702 results about "Recognition algorithm" patented technology

The invention relates to the technical field of building construction monitoring, and discloses a method for evaluating mechanical properties of a steel structure building construction whole process based on digital twinning. The method comprises the steps of establishing a digital twin model fusing multi-source information, and performing real-time linkage with a sensor network arranged on site. Collected data such as deformation, temperature and wind speed are processed through dynamic fusion and an anomaly recognition algorithm, model parameters are continuously corrected, and real-time dynamic high-fidelity mapping of the mechanical state in the construction process is achieved. And based on the updated model, the intelligent analysis module performs cooperative calculation, autonomously identifies construction abnormity, quantitatively predicts potential risks, generates a process optimization decision instruction and feeds back the process optimization decision instruction to a site. According to the method, a closed-loop regulation and control mechanism from data perception, model analysis to decision execution is constructed, and accurate online evaluation of construction mechanical properties and active prediction control of safety risks are realized.

The invention belongs to the field of edible mushroom planting control, and particularly relates to a robot collaborative operation control method and system in the edible mushroom planting process, and the method comprises the steps: an inspection robot scans a mushroom bed, judges the maturity of edible mushrooms in combination with an image recognition algorithm, calculates the regional to-be-picked density, and generates a to-be-picked task list; according to the list, a preset parameter table and a curve, dynamically generating a coordinated picking task control instruction including the target coordinate, the number of the robots and the specific clamping posture, force and lifting speed of each robot; after the picking robot responds to the instruction to execute operation, a signal is fed back to the nearest inspection robot, secondary imaging confirmation is triggered, if residues exist, a supplementary picking list is generated according to the residue density, execution is repeated, and therefore efficient, accurate and lossless automatic cooperative harvesting of the edible mushrooms is achieved.

The invention provides a driving authority management method and device based on biological feature verification and a medium, and belongs to the technical field of vehicles. The method comprises the steps of collecting facial features, voiceprint data and driving behavior data by detecting a starting operation of a current driver; the identity of the current driver is recognized by using a multi-modal biological recognition algorithm, and the accuracy and anti-counterfeiting capability of identity verification are improved by adopting a bimodal fusion scheme of face recognition and voiceprint recognition; when it is determined that identity recognition of the current driver succeeds, a driving account is determined to judge whether the driver has the use permission or not, a binding relation of driver identity-account-function permission is established, and accurate matching of the intelligent driving permission and the driver qualification is achieved; the open state of the intelligent driving function is dynamically controlled based on a permission verification result, if the permission exists, the intelligent driving function is automatically prepared, if the permission does not exist, the function is locked, a clear prompt is given, and the safety risk that the intelligent driving function is used before being learned is avoided from the source.

The invention discloses a frozen soil temperature feature recognition algorithm and a moisture and electric field state prediction model, and the method comprises the steps: synchronously collecting temperature, moisture, resistivity and meteorological data through arranging a multi-depth sensor array, and constructing a multi-source data set with time-space alignment; utilizing continuous wavelet transform and cross-modal coherence analysis, and combining a hidden semi-Markov model to extract a depth-periodic characteristic profile; a time domain fusion Transform model embedded with physical constraints is adopted to realize frozen soil multi-parameter physical consistency prediction and uncertainty quantification; and finally, outputting a prediction result through a visual interface and setting an automatic alarm. The problem of cycle identification caused by weak and non-stable deep frozen soil signals is solved, the prediction accuracy and the physical rationality are remarkably improved, an end-to-end solution from data quality control to early warning is formed, and the reliability and the practicability of frozen soil engineering monitoring are greatly enhanced.

