74results about How to "Quick and accurate identification" patented technology

The invention discloses a high-precision industrial quality control system based on an AI large model and an application method, which are characterized by being applied to detection and measurement of industrial products. Comprising a product data acquisition module in an industrial production process, a data preprocessing module, an AI large model suitable for industrial quality control, a quality detection module, a result feedback and decision support module, a control module and a system management module, the application of the AI large model suitable for industrial quality control comprises the steps of constructing the AI large model suitable for industrial quality control, calling the AI large model suitable for industrial quality control, executing the control system, carrying out multi-modal data fusion, and carrying out multi-modal data fusion. The product data acquisition module in the industrial production process acquires data of multiple modes for quality detection and measurement, the AI large model suitable for industrial quality control adopts a data fusion algorithm, and the feedback module provides a visual detection result. And visually displaying the detected quality problem to a user in the form of an image and a chart.

The invention discloses a method and a system for improving Mura defect classification precision, electronic equipment and a storage medium, which are used for realizing accurate classification of Mura defects of low-gray-scale images. The method for improving the Mura defect classification precision comprises the steps of obtaining an image on a display screen to be detected; performing texture filtering processing on the image to obtain a filtered image; performing graying processing and defect segmentation processing on the filtered image to obtain a to-be-detected image; inputting the to-be-detected image into the Mura classification model to obtain a classification result; wherein the Mura classification model is obtained through the following training method: obtaining first data collected by a plurality of sensors; preprocessing the first data to obtain second data; extracting feature data in the second data; fusing the feature data according to an orthogonal sequence fusion algorithm to obtain a fusion result; and training according to a fusion result to obtain a Mura classification model.

The invention discloses an unmanned aerial vehicle visual guidance obstacle avoidance system and method based on polarization information, and relates to the technical field of unmanned aerial vehicle navigation and obstacle avoidance. Polarization image information in a flight scene is collected in real time through a polarization sensor, and the perception capability of a complex environment is enhanced. Polarization parameters are extracted through a compression observation method, the efficiency and accuracy of feature extraction are improved, a visual attention mechanism is combined, an obstacle background area is accurately recognized, a self-adaptive decision fusion model is constructed, the feature weight is optimized, and the reliability of target contour features is enhanced. Preliminary path search is carried out based on target contour features, and a centralized optimization sampling method is adopted, so that the probability of generating an optimal flight path in a complex environment is effectively improved. And the obstacle avoidance flight module is used for smoothing the flight path, so that the smoothness and controllability of the flight path are ensured, and the navigation and obstacle avoidance capabilities of the unmanned aerial vehicle in a complex environment are effectively improved.

The invention relates to the technical field of bus joint temperature measurement, and discloses a method for automatically monitoring the temperature of each phase busbar for infrared temperature measurement of a bus joint, and the method comprises the steps: collecting the surface temperature data and real-time images of each phase busbar of the bus joint in real time through installing an infrared temperature measurement sensor and image collection equipment. After preprocessing, temperature data is converted into an image form, and the image form and a bus bar real-time image are superposed to generate a fusion image, so that the actual form and the temperature distribution condition of the bus bar can be observed at the same time. Through presetting a high temperature early warning threshold value and a temperature gradient change threshold value, a temperature gradient direction is calculated and a change trend is analyzed so that prediction and early warning of temperature abnormity are realized. And when the temperature or temperature gradient change exceeds a preset threshold value, an early warning mechanism is triggered, an alarm signal is sent out, and an abnormal position and related information are displayed through a visual interface. According to the invention, the use safety and reliability of the bus duct are improved, and the fault risk caused by abnormal temperature is reduced.

The invention discloses a strong earthquake risk assessment method based on focus parameters, and relates to the technical field of earthquake risk analysis. The method comprises the following specific steps: acquiring real-time seismic waveform data received by a seismic station located in a strong earthquake risk area; constructing a seismic source mechanism inversion model based on the real-time seismic waveform data and the seismic phase file; calculating focus parameters in real time through the focus mechanism inversion model; and inputting the seismic source parameters into a trained seismic magnitude prediction model, and outputting seismic magnitude prediction information. The physical process of the seismic source and the propagation characteristics of seismic waves are analyzed through the method, the strong earthquake risk can be evaluated more scientifically and more accurately, an important basis is provided for disaster prevention and reduction, and the possible influence and damage degree of the earthquake are predicted.

