44 results about "Visual tool" patented technology

The invention relates to a conveying belt tearing positioning and tracing system based on RFID and multi-sensor fusion, which effectively overcomes the influence of belt slipping and elastic deformation through multi-sensor fusion and motion compensation, and realizes accurate positioning of a tearing point. Meanwhile, through cross validation of vibration, vision and infrared data, the recognition accuracy is improved to 99.5% or above, false alarms caused by environmental factors (such as dust and moisture) are greatly reduced, stable operation in a severe industrial environment is ensured, and high reliability and low false alarm rate are achieved. And thirdly, the system supports long-term data tracing and intelligent operation and maintenance, and a user can analyze a tearing trend and identify a high-frequency fault area by storing historical data of at least 180 days and providing a visual tool, so that a maintenance plan is optimized, predictive maintenance is realized, and faults are reduced from the source.

The invention discloses a social group discovery system based on fused semantic information. The system comprises a presentation layer, a presentation layer, a service layer and a data layer. The presentation layer supports checking of a group division result; the display layer displays a user structure network, group dynamic division and group characteristic indexes through a visual tool; the service layer is composed of a semantic tag generation module, a user structure analysis module, a social group division module and a social group analysis module, semantic and structural features are fused through a graph attention network, and accurate group division is realized in combination with an adaptive embedded graph aggregation algorithm; and the data layer stores the original data set of the social platform and the processed data. The system inputs user texts into a large language model to generate semantic tags, node embedding is generated by calculating node association weights, group division is completed by combining an adaptive embedding graph aggregation algorithm, and finally group characteristics are quantitatively analyzed from the aspects of cohesion, conductivity and the like. The system is suitable for scenes such as vertical domain community mining.

The invention discloses a diabetic intestinal preparation quality prediction method based on SHAP, and relates to the technical field of medical artificial intelligence, and the method comprises the following steps: collecting clinical feature data of a diabetic, and carrying out the preprocessing of the clinical feature data; constructing a diabetic intestinal preparation quality prediction model, taking the preprocessed clinical feature data as model input, and taking the intestinal preparation qualification rate as output; and an SHAP interpretable artificial intelligence framework is introduced to analyze and attribut the prediction model. According to the method for predicting the intestinal preparation quality of the diabetic patient, the prediction model is constructed by integrating multi-dimensional clinical data and applying an ensemble learning algorithm, so that the prediction accuracy and stability are remarkably improved, the limitation that the traditional method depends on doctor experience is broken through, the judgment inconsistency caused by individual difference and diagnosis and treatment pressure is reduced, and the prediction accuracy and stability are improved. The interaction between the features is displayed through a visualization tool, and the interpretability and transparency of the model are enhanced.

The invention discloses an intelligent construction site multi-dimensional collaborative management method based on BIM and Internet of Things, and the method comprises the steps: carrying out the grouping processing of data through employing a clustering algorithm according to a unified data set after fusion, classifying the data of similar processes, and carrying out the recalculation through adjusting a clustering center if a grouping result shows that the data distribution is not uniform, determining an inter-process logic dependency relationship graph; the method comprises the following steps: acquiring real-time field update data as input through a complex coupling process sequence, predicting the completion time of each process in the sequence by adopting a neural network model, judging whether a prediction deviation is within an allowable range or not, and obtaining an accurate progress prediction result; and deducing a result through a final dynamic resource demand, generating a summary report for a project end process, carrying out mapping display on the result by adopting a visual tool, judging whether a construction period requirement is met or not, and obtaining complete construction management output.

The invention discloses a dialogue type data analysis method and system based on a large model application technology in the field of rail transit, and the method comprises the steps: extracting data from a rail transit multi-source system through a streaming or batch processing mode, constructing an original data layer (ODS), and carrying out the automatic processing based on a materialized view technology to generate a detail layer (DWD), a summary layer (DWS) and an application data layer (ADS). And efficient hierarchical management and nearly real-time updating of the data are realized. The system receives a data query demand input by a user in a natural language form, performs intention recognition and semantic understanding by using a large model technology, matches a related data table structure in combination with vector retrieval, automatically generates a safe and accurate SQL query statement, and executes a query operation. And the query result is returned in a JSON format, and the front end dynamically generates a chart through visual tools such as ECharts and the like, so that a dialogue type data analysis result which is analyzed as required is realized.