The invention relates to a wall-climbing robot system for detecting a coating of a steel pipe truss structure and an application method, and the system comprises a self-adaptive magnetic adsorption mobile platform which can be adaptively and stably adsorbed on steel trusses with different pipe diameters to execute an inspection task; the robot platform is provided with an autonomous positioning navigation system composed of a laser radar, binocular vision and an airborne computer, and the position of the robot is positioned in the steel pipe truss without satellite signals through real-time mapping and a positioning algorithm of tight coupling of laser and vision. The front part of the robot is provided with an electrically foldable surrounding type array camera, so that full-surface coverage type shooting of a steel pipe component is realized, and real-time identification of coating apparent defects is executed in combination with an artificial intelligence disease identification algorithm in an airborne computer; and the lower part of the robot is provided with a telescopic elastic electromagnetic eddy current probe for detecting the thickness of the coating. The method has the advantages that the apparent diseases and the thicknesses of coatings of steel structures such as ceilings of high-speed rail stations and ceilings of large stadiums can be rapidly detected.

The invention relates to the technical field of general control or regulation systems, in particular to a critical patient respirator parameter adaptive optimization control method, which comprises the following steps of: acquiring and processing physiological signals; constructing a dynamic respiration mechanical model based on a sliding time window recursive recognition algorithm; a man-machine coordination index is calculated, and the desynchrony degree is quantified; generating a ventilation parameter adjustment scheme through the reinforcement learning strategy network; and executing smoothing parameter adjustment, and introducing a safety boundary and a secondary fine adjustment mechanism. The breathing state of the critical patient can be accurately perceived and intelligently regulated and controlled, the ventilation efficiency is guaranteed, meanwhile, the man-machine synchronism and safety are improved, the medical care burden is relieved, and the prognosis of the patient is improved.

The invention provides a metal imprinting character recognition method and system fused with multi-direction double-exposure image deep learning. The method comprises the following steps: step 1, collecting double-exposure images in four directions on the surface of a metal imprinting character; 2, fusing the double exposure images based on a differentiated weight distribution strategy of the regional information amount; step 3, constructing a lightweight improved EAST imprint character target detection model, and realizing detection of an imprint character region through a trunk feature extraction network, a self-adaptive channel attention mechanism, a fusion network and an output layer; and 4, character recognition: adopting a text recognition algorithm SVTR based on a single vision model, and completing the recognition of an embossed character sequence through convolution down-sampling, three-stage feature extraction and linear layer prediction. According to the technical scheme, high-precision and high-efficiency identification of the metal imprinting characters is realized by optimizing image acquisition, fusing a strategy and a deep learning model.

The invention provides an intelligent inspection system based on a robot and a camera, and relates to the field of intelligent manufacturing, the system adopts a four-layer architecture including a sensing layer, a transmission layer, a platform layer and an application layer, the sensing layer is composed of a fixed camera and an intelligent inspection robot; the transmission layer adopts optical fibers and 5G hybrid networking; the platform layer is integrated with a distributed server cluster, a data platform and an AI inference engine; the application layer provides a Web and mobile terminal management interface. The core of the system adopts a plurality of AI algorithms, including an instrument reading recognition algorithm based on Hough transform and CRNN, a leakage detection algorithm combined with background difference and an optical flow method, and a multi-device collaborative scheduling algorithm based on reinforcement learning, and constructs a digital twinborn inspection scene of virtual-real mapping. According to the invention, automatic and intelligent inspection and visual monitoring of power plant instruments, pipelines, power distribution rooms and other areas are realized, the inspection efficiency and accuracy are effectively improved, and the labor cost and the safety risk are reduced.

The invention discloses a food packaging bag directional conveying device and method, and relates to the technical field of automatic conveying and logistics equipment.The food packaging bag directional conveying device comprises a position and posture detection module, real-time position and posture data of a packaging bag at a key node are captured through a sensor array, the data are processed through an image recognition algorithm to extract steering angle features, and the steering angle features are obtained; the steering correction module is used for comparing the obtained current direction deviation value of the packaging bag with a preset standard angle, if the deviation value exceeds a threshold value, a steering correction mechanism is activated, and corrected packaging bag posture data are obtained by adjusting the angle of a mechanical arm at a node; according to the oriented conveying equipment and method for the food packaging bags, it is guaranteed that the packaging bags are efficiently and stably conveyed in a complex environment, and the reliability and adaptability of an automatic production line are improved.