The invention provides a PD-L1 immunohistochemical section tumor identification detection method and system, and the method comprises the steps: collecting various organ PD-L1 sections of a clinical pathological experiment, carrying out high-resolution scanning, pre-identification screening and cloud data storage, and obtaining a PD-L1 section data set of various organs; automatically labeling each tumor cell labeling pixel-level mask, automatically generating a labeling bounding box, and obtaining a tumor cell region labeling data set; according to the tumor cell region labeling data set, preprocessing all positions labeled as effective regions, creating a YOLO v8 deep learning model, training to gradually adjust network weight, and outputting a prediction result of the YOLO v8 deep learning model; and analyzing and calculating whether the prediction evaluation index information of the YOLO v8 deep learning model accords with a training target evaluation index, completing model training, and carrying out accurate positioning and pixel-level region identification detection on PD-L1 immunohistochemical pathological section tumor cells.

The invention provides a method for rapidly detecting the type of a UPS rear-end load and related equipment, which can rapidly and accurately identify the type of the UPS rear-end load. The method comprises the steps that a UPS output current waveform in a current period is acquired, whether the rear end of the UPS is connected with a target load or not is judged according to the UPS output current waveform, and the target load is a linear load (R load) or a non-linear load (RCD load); if the rear end of the UPS is connected with the target load, acquiring current instantaneous value data of a preset number of times based on a current waveform output by the UPS, and obtaining a current instantaneous value set; grouping the current instantaneous value set to obtain a grouped data set; calculating a slope between adjacent data points in each group in the grouped data set to obtain a slope set corresponding to each group; and determining the load type of the UPS rear end according to the slope set corresponding to each group.

The invention discloses a scene text recognition method and system for multi-granularity training based on global vectors, and relates to the technical field of scene text recognition. The method comprises the following steps: acquiring a to-be-recognized image, and cutting the to-be-recognized image to obtain a part containing a text; the scene text recognition model is used for carrying out scene recognition on the part containing the text to obtain a recognition result, and training is carried out on the scene text recognition model stage by stage and comprises the steps that training samples are preprocessed, the preprocessed samples are encoded through an encoder, global vectors containing multi-granularity information are extracted, and the global vectors containing the multi-granularity information are obtained; and decoding the global vector by using a decoder to obtain an identification result. According to the method, the information of different granularities of the image can be better utilized, the information of the image can be fully learned from easy to difficult, and a better effect can also be achieved for complex texts and curved texts.

The invention belongs to the technical field of fish genetic identification, discloses a multi-site large yellow croaker wild domestication population identification system based on an agarose gel electrophoresis technology, and particularly discloses a group of InDel molecular markers for large yellow croaker wild and domestication population identification, and the InDel molecular markers comprise InDel11-5053756, InDel20-14004876, InDel3-28345321 and InDel21-17327392. The InDel molecular marker detection primer is used for amplification to obtain a PCR product, agarose gel can be used for detecting strip polymorphism, extra restriction enzyme digestion or complicated polyacrylamide gel electrophoresis operation is not needed, and identification of large yellow croaker wild and domesticated populations can be rapidly and accurately achieved.

The invention relates to the technical field of semiconductors, and discloses a defect detection device and method for a semiconductor chip. The defect detection device for the semiconductor chip comprises an image acquisition module which is used for acquiring a high-definition image of the semiconductor chip and transmitting the high-definition image to a defect identification module; the defect identification module is used for receiving the high-definition image, identifying the position of a potential defect in the high-definition image by adopting an image identification technology and automatically marking in the high-definition image; according to the invention, by using the high-definition image acquisition module, a high-resolution chip image can be obtained, which provides conditions for identifying tiny or low-contrast defects. And the defect identification module automatically marks the defect position by using an image recognition technology, so that the accuracy and the automation level of detection are improved.