The invention discloses a digital human live broadcast microphone connection and video generation method based on computer vision, and belongs to the technical field of artificial intelligence, and the method comprises the following steps: 1, real-time vision capture and action expression driving: employing a computer vision tool to carry out face key point detection, facial feature points of lips, eyes and the like of a live anchor or a microphone-connected guest are accurately captured; and then a voice conversion model is matched, microphone-connected audio features and lip key point movement are bound, precise synchronization of the sound and the lip is achieved, and for limb actions, limb joint data can be collected through a monocular camera in combination with a posture estimation algorithm or a simple action capture device and mapped to a digital human skeleton model. According to the digital human live broadcast microphone connection and video generation method based on computer vision, through multi-dimensional technical innovation and systematic design, the problems of insufficient sense of reality, interaction lagging, poor scene adaptability and the like existing in current digital human application are effectively solved.

The present invention is directed to a system and method for displaying interior design proposals for a given room according to known design principles and user preferences. The system is operative to convert user-input sketches and / or blueprints of two-dimensional space geometry into digital floor plans and identify and propose blocks of appliances and / or articles of furniture as design elements defined by their dimensions and any limiting characteristics. Each block is functionally and dimensionally appropriate for the floor plan based on user preferences, as filters, and floor plan limitations and may be defined by minimum space requirements, anchor point restrictions, and configuration restrictions. The system and method may further comprise visual tools for conveying a user's creative taste, style, and aesthetic for each floor plan, and be operative to display hyperlinked shopping lists in accordance with filtered blocks of design elements.

The invention provides a calibration piece, which is used for calibrating the installation position of a cutter on a cutter base, and comprises a sleeve mechanism, the sleeve mechanism is provided with a measuring tool part and a sleeve hole penetrating along a first direction, the measuring tool part is located at one side of the sleeve hole in the first direction, and the measuring tool part is located at the other side of the sleeve hole in the first direction; the sleeve mechanism can be arranged on the periphery of the tool base in a sleeving mode based on the sleeve hole. The stopping part is movably arranged on the measuring tool part in the first direction, and in the state that the sleeve mechanism is arranged on the tool base in a sleeving mode and the stopping part is positioned on the measuring tool part, a tool bit of the tool can abut against the stopping part so as to calibrate the protruding length of the tool bit relative to the tool base; the problems that the efficiency is too low, the precision is not high and the consistency is poor due to the fact that an operator uses a measuring tool to conduct visual tool setting in an existing traditional tool installation mode are solved.

An electric spark machining coordinate collecting and executing method comprises the steps that S1, in a calculation environment with a three-dimensional geometric processing function, a first coordinate position where interference-free entering and a second coordinate position where machining is terminated are obtained; s2, mapping from the reference zero point to a physical processing zero point on an electric spark machine tool; s3, after the physical electrode is automatically replaced on the machine tool, the geometric deviation of the physical electrode is measured, and the deviation serves as a tool offset parameter to be written into a numerical control system; and S4, an electrode is controlled to execute electric discharge machining from the first coordinate position to the second coordinate position. According to the method, coordinates are obtained through simulation, the risk of motion interference of the electrode in a complex inner cavity is eliminated in advance, and the problem that visual tool setting cannot be achieved during blind hole machining is solved. A large amount of coordinate acquisition work is transferred to the computer end, the machine tool end only needs to execute automatic calibration and machining, the shutdown waiting time is remarkably shortened, and continuous automatic production is achieved.

Disclosed systems and methods may receive real-time, external credit line information into one or more source applications. Thereafter, the real-time, external credit line information may be merged with internal credit line information and legacy information via a data orchestrator. The merged information may be further consolidated via the data orchestrator resulting in one or more micro batches of merged credit line information that are transmitted to a credit line decisioning system at frequent intervals. At least one exclusion rule may be applied to the transmitted credit line information, and thereafter, the transmitted credit line information may be provided to an interactive visual tool for real-time credit line management. The interactive interface allows for a user to incorporate real-time credit line information into credit line management decision-making. Thus, the credit line management system described herein provides real-time credit line information to enable real-time decision-making regarding credit line increase, decrease, and closure.