The invention discloses a method and system for predicting the scouring trend of a full-hall long-piled wharf, and belongs to the field of port engineering monitoring. The method comprises the following steps: firstly, constructing an embedded ultrasonic array monitoring system, arranging probes according to regions to realize accurate deployment and control, and synchronously acquiring acoustic data and environmental parameters; noise reduction, inversion and fusion processing are carried out on the data, and normal deposition fluctuation and abnormal scouring events are distinguished by combining shallow water acoustic modeling and a scouring mechanism recognition algorithm; and finally, predicting the current siltation state through an improved DWMA model, predicting the scouring trend by means of an improved LSTM physical hybrid model, and realizing risk visualization based on a grading early warning mechanism. The system can adapt to the dense pile foundation environment of the full-space high-pile wharf, effectively solves the problems of large blind area and high error in traditional monitoring, remarkably improves the monitoring precision and early warning timeliness, and provides powerful support for safe operation and maintenance of the wharf.

The present disclosure relates to systems, non-transitory computer-readable media, and methods for generating a fraud score for survey response data and updating a dataset of responses of a digital survey. In particular, in one or more embodiments, the disclosed systems utilize a fraud indicator identifying algorithm to determine fraud indicators and generate a fraud score for the survey response data. In addition, in one or more embodiments, the disclosed systems utilize a fraudulent response identifying machine-learning model to generate a fraud score. The disclosed systems then utilize the fraud score to generate a label for survey response data and update a dataset of responses to a digital survey based on the label. In one or more embodiments, based on the disclosed systems generating a fraudulent label for the survey response data, the disclosed systems remove survey response data from the dataset.

The invention provides a transformer substation noise traceability and diffusion simulation method fusing multi-source remote sensing data. Belongs to the technical field of power environment protection and remote sensing science. The method comprises the following steps: acquiring thermal infrared remote sensing image data of a transformer substation, after preprocessing, finding out a temperature anomaly region of main noise source equipment by using a temperature anomaly recognition algorithm, determining the position of the main noise source equipment in combination with equipment layout information, and generating position data of the noise source equipment; a laser radar is used for scanning a transformer substation and surrounding areas to obtain high-precision point cloud data, the high-precision point cloud data are processed to generate high-precision digital elevation model data, and a transformer substation three-dimensional scene model containing detailed topographic features and noise source positions is constructed in combination with noise source equipment position data. Temperature abnormity of main noise source equipment is identified by means of thermal infrared images, noise sources of specific equipment such as a main transformer and an electric reactor can be accurately positioned, equipment-level traceability is realized, the situation that traceability is difficult in a traditional method is changed, and a clear direction is provided for targeted noise treatment.

The invention discloses a design method of an SCPI-based instrument automatic identification and unified control system, and the method comprises the steps: constructing a multi-interface instrument resource unified search framework through integrating a VISA library function and a multi-interface adaptive logic; and designing a multi-feature fusion instrument automatic identification algorithm, and reading instrument identification information through an SCPI standard command to realize accurate classification of general instruments. A layered resource management framework and a unified control interface are established, instrument core resource information is stored in a centralized mode, SCPI instruction differences of different manufacturers and different instruments are shielded, and recognition-management-control whole-process closed-loop control is achieved. Through deep fusion of a multi-interface instrument resource search framework, an instrument automatic identification and classification system, layered resource management and a unified control interface, automatic search and automatic identification of multi-model instruments and unified management and flexible switching of instrument resources are realized, and the compatibility and test efficiency of the test system to the multi-model instruments are improved.