The invention relates to the technical field of pulse state recognition, in particular to an electric spark discharge pulse state recognition and control system and method.The electric spark discharge pulse state recognition and control method comprises the following steps that gap voltage is obtained, and whether the gap voltage is rising edge voltage or not is judged; comparing the gap voltage which is judged to be the rising edge voltage with a preset threshold voltage; when the gap voltage is the same as any one of the preset threshold voltages, a trigger state identifier is set to be 1, and counting is started from a preset value A1; when any one of counting stopping conditions is met, counting is stopped, the state identifier is reset, and the counting value is returned to the preset value A1; calculating delay time between two time points of counting starting and counting stopping; judging the delay time and outputting a judgment result; and performing control according to the judgment result. According to the invention, based on the pulse voltage threshold and the time characteristic, the short-circuit arc pulse and the arc discharge pulse are rapidly and accurately identified.

The invention discloses a focusing method and device and a storage medium, and relates to the technical field of video image focusing, and the method comprises the steps: determining a to-be-focused target block in a current frame of video data and a plurality of target sub-blocks in the target block, and obtaining a to-be-focused region set, each target sub-block being obtained through the division of the target block; an image original pixel of the target block is acquired, a phase difference value of each to-be-focused area in the to-be-focused area set is determined based on the image original pixel, and each to-be-focused area comprises the target block and each target sub-block; and determining a target focusing area according to each phase difference value, and focusing the target focusing area. According to the method, the target block needing to be focused is further subjected to region division, and the phase difference values of the divided regions are evaluated, so that a better region of interest of the target block can be quickly focused in a complex scene, and the focusing efficiency and the imaging quality of the automatic focusing system are improved.

The invention provides a fault positioning method and device for a box-type substation, and the method comprises the steps: obtaining a topological structure of the box-type substation, and recognizing all switch nodes and terminal equipment nodes; according to the current flow direction and the electric energy transmission path, establishing a directed graph model of the box-type substation; a sensor array is deployed on the switch node and the terminal equipment, electrical parameters of a deployment area are monitored in real time, fault detection and signal analysis are carried out according to a set threshold value, and a fault signal is determined; according to the current flow direction in the topological structure and a fault signal in the sensor data, analyzing the current flow direction at the fault moment, determining a monitoring area where the fault signal is located, and preliminarily positioning a fault area; and according to the change rate of the voltage and the current of the preliminarily positioned fault area, the distance from the fault point to the nearest sensor is calculated by combining cable parameters and transmission delay, the fault position is accurately positioned, and rapid and accurate positioning of the fault is realized.

The invention provides an energy-saving machine room environment data analysis method, a monitoring system and a storage medium. The method comprises the steps that energy-saving machine room environment data is collected, a cloud platform sets a threshold value, and a Z-Score algorithm is used for calculating the distance between a feature point and an average value to judge an abnormal value. And constructing a random decision tree to isolate abnormal values, setting an abnormal value standard to judge environmental data, and analyzing the environmental data to obtain three analysis results. Based on the three analysis results, abnormal value information is output if all the analysis results are abnormal, data alarm information is output if two analysis results are abnormal, and data deviation information is output if one analysis result is abnormal. And sending an analysis result to a machine room manager. According to the invention, environmental data is comprehensively analyzed, abnormal values are rapidly identified, real-time feedback is realized, rapid response of managers is ensured, the labor cost is reduced, and the intelligent level is improved.

The invention provides a host load anomaly detection method and system, electronic equipment and a storage medium, and relates to the technical field of network engineering. The method comprises the following steps: acquiring monitoring index data of a target host; performing type classification on the monitoring index data according to a fixed index, a non-black and non-white index and a dynamic change index, and constructing a data set under the corresponding classification; and inputting the monitoring index data classified into the dynamic change indexes into an anomaly detection model for detection so as to understand position information in a corresponding data sequence through position coding, learning representation of an input sequence by using a multi-head self-attention mechanism of a Transform encoder and a full-connection feedforward network, and outputting an anomaly judgment result through a linear layer and a softmax layer. According to the technical scheme, the host load abnormity can be quickly and accurately detected.

