state, and the user experience is affected is solved.

The invention relates to the technical field of power system information management, and provides a power system alarm information processing method and device, electronic equipment and a storage medium. The power system alarm information processing method comprises the steps that basic data of at least one information system is acquired, and the information system is a system used for monitoring substation information; a fault event experience library is constructed based on the basic data, and the fault event experience library stores alarm logic and disposal suggestions of fault events; in response to the triggering operation of the alarm signal, generating a fault diagnosis result based on a fault event experience library and a preset rule engine; and carrying out distribution processing on the fault diagnosis result to obtain an information alarm processing result. According to the method, fault diagnosis is carried out in combination with the fault event experience library and the rule engine, automatic diagnosis and intelligent management of the alarm information of the power system are achieved, processing of the alarm information of the power system is more flexible and efficient, and the method can adapt to demand changes in different scenes.

The invention relates to the technical field of data retrieval, in particular to a hardware image retrieval method based on deep learning. The method comprises the following steps: acquiring a multi-angle image of hardware to be retrieved, and generating hardware feature data after geometric correction and feature enhancement; and performing shape fuzzy retrieval on a preset hardware image database by using the feature data to generate a shape-like hardware data set. And carrying out surface grating structure construction and optical response simulation on the shape-like data set to generate hardware grating data. Based on the grating data, simulating a light field and collecting four-dimensional reflected light to obtain a four-dimensional data tensor; and associating the tensor with a preset hardware material attribute, and establishing a spectrum-material attribute relationship. Feature matching is carried out according to the relation, and target hardware data are accurately retrieved; according to the hardware image retrieval method, the hardware image retrieval method can be faster and more accurate.

The invention relates to a multichannel PCM data real-time acquisition and output system and method based on an ALSA framework, and the system comprises a multichannel analog quantity input unit and a multichannel analog quantity output unit, and also comprises a system configuration unit, a memory resource unit, an AD acquisition conversion unit, an audio routing unit, a DA output conversion unit and an ALSA DMA cache region distribution unit. The ALSA DMA cache region allocation unit is used for building a PCM data path of a user space and a kernel space and providing an annular cache region queue for storing DMA resource descriptors, and the audio routing unit obtains the returned DMA resource descriptors from the ALSA DMA cache region allocation unit and allocates the DMA resource descriptors to an output end to complete synchronous multichannel input and output of PCM data. According to the system provided by the invention, the ALSA DMA buffer area distribution unit is adopted, so that the DMA configuration process is simplified, and the data transmission efficiency is improved.

The invention relates to the technical field of vehicles, and discloses a multifunctional steer-by-wire hand feeling simulator which comprises a steering column and an upper steering mandrel assembly arranged in the steering column, and the upper steering mandrel assembly is connected with a steering wheel through a spline. A speed reducing mechanism, a telescopic adjusting mechanism, an angle adjusting mechanism and an angle limiting mechanism are arranged on the steering column, the speed reducing mechanism is used for providing damping for the steering wheel to guarantee the road feeling, the telescopic adjusting mechanism is used for automatically adjusting the telescopic steering wheel, and the angle adjusting mechanism is used for adjusting the angle of the steering wheel. The angle limiting mechanism is used for limiting the number of turns of the steering wheel. The invention has the following advantages and effects: the driving experience of a driver can be improved, the driving space is saved, and the comfort level is improved.

The invention relates to a time-delay variable-gain robot virtual-real interaction remote control system and method. The method comprises the following steps: acquiring data of wrist, finger and pedal actions; mapping the collected action data to obtain a motion instruction of the slave end robot, and selecting and recording the motion instruction in a certain time period according to different treading states of the pedal action to obtain a corresponding motion instruction sequence; a motion instruction sequence is received, and according to the distance between a slave-end robot virtual simulation model and a virtual simulation environment space obstacle and the time delay generated when the motion instruction is sent from the master-end controller to the slave-end controller to be processed and fed back, mapping proportion adjustment is conducted on the motion instruction to obtain an operation instruction; and according to the operation instruction, the slave end robot is driven to execute a corresponding action, and action state feedback is carried out. The problems that in the current robot teleoperation technology, operation is unstable and operation is prone to collision due to unstable time delay jitter and sensing limitation are solved.