The invention discloses a server cluster monitoring data processing method and device and a medium, and the method comprises the steps: scanning a target directory structure to analyze a directory name, and dynamically recognizing a server node address based on a predefined rule to generate server node list data; executing the data acquisition script on each server node to obtain original monitoring index data, and converting the original monitoring index data into a standardized monitoring data file; based on the historical processing record data and the characteristic value of the monitoring data file, whether the monitoring data file is a to-be-transmitted incremental file is judged through an incremental recognition algorithm, and if yes, the incremental file is transmitted to a central processing node based on a pre-constructed secure connection channel to execute transmission integrity verification, and transmission state data is generated; and performing standardization processing on the transmission state data, importing the processed to-be-imported data file into a target database, and comparing the number of records before and after importing to generate a data integrity report.

The invention relates to the technical field of engineering safety monitoring, in particular to an engineering safety monitoring system and method.The method comprises the steps that an intelligent monitoring terminal collects original data in a graded mode through a multi-mode sensing array, and core parameters are reserved preferentially after the original data are preprocessed through an edge computing unit; the dynamic data transmission module implements differential transmission according to data importance, high-priority parameters are transmitted through an encryption channel, and conventional parameters are transmitted through a universal Internet of Things channel; the self-adaptive data processing unit is internally provided with a dynamic threshold model fusing engineering mechanics characteristics and an anomaly recognition algorithm, and is used for processing transmission data to recognize anomaly; the grading early warning and linkage disposal module receives the abnormal result through a data interaction protocol, generates a corresponding disposal instruction and is in butt joint with an existing emergency system to achieve reuse. Therefore, the problems of low data processing efficiency, insufficient anomaly recognition accuracy, poor system suitability and the like in existing engineering monitoring are solved, and the reliability and response speed of engineering safety monitoring are improved.

The invention discloses a mechanical arm double-shaft-hole bolt inserting method based on three-dimensional visual positioning and admittance force control, and belongs to the technical field of mechanical arm intelligent assembly.The mechanical arm double-shaft-hole bolt inserting method comprises the steps that bolt inserting position estimation is conducted through an image recognition algorithm, and a shaft hole recognition strategy combining a YOLOv8 neural network model and Hough circle detection is adopted; extracting a shaft hole normal vector by using principal component analysis, and constructing a spatial constraint to solve a tail end target attitude; formulating a two-step visual guidance strategy to avoid an observation blind area, and controlling the six-degree-of-freedom mechanical arm to move to a bolt preparation point position in stages; and finally, switching an admittance force control mode to realize fine and smooth bolt operation. According to the method, high-robustness assembly of the metal shaft hole in the unstructured environment is realized, and the problem of positioning failure caused by specular reflection interference and a depth observation blind area is effectively solved.

The invention discloses a condom cleaning and drying system based on machine vision and a control method, and relates to the technical field of automatic control of condom production and processing equipment. Based on a preprocessed image, color features, texture features and morphological features are extracted through a multi-feature fusion stain recognition algorithm; inputting into a pre-trained stain level judgment model, and analyzing based on a stain coverage area calculation result to obtain a stain level; step 3, the main control module executes a targeted cleaning and drying control strategy according to the judged stain level in combination with the real-time operation parameters collected by the feedback regulation module; the condom image is collected and analyzed in real time through machine vision, the stain level is accurately judged in combination with a multi-feature fusion algorithm, targeted adjustment of a cleaning and drying strategy is achieved, and the condom cleaning cleanliness is effectively improved; and by dynamically optimizing control parameters, excessive cleaning and drying are avoided, and energy consumption is remarkably reduced.