The application discloses a kind of general software development construction methods suitable for multiple scene requirements, belong to software development technical field.The method includes four steps: one is abstracted multiple scene common characteristics, peeling specificity, establishes extensible reusable standardized metamodel system;Second, based on the metamodel system, establish scene adaptation mechanism, realize accurate mapping with different scenes through configuration;Three are using pre-packaged functional components and visual tools, arrange the application process that meets target scene;Four are starting automated build release engine, automatically complete code generation, output directly running application.The method also includes unified access management and dynamic update extension steps, can greatly improve software development efficiency, reduce cost, enhance scene adaptation flexibility and expansibility, reduce maintenance iteration difficulty, suitable for the differentiation scene of multiple industries, multiple terminals, multiple user groups, realize the rapid construction and efficient iteration of general software.

The application provides a multimodal fidelity hallucination detection method and system based on closed-loop multi-dimensional verification, belonging to the technical field of large visual language models, comprising: S1: decomposing the user instruction input into the LLM and the answer of the LLM into structured units; S2: through the collaborative mode of the LLM and the visual tool, the consistency score of the assertion is calculated from three dimensions of text-instruction consistency, text-image consistency and logical consistency, all verification results are bound to a unique evidence ID, and the assertion without evidence support is determined as unknown; S3: performing type-specific calibration on the consistency score, and then generating a unified hallucination score H through weighted fusion; S4: providing traceable evidence support for the detection result, and triggering the abstention mechanism when the evidence is insufficient, so as to ensure the credibility and practical value of the detection result. The method realizes accurate detection of three types of hallucinations, i.e., instruction inconsistency, text-image separation and logical inconsistency.

The invention relates to the technical field of building engineering, and discloses a building construction whole process intelligent monitoring system based on a BIM internet of things. According to the system, environment data and structure state data of a construction site are acquired in real time through the data acquisition module and uploaded to a cloud end through the Internet of Things technology, the timeliness and accuracy of the data are ensured, the data are deeply analyzed through the data analysis module, potential construction hidden dangers are recognized, a construction progress analysis report is generated, and the construction progress is analyzed through the data analysis module. The monitoring and early warning module can monitor the state of a construction site in real time, set a risk early warning threshold value, give an abnormal alarm in time and ensure construction safety, and the decision support module formulates a coping strategy according to alarm information, assists a management layer in evaluating the effects of different countermeasures through a visual tool, and finally converts the effects into specific adjustment measures of the execution module. Therefore, potential safety hazards and resource waste are effectively reduced, the construction progress is optimized, the overall construction quality is improved, and intelligent and refined management of the building industry is truly achieved.

The invention discloses a multi-scene building visualization system based on a BIM technology and virtual reality, and relates to the technical field of building informationization, and the system comprises the steps: collecting three-dimensional point cloud data and RGB-D image streams of a scene building, and carrying out the linkage of a BIM collaborative model, and obtaining a building digital gene pool; activating a role customization mode in a multi-source fusion interactive scene based on the user identity token, and dynamically loading a supervision visualization tool set matched with the user identity; operating and supervising the visual tool set in the terminal equipment, generating a design change instruction, returning the design change instruction to the BIM collaborative model, triggering version updating of the building digital gene pool, and generating a visual multi-scene building; the supervision tool set is dynamically loaded based on the user identity, design change instruction return and version updating of the building digital gene pool are supported, and the precision, efficiency and intelligent level of multi-scene collaborative management in the whole life cycle of the building are improved.

The invention discloses a modularization software integration deployment method of an underwater acoustic system, and belongs to the technical field of underwater acoustic system software development and deployment, and the method comprises the steps: packaging an underwater acoustic system function module assembly, generating a container mirror image, and storing the container mirror image into an assembly warehouse after multi-dimensional verification; arranging a business process through a visual tool based on the business demand and generating an executable business link configuration file; and a deployment plan is generated through a resource matching algorithm in combination with deployment environment constraints, and finally deployment is implemented and verification is performed through a container arrangement tool. According to the method, data closed-loop management of'development-deployment-operation and maintenance 'is realized, and the success rate and the stability of underwater acoustic system software integration deployment are effectively improved.