According to the multi-target tracking method and device, the storage medium and the computer equipment provided by the invention, in the to-be-tracked video of the non-first frame, the to-be-tracked target and the historical detection frame thereof are determined from the previous frame, the target position of the current frame is further predicted, and the target position of the current frame is determined according to the confidence of each detection target in the current frame. The detection targets with the confidence coefficients lower than a threshold value serve as candidate detection targets, and the detection targets with the confidence coefficients not lower than the threshold value serve as main detection targets; on the basis of the historical detection frame and the current prediction frame of each to-be-tracked target and the current detection frame of each main detection target, performing multiple times of matching on each current prediction frame and the current detection frame of each main detection target; and matching the current prediction frame which is not matched with the current detection frame of each candidate detection target, and taking each main detection target and the successfully matched candidate detection target as a to-be-tracked target of the current frame. Therefore, the tracking error can be reduced to the greatest extent under the conditions of shielding and detection loss.

The invention discloses an intelligent control method and device, electronic equipment and a readable storage medium, and belongs to the technical field of intelligent control. Under the condition that the gesture instruction does not conflict with the voice instruction, performing multi-dimensional normalization on the gesture feature vector to obtain a to-be-fused gesture vector; performing normalization processing on the voice feature vector based on context semantic perception to obtain a to-be-fused voice vector; fusing the to-be-fused gesture vector and the to-be-fused voice vector to obtain a fusion vector, and determining a fusion instruction corresponding to the fusion vector as a target execution instruction; and under the condition that the gesture instruction conflicts with the voice instruction, selecting a target execution instruction according to a first confidence coefficient corresponding to the gesture instruction and a second confidence coefficient corresponding to the voice instruction. The self-adaptive normalization strategy is designed for different modal data, the data fusion accuracy is improved, meanwhile, the gesture and voice instruction conflict can be intelligently and flexibly processed, and the accuracy of determining the real intention of the user is improved.

The invention relates to the technical field of data processing, and discloses a BIM-based intelligent weekly progress visual display system and method, and the system comprises a weekly report management module which is used for uploading and storing weekly report progress data of a target project; the model management module is used for uploading and storing the building information model of the target project; the model association module is used for associating the weekly report progress data with the building information model; the coloring module is used for setting colors of different components in the building information model according to the construction progress in the weekly report progress data to obtain a construction progress three-dimensional model; and the display window module is used for displaying the building information model and the construction progress three-dimensional model. According to the method, the weekly report progress data and the building information model are automatically associated and dynamically updated in real time, the weekly progress of the target project is visualized, fine management of the project construction progress is assisted, the project management level is improved, predictable management and control correction is achieved, and it is guaranteed that the project construction period is achieved as much as possible.

The invention relates to the technical field of data processing, and provides a data processing method and device based on an SaaS platform and a storage medium, and the method comprises the steps: collecting data from different data sources through a unified data collection interface provided by the SaaS platform; automatically identifying and converting data acquired from different data sources to obtain target format data; storing the target format data to a cloud database by adopting a distributed storage scheme; analyzing the target format data stored in the cloud database based on a configurable data processing pipeline provided by the SaaS platform; and displaying the analysis result of the target format data to a user in a visual mode. According to the technical scheme provided by the invention, data with relatively large data volume can be flexibly and efficiently processed at low cost.

The invention discloses an unmanned aerial vehicle group path planning method based on dual-population multi-constraint evolution. The method comprises the following steps: determining a first population and a second population according to unmanned aerial vehicle paths; generating filial generations based on the first population and the second population to obtain a first combined population and a second combined population; and sequentially determining the number of feasible solutions in the first combined population and the second combined population, performing updating in combination with a contribution value threshold, performing path planning on the unmanned aerial vehicle group, and judging whether a termination condition is reached, if not, returning to continue iteration, and if yes, outputting a non-dominated solution according to the updated first combined population and the second combined population. Based on the population contribution degree, an appropriate constraint calculation mode is adaptively selected to obtain a feasible solution mode, the solution efficiency and the solution quality can be remarkably improved, and meanwhile the stability and adaptability of the algorithm are enhanced.

The invention relates to the technical field of computer machine vision, and discloses a plant growth condition monitoring method, system and device and a storage medium, and the method comprises the steps: obtaining a target plant image, inputting the preprocessed target plant image into an improved YOLOv9 model (comprising a backbone network, a feature fusion network and a prediction classification network), and obtaining a target plant image; obtaining a target plant detection image; performing feature extraction on the preprocessed target plant image through a backbone network to obtain plant multi-scale features; performing feature conversion, branch processing and feature fusion on the plant multi-scale features through a feature fusion network to obtain plant fusion features; performing classification processing on the plant fusion features through a prediction classification network to obtain a target plant detection image; and inputting the target plant detection image into an image semantic segmentation model (including an image super-resolution network and an image segmentation network) to obtain a plant binary image. The method effectively improves the accuracy of plant growth condition monitoring.