The invention discloses a rock particle interface deformation identification method based on an image identification algorithm, and belongs to the field of rock particle interface deformation identification, and the method comprises the steps: obtaining a particle distribution range through a high-resolution image collection and multi-scale scanning technology, and when the scale span is detected to exceed a preset threshold value, determining the particle distribution range; and separating particle regions with different sizes by adopting a layered segmentation method. The texture mode of each layer is analyzed by using a convolutional neural network, a deformation feature vector is extracted, and algorithm granularity parameters are adjusted through iterative optimization so as to match the actual particle size. And when the matching degree is insufficient, a feedback loop mechanism is introduced to reprocess the image data and a noise filtering link is fused, so that the deformation identification precision is remarkably improved. Finally, an integrated calculation method is adopted to fuse multi-scale analysis results, accurate judgment of the deformation state of the whole particle interface is achieved, reliable comprehensive indexes are provided for rock stability evaluation, and the technical problem that a traditional method is insufficient in precision in cross-scale particle deformation recognition is solved.

The invention discloses a smart education guidance evaluation system and method based on big data, and belongs to the technical field of smart education and big data analysis. The system comprises a multi-source heterogeneous education data acquisition module, a data preprocessing and feature fusion module, a multi-dimensional capability evaluation module, a learning behavior sequence analysis module, a personalized guidance recommendation module and a multi-role visual display module. A subjective and objective fusion weighting algorithm and a double-flow attention knowledge tracking algorithm are adopted to realize five-dimensional capability comprehensive evaluation, a key behavior node identification algorithm is utilized to carry out learning behavior time sequence modeling, and a cognitive load perception path optimization algorithm is utilized to generate a personalized learning path. The problems that in an existing education evaluation system, the evaluation dimension is single, data collection fragmentation is achieved, and personalized guidance lacks data support are solved.

The invention discloses a contour and filling direction adaptive path planning and SLM support-free forming method, which has the following beneficial effects that a low-angle overhanging region is separated based on a region division algorithm, a scanning vector is controlled by a contour on the basis of contour features of complex structure features and an overhanging direction recognition algorithm, and the accuracy of the contour and filling direction adaptive path planning and SLM support-free forming method is improved. The expansion direction of the lower surface area is a filling direction, and adaptively generating a scanning path and the filling direction; different from a conventional path filling strategy, the method can meet the complex structure characteristics of a curved surface, a round hole, rotation, a special-shaped runner, a curved surface and the like, and the scanning line vector and the filling direction of a lower surface area adapt to the structure contour characteristics, so that the influence of the placing position and the scraper direction on the forming thermal strain accumulation and the lower surface forming quality is reduced; according to the method, the high-quality and stable forming requirements of the low-angle feature of 30 degrees or below of existing mainstream metal materials can be met, and the data processing time cost, the powder material consumption and the post-processing time in the whole forming process are effectively reduced.

The invention provides a multi-mode unmanned aerial vehicle countering method and system based on ground-air cooperation, relates to the technical field of unmanned aerial vehicle countering, and solves the problems that an existing unmanned aerial vehicle countering technology is low in cooperation efficiency, weak in environmental adaptability and insufficient in real-time response capability. In the method, a ground station uses a radar to scan a target unmanned aerial vehicle in a monitoring airspace, calculates basic parameters of the target unmanned aerial vehicle, constructs a threat degree evaluation model, obtains the threat degree of the target unmanned aerial vehicle, and uses a greedy algorithm to carry out target distribution between a countering unmanned aerial vehicle and the target unmanned aerial vehicle to form a distribution scheme; the distribution scheme is uploaded to the countering unmanned aerial vehicle, and a target unmanned aerial vehicle distributed to the countering unmanned aerial vehicle in a monitoring airspace is recognized through a built-in lightweight visual recognition algorithm; a target unmanned aerial vehicle is continuously tracked through a built-in visual tracking algorithm, and a corresponding countering means is dynamically selected based on recognition and tracking states. According to the invention, the countering response speed of the unmanned aerial vehicle and the stability of target tracking in a complex environment are smoothly improved.