The invention provides a text generation method based on a large model and multi-source heterogeneous data, which comprises the following steps: collecting multi-source unstructured data such as an enterprise internal system and public network resources, and cleaning, decoupling and labeling metadata to form a high-quality data set; converting the text into a vector by utilizing a pre-training language model, and constructing an RAG knowledge base supporting efficient retrieval; modular components are arranged through a visual tool to form a workflow, and the workflow is scheduled by a workflow engine; combining the large model integrated with the RAG mechanism with field data to generate a text after fine tuning; defining rule constraint generation content based on industry specifications; and a feedback loop optimization model is formed through automatic evaluation and manual review. The system correspondingly comprises a data processing module, an RAG knowledge base module, a business process arrangement module and the like. According to the scheme, multi-source data are integrated, efficient and high-quality generation of the cross-domain unstructured text is realized by utilizing the large model and the RAG technology, and the method has the advantages of high flexibility, high automation degree, strong generated text professionality and the like, and is suitable for various complex text processing scenes.

The invention discloses a reasoning segmentation method based on alternate cross-modal reasoning and tool calling intelligent agent, which comprises the following steps: modeling a reasoning segmentation task into a multi-round decision process in a pixel space, and constructing an interaction environment consisting of the intelligent agent and an external visual tool. And the intelligent agent gradually generates a segmentation result through multiple rounds of interaction with an external visual tool. And obtaining a training data set including an input image, a natural language instruction and a real segmentation mask, and training the intelligent agent by using cold start supervision fine tuning and reinforcement learning. In the cold start stage, an intelligent agent is guided to generate a preliminary segmentation result through a supervision signal, and in the reinforcement learning stage, a tool calling strategy is optimized through a reward mechanism, so that the adaptive segmentation capability is realized. Based on the trained intelligent agent and an external visual tool, the intelligent agent can effectively complete an image segmentation task in multiple rounds of interaction. According to the method, by combining cross-modal reasoning and tool calling, the adaptivity and accuracy of the intelligent agent in a complex visual task are improved.

The invention provides a method and a system for building a scenic spot tourist flow prediction model based on artificial intelligence. The method mainly comprises the steps of data collection and processing, feature engineering and initial model training, model optimization and key factor screening, data feedback and continuous model improvement, prediction result display and explanation and the like. The system integrates structured data and unstructured data, and performs data training and model optimization by using a deep learning algorithm. In addition, the data display and interpretation module enhances the interpretability of the model, displays the data processing flow and the prediction result through a visual tool, and helps the user to understand the model logic. Compared with a traditional method, the dynamic change of the popularity of the scenic spot can be more accurately captured, the prediction precision and efficiency are improved, and the method is suitable for a rapidly changing tourism market.

The invention relates to the technical field of tool detection, in particular to a 3D vision tool end face axial run-out detection mechanism and an outer diameter online detection method. The mechanism comprises a rack, a back plate, a test seat, an adjusting assembly and a protective cover. The back plate is slidably arranged in a strip-shaped hole of the rack, and the adjusting assembly comprises a tool mounting base and a multi-dimensional moving mechanism and is used for adjusting the position of a tool. The testing seat is arranged on one side of the machine frame in a turnover mode, and a laser displacement sensor and a visual detection unit are arranged in the testing seat and used for detecting end face run-out and the outer diameter size of the tool respectively. The protective cover is slidably arranged outside the test seat in a covering mode. The method comprises the steps that tool setting is conducted through the adjusting assembly, end face run-out is detected through the laser displacement sensor, images are shot through the industrial camera, the outer diameter size is analyzed, and multi-angle measurement can be achieved by rotating the shooting pose. The mechanism can realize high-precision detection of multiple parameters on one device, has protection and position adjustment functions at the same time, and adapts to detection requirements of cutters of different specifications.

The utility model relates to the technical field of intelligent cockpits, and discloses a visual tool of an intelligent cockpit of a power generation enterprise, which comprises a seat, side plates are symmetrically and fixedly connected to the outer side wall of the seat, and supporting tables are fixedly connected to the sides, away from the seat, of the two side plates. The first auxiliary panel and the second auxiliary panel are driven to move towards the two sides through the driving mechanism arranged in the supporting assembly, at the moment, the first auxiliary panel and the second auxiliary panel are located on the two sides of the main panel, and the first auxiliary panel, the second auxiliary panel and the main panel can be expanded; a user can better use a plurality of screens, the situation that the viewing effect is reduced due to limitation of a single screen is avoided, the first auxiliary panel and the second auxiliary panel are synchronously driven by the driving mechanism, and when the first auxiliary panel and the second auxiliary panel are not used, the first auxiliary panel and the second auxiliary panel can be contracted to the back face of the main panel and can be rapidly unfolded when used.