The invention relates to a fundus image classification method and system based on a dual-path neural network and a dynamic weight adjustment fusion module, and belongs to the technical field of medical image processing. Obtaining an original fundus image, and preprocessing the fundus image; the preprocessed image is input into a feature extraction module for processing, the feature extraction module comprises a local path and a global path, and the local path module extracts disease detail information by using a convolutional neural network; the global path module extracts disease contour information by using a visual converter; fusing the images passing through the local path and the global path through a dynamic weight adjustment fusion module; and the fundus images are graded through a full connection layer of a classifier and an activation function Softmax layer of the fused images. According to the method, the advantages of the neural network and the image processing technology are combined, the dynamic weight adjustment fusion module is utilized to intelligently distribute the weight, the feature integration process is optimized, and the accuracy and efficiency of eye fundus image grading are improved.

The invention relates to the technical field of molecular dynamics, in particular to a molecular system simulation method and system, computer equipment and a storage medium. The molecular system simulation method at least comprises the steps that virtual site information of virtual sites in a molecular system is calculated in parallel based on all calculation units, and the virtual site information and charge information of the virtual sites are updated to a virtual site information table in a storage unit; calculating the atomic information of each atom in the molecular system in parallel based on each calculation unit; updating the virtual site information of each virtual site based on the obtained atomic information to obtain new virtual site information; distributing the obtained new virtual site information to each atom to obtain new atom information of each atom; and carrying out constraint processing on the boundary of the dynamic system where the molecular system is located based on the new atomic information, and outputting integrated data of the dynamic system. The invention provides a molecular system simulation method with higher efficiency for a large-scale system.

The invention provides a data processing method, electronic equipment and a storage medium, and relates to the technical field of data processing. The method comprises the steps that on the basis of preset mapping relation information and multiple self-defined preset data processing statements, multiple alternative semantic trees corresponding to the preset data processing statements are constructed, and the preset mapping relation information is information representing the mapping relation between the preset data processing statements and execution statements in a database to be used; according to the obtained format of the to-be-processed data, matching with a plurality of alternative semantic trees, and determining a target semantic tree corresponding to the to-be-processed data; determining a query statement corresponding to the to-be-processed data according to a target semantic tree corresponding to the to-be-processed data; and querying the to-be-used database based on the query statement corresponding to the to-be-processed data to obtain a data query result. And the processing steps of the business data are simplified, so that the data query result reflects the change condition of the business data.

The utility model discloses an RCO heat accumulating type catalytic combustion device which comprises a combustion tank, air pipes communicated with the two sides of the combustion tank and catalytic materials filled in the combustion tank, an adjusting structure is arranged in the combustion tank, and a transmission structure located at the bottom of the adjusting structure is arranged in the combustion tank. The adjusting structure comprises a bearing plate fixedly connected to the interior of the combustion tank. The bearing plate is extruded through the transmission structure, so that the bearing plate is shifted in the combustion tank, catalytic materials on the surface of the bearing plate are pushed in the shifting process of the bearing plate to complete the displacement and transposition state, the catalytic materials with the changed positions can expand the contact uniformity with waste gas, the catalytic effect is improved, and the catalytic effect is improved. According to the heat accumulating type catalytic combustion device, the storage mode of a catalyst inside an existing heat accumulating type catalytic combustion device can be improved, a direct spreading mode is replaced for spreading, the catalyst at the bottom can make full contact with internal waste gas in the heating process, and the release effect of the catalyst is prevented from being affected.

The invention discloses a metal sheet carrying device and belongs to the technical field of sheet carrying, the metal sheet carrying device comprises a cross beam, the top of the cross beam is fixedly connected with two pieces of supporting steel, the tops of the two pieces of supporting steel are fixedly connected with hanging steel, the tops of the hanging steel are fixedly connected with hanging rings, and the front end and the rear end of the cross beam are provided with adjusting mechanisms. An adsorption mechanism is installed on the top of the cross beam, and a rotating mechanism is installed on one side of the adjusting mechanism. By designing the adjusting mechanism, the adsorption mechanism and the rotating mechanism, a metal sheet can be adjusted according to the width of the sheet during carrying work, the edge of the sheet can be better attached by adjusting the distance between suction cups, the suction force is increased, the risk of slipping or falling in the carrying process is reduced, and the safety is improved; the metal iron plate side-standing carrying device can carry metal iron plates in a side-standing mode, effectively utilizes the vertical space, can carry the metal iron plates more flexibly especially in a narrow working environment, and improves the use efficiency of the space.