The invention relates to the technical field of forensic examination, and discloses a forensic diatom automatic scanning and interpretation system based on optical microscopic imaging and AI, which comprises an automatic scanning imaging optical platform, an embedded high-performance all-in-one computer and an artificial intelligence recognition algorithm module. The all-in-one computer is configured to control the optical platform to execute XY-axis automatic scanning and Z-axis automatic focusing, and the automatic focusing is realized based on a definition evaluation function and a search algorithm to obtain a clear field-of-view image; and the artificial intelligence recognition algorithm module is configured to call a diatom target detection model and a diatom form classification model to carry out automatic detection and morphological classification on diatom targets in the clear view field image, and can also comprise an online learning and image library expansion mechanism. The problems that forensic diatom inspection depends on manpower, efficiency is low, interpretation is subjective, and an automatic system lacks accurate focusing and self-adaptive capacity are solved.

This invention relates to the field of fault location technology and discloses a fault location method for a long-distance conveying inspection robot. The method includes: when the inspection robot performing long-distance conveying is detected to have reached an entrance equipped with an RFID tag, reading the RFID tag to determine the section of the long-distance conveying route to which the inspection robot belongs, and calculating the distance between the inspection robot and the entrance; acquiring an image of the idler roller taken by a camera mounted on the inspection robot, and using an image recognition algorithm to identify the idler roller code; when a fault signal is received, concatenating the section of the long-distance conveying route to which the inspection robot belongs, the distance between the inspection robot and the entrance, and the idler roller code identification result to obtain the position of the inspection robot in the long-distance conveying, and generating an alarm message. This invention combines RFID and visual recognition to accurately locate the inspection robot's position, improving positioning accuracy and thus improving fault handling efficiency.

The application belongs to the technical field of intelligent security and protection home, and relates to a smart floor system capable of identifying falling behaviors, which comprises a smart pressure sensing floor, a data acquisition system and an upper computer, the smart pressure sensing floor is a composite structure of a solid floor and a flexible sensing fabric, the flexible sensing fabric contains 40 electrodes and 40*40 array distributed sensors, and a piezoresistive sensing layer is constructed by a graphene-carbon black mixed filler to form a conductive network; the data acquisition system collects signals through FPC flat cable, transmits data to the upper computer through a sensor router and a public network router; the upper computer adopts a "cloud-edge-end" architecture, carries a machine learning identification algorithm, can accurately identify falling and intrusion behaviors, the falling identification accuracy is more than 98%, and the response time is less than or equal to 1 second; the terminal side is associated with a cloud program to realize alarm pushing and data viewing functions. The disclosed system simultaneously integrates two core functions of old people falling monitoring and home safety protection, and meets the diversified needs of users living alone.

The application discloses an image processing method for displaying a back picture in real time, comprising the following steps: after a camera observes a human body, collecting and processing position information of a user relative to a display mirror, and outputting a distance of the human body from a screen; calling a camera to dynamically extract a back feature value of the human body by using the distance of the human body from the screen, fitting an eye height of the user, so as to determine a specific position of the camera; obtaining back image data of the user by using the camera, filling after human back contour recognition by using a human back recognition algorithm, obtaining processed picture information and transmitting the processed picture information to the display mirror, and displaying the processed back information picture on the screen. The application further discloses an image processing system for displaying a back picture in real time. The application solves the problem that a user is difficult to obtain a back picture in real time when facing the intelligent mirror group by using simple equipment combination and use, and has important production value and application space.

The invention provides an intelligent analysis method and device for population around a nuclear energy project factory site, and the method comprises the steps: constructing a factory site peripheral region image database through integrating a satellite map, GIS data and an unmanned aerial vehicle aerial image; automatically distinguishing buildings from non-buildings by using an image recognition algorithm, and intelligently classifying the buildings according to civil, office, commercial, industrial, cultural and other types; the population density or empirical parameters of various buildings are combined to calculate the total population number of single buildings and regions, and future population prediction based on the annual growth rate is supported. The system is further provided with a visual interaction interface, and parameter input, data storage, result query and report automatic generation are supported. According to the method, the automation, precision and high-efficiency level of demographic statistics around the nuclear power plant site is improved, the manpower and time cost is reduced, and scientific, dynamic and sharable data support is provided for nuclear power project site selection decision.