This invention discloses a user-defined template management system and method. The system allows users to intuitively construct table templates through a visual tool module. The server-side control module sends this template data and the original data from the target terminal to a comparison service module via a protocol call comparison command. The comparison service module first compares the table sample data and the data sample data, accurately identifying the protocol data of the table sample data and the protocol data of the data sample data to be updated. Then, it packages and encapsulates the data into encapsulated difference data according to a private protocol format and feeds it back to the server-side control module via a message queue. Finally, the server-side control module issues a configuration update command to the target terminal based on the encapsulated difference data. This achieves full automation from user-defined templates to accurate terminal data updates, effectively improving the flexibility of template management and the accuracy and timeliness of data updates, and avoiding errors that may be caused by manual operation.

This invention discloses a component-based software integration and deployment method for underwater acoustic systems, belonging to the field of underwater acoustic system software development and deployment technology. The method involves encapsulating functional module components of the underwater acoustic system and generating container images, which are then stored in a component repository after multi-dimensional verification. Based on business requirements, a visual tool is used to orchestrate business processes and generate executable business link configuration files. Combined with deployment environment constraints, a deployment plan is generated using a resource matching algorithm. Finally, deployment is implemented and verified using a container orchestration tool. This invention achieves closed-loop data management of 'development-deployment-operation', effectively improving the success rate and stability of underwater acoustic system software integration and deployment.

This invention discloses a variable diameter tool holder and its machining system, relating to the field of machining tool technology. It includes a main tool holder structure, a rotary telescopic variable diameter mechanism, a rotary transmission mechanism, and a hydraulic clamping mechanism. It employs upper and lower double-layered staggered blades, driven by rotation to radially extend and retract, achieving small-diameter feed when contracted and large-diameter machining when expanded. Hydraulically driven upper plate lifting and lowering allows the upper and lower blades to engage and clamp, improving rigidity. The lower blades have tool positions, while the upper blades are equipped with a macro-light camera for visual tool setting. This invention solves problems such as interference, poor rigidity, tool vibration, and cumbersome assembly and disassembly in machining narrow-mouth internal cavities, significantly improving machining quality and efficiency. It is suitable for precision machining of narrow-mouth tank internal cavities on vertical lathes.

The utility model discloses a financial transaction risk early warning information visualization display device which comprises a supporting frame, the supporting frame is a rectangular structure frame, a plurality of split screen supports are arranged on the supporting frame in a matrix mode, a plurality of distributed displayers are arranged on one sides of the split screen supports, and the distributed displayers are arranged on the other sides of the split screen supports. The utility model relates to the technical field of financial visual tools, and the financial transaction risk early warning information visual display device dynamically adapts to the position change of an observer through three-stage cooperative control of horizontal rotation, pitching adjustment and spherical compensation, and when personnel on a station carry out adjustment, a control system synchronously drives horizontal steering and pitching angle adjustment, so that the visual display of financial transaction risk early warning information is realized. By matching with the micro attitude compensation of the spherical joint, the normal of the display surface accurately points to the viewpoint of an observer, the plane consistency can be kept during multi-screen linkage, the spatial pose of each unit is independently adjusted in a split-screen mode, the organic unification of rigid support and flexible adaptation is realized, and the visual comfort of multi-angle observation is remarkably improved.

The application discloses a DCLK1 fluorescent probe or a pharmaceutically acceptable salt thereof, and a structural general formula is as follows: The application fuses a high-selectivity DCLK1 inhibitor DCLK1-IN-1 as a pharmacophore with an environment-sensitive fluorophore, and designs an environment-sensitive DCLK1 fluorescent probe. When the probe molecule is combined with the DCLK1 protein, the hydrophobic environment of the protein cavity enables the fluorescence of the probe to be turned on, the intensity is increased, and the signal-to-noise ratio is significantly improved, thereby realizing selective and sensitive imaging of the DCLK1 protein, and the DCLK1 fluorescent probe is expected to become a visual tool for in-depth research on the biological and pharmacological effects of the DCLK1 protein.