The application relates to the technical field of power detection, in particular to a tower partial discharge detection method and system based on an acoustic imager. The method comprises the following steps: acquiring tower acoustic data collected by the acoustic imager; analyzing the tower acoustic data to determine an acoustic characteristic map; analyzing the acoustic characteristic map based on a preset discharge identification algorithm to determine a partial discharge type; matching the partial discharge type with a preset detection mode to obtain a matching relationship; and adjusting a detection frequency band of the acoustic imager according to the matching relationship. The method provides multi-dimensional discrimination basis for partial discharge type identification, matches the partial discharge type with the preset detection mode, improves the synchronization rate of the detection time of intermittent discharge, reduces the missed detection rate, helps to eliminate positioning errors in a high-humidity environment, reduces the equivalent propagation distance calculation error, and meets the high-precision positioning demand in a complex environment.

The invention belongs to the technical field of e-commerce, and particularly discloses an enterprise e-commerce process customization and substituted operation service system. The system comprises a user intention real-time analysis module used for outputting a user intention quantization vector through a deep behavior pattern recognition algorithm; the strategy knowledge graph construction module is used for constructing and dynamically updating a multi-dimensional association network containing commodities, user tags and marketing actions; the dynamic strategy coupling engine is used for inputting the intention vector into a knowledge graph and generating a matched composite marketing strategy instruction set through graph neural network reasoning; the strategy instruction atomization execution unit is used for decomposing the instruction set and scheduling the instruction set into atomic operation instructions; and the full-link effect tracking and feedback module is used for collecting effect data and feeding back the effect data to the knowledge graph to optimize the weight. According to the system, end-to-end intelligent automatic closed loop from user intention recognition to marketing strategy generation and execution is realized, and the timeliness, accuracy and self-adaptive capability of marketing response are improved.

The invention discloses an intelligent detection method for a reservoir piping phenomenon, and the method comprises the following specific steps: firstly, collecting piping phenomenon pictures for different reservoir scenes, constructing a training data set, and training the training data set to obtain a piping recognition model; secondly, performing candidate area detection on a video to be detected by adopting a piping recognition algorithm to obtain a candidate area containing a potential water flow target; then, in each candidate area, a water flow velocity detection algorithm is used for obtaining a horizontal velocity characteristic value between continuous frames; extracting a texture contrast characteristic value based on a texture analysis technology; and finally, constructing a comprehensive scoring system through weighted fusion of the horizontal speed characteristic value and the texture contrast characteristic value of the water flow, and determining an effective piping detection result only when a fusion score exceeds a preset threshold value. According to the method, deep learning and image analysis technologies are fused, the limitation of single feature detection is effectively solved, and the reservoir piping phenomenon is accurately recognized.

The application discloses a Helicobacter pylori breath test signal denoising fitting method and system, and belongs to the technical field of medical detection and signal processing. The method processes the continuous monitoring signal of the 13CO2 / 12CO2 isotope ratio in exhalation: firstly, the original signal is subjected to band-pass digital filtering to suppress high-frequency noise; then, low-frequency drift is eliminated through adaptive baseline correction to generate a drift-removed signal; subsequently, a robust peak value recognition algorithm is used to extract physiological response characteristics; finally, the identified peak value is taken as the center, a nonlinear function with physiological constraints is used for local fitting, and the Delta Over Baseline value is calculated based on the generated fitting curve with high precision. The application effectively solves the signal distortion problem caused by breathing fluctuations, instrument noise and the like in in vivo detection, realizes high-fidelity restoration of the Helicobacter pylori metabolic response, and significantly improves the detection accuracy and reliability of the breath test.