The application discloses a chip system test requirement analysis method, device and medium, relates to the chip system test technical field, and aims at the problem of incomplete analysis in the chip system test requirement analysis at present, and provides a chip system test requirement analysis method. The test analysis process is completed by a test personnel with the aid of a visual tool in a four-layer step-by-step test analysis mode, the level and experience difference of different test personnel are isolated, and the quality of test analysis is improved. In addition, the method can also guide the test personnel to input corresponding description information during requirement analysis at each layer, so that all necessary elements required by a requirement document in a test analysis process are acquired, and omission is avoided. Furthermore, test case description is automatically generated based on the acquired description information, and is output as a requirement document of the test project, so that the problems of omission and the like caused by manual output of the requirement analysis document can be effectively avoided.

The invention discloses a rapid detection system and method based on pesticide toxicity, and relates to the technical field of pesticide detection analysis, and the system comprises a data acquisition module, a data preprocessing module, a multi-dimensional feature extraction module, a prediction model construction module, a toxicity grade prediction module and a visual aided design module. According to the method, invalid tests of pesticides with extremely low toxicity or extremely high toxicity can be preferentially excluded, so that the accuracy of LC50 value prediction results is higher through integration of multi-dimensional characteristics, the influence of soil environment parameters on the toxicity can be accurately captured, the prediction results better meet the actual requirements of pesticide ecological risk assessment, and the method is suitable for popularization and application. Meanwhile, a visual tool is adopted to directly output the dose = the toxicity response relation and the recommended dose range, the blindness of traditional empirical dose design is solved, the method can be migrated to toxicity prediction of other organisms, a new model can be constructed only by replacing corresponding historical test data, and meanwhile the method can also be applied to toxicity screening in the pesticide research and development stage.

The invention relates to the technical field of crossing of embedded systems and machine vision, in particular to an FPGA-based DVS camera integrated use method, which comprises the following steps: constructing an embedded Linux system at a PS end of an FPGA, and configuring a USB driving environment by starting libusb1 and libusb-complex libraries and compiling a system mirror image; performing cross compiling on an equipment interface library and a visual processing library of the DVS camera aiming at the ARM architecture; developing an application program based on an ARM architecture, pulling event data through the equipment interface library and deploying the event data to the FPGA; and monitoring logs through a serial port tool and verifying data integrity and real-time performance through a visual tool. The problems that a traditional DVS camera only supports an x86 architecture platform and is poor in compatibility with an FPGA embedded system are solved, a set of complete FPGA-DVS integration scheme is constructed through driving adaptation, architecture compatible library compiling and application layer interface development, the deployment threshold of the DVS in an embedded scene is lowered, and the real-time performance and the integration degree of the system are improved.

The invention relates to the technical field of Web front-end performance monitoring, and particularly provides a Vue component rendering performance monitoring method and system, a terminal and a medium, and the method comprises the steps: extracting a response type dependency graph through an abstract syntax tree in the construction stage of a Vue component; injecting a performance probe into the life cycle hook of the Vue component; when the Vue component runs, acquiring rendering performance indexes of the Vue component through the injected performance probe, including component-level rendering time consumption, a memory occupancy peak value, an Element-Plus specific index and a mobile terminal special index; and executing a cross-level attribution algorithm according to the acquired rendering performance indexes to generate an attribution result, and generating a visual rendering performance result according to the attribution result. According to the method, Vue component-level performance accurate monitoring is realized through AST analysis, Element-Plus and mobile terminal performance acquisition is provided, and the component rendering performance optimization efficiency is improved by combining intelligent analysis and a visual tool.

The application discloses a double-station five-axis mechanism correlation calibration method, solves the calibration problem of the cooperative work of the double-station five-axis mechanism, realizes the accurate correlation of a visual tool and a work tool by establishing a unified base coordinate system and a shared mark point calibration reference, greatly improves the cooperative work precision of the double-station, and has the advantages of simple calibration process and strong operability, does not need complex equipment modification and additional calibration tools, can be completed only by conventional visual calibration and position collection, reduces calibration cost and operation difficulty, and is suitable for various types of tool head combinations and can meet the needs of different process scenes.