The embodiment of the invention discloses a jade category identification method and device, electronic equipment and a storage medium, and relates to the technical field of jade identification. The method comprises the following steps: acquiring a first jade image of target jade acquired by jade acquisition equipment, and compressing the first jade image to obtain a first compressed jade image; determining a reference jade area in the first compressed jade image, and determining whether the reference jade area is a fuzzy area; under the condition that the target jade region is determined to be a non-fuzzy region, determining a target jade region matched with the reference jade region in the first jade image; and determining the category of the target jade according to the target jade area. According to the scheme provided by the embodiment of the invention, the problems of low accuracy and low efficiency of a jade detection scheme can be solved, the category of the jade product can be quickly and accurately identified, and the authenticity of the jade product can be verified.

The invention relates to the technical field of building construction supervision, in particular to a safety early warning method and system for a concrete construction process, and the method comprises the steps: firstly obtaining data of a monitoring point of a support frame and an adjacent monitoring point, and constructing a deformation feature vector of the support frame; the support frame deformation feature vector comprises a distance feature, an angle feature, a deformation feature and a vibration frequency feature; then determining the number of clustering clusters according to a BIM model of a building, carrying out clustering analysis on the deformation feature vector of the support frame, and judging whether an abnormal cluster exists or not; support frame monitoring points in the abnormal cluster are obtained, and an abnormal support frame monitoring point set is obtained; and obtaining the distance feature, the angle feature, the pouring feature component and the vibration feature component of the abnormal support frame monitoring point to construct an abnormal evaluation index, and if the abnormal evaluation index is greater than a preset abnormal threshold value, triggering safety early warning. By means of the method, real-time monitoring of the formwork supporting frame in the concrete construction process is achieved, and construction safety is improved.

The embodiment of the invention provides a fabric weave structure intelligent detection method and a related device, and the method comprises the steps: obtaining a to-be-detected fabric image from a terminal, the terminal being a Web terminal or a mobile terminal, and the fabric image being obtained through the shooting of a portable microscope; preprocessing the fabric image to obtain a target fabric image; the target fabric image is input to a pre-trained deep learning model, a fabric weave structure result is output, the deep learning model is constructed based on a target detection algorithm and is suitable for weave structures of various woven fabrics, knitted fabrics and non-woven fabrics, and the fabric weave structure result is obtained. The deep learning model comprises a target detection layer, a feature extraction network and a classifier. On the basis of the method, the organization structures of various textile fabrics can be quickly and accurately recognized, the detection efficiency and accuracy are improved, the manual detection cost is saved, the labor intensity is reduced, and automatic and intelligent recognition of the fabric organization structures is achieved.

The invention relates to a circuit reverse device for three-dimensional simulation of an electric energy meter. The circuit reverse device comprises a circuit board placement platform, an image acquisition module and an industrial personal computer. The circuit board placing platform is used for placing a circuit board of a to-be-identified electric energy meter; the image acquisition module is arranged around the circuit board placing platform and is used for acquiring image information of the circuit board; the image information comprises circuit board color, circuit board texture, circuit board shape and circuit board local feature points; and the industrial personal computer is connected with the image acquisition module and is used for acquiring the image information of the circuit board and comparing the image information of the circuit board with image data in a preset electric energy meter image model library so as to determine an electrical connection relation result of the circuit board of the electric energy meter to be identified. By adopting the device, the elements on the circuit board and the electrical connection relation thereof can be quickly and accurately identified.

The utility model relates to the technical field of power case cabinet cable management, and discloses a power case cabinet with high versatility, which comprises a cabinet body and a cabinet panel, the outer side of the cabinet body is movably connected with the cabinet panel through a hinge, and when a cable management structure is in operation and use, cables needing to be arranged are folded, and then the cables are stored in the cabinet body. One group of silicon rubber cases is dragged, a cable is placed on the arrangement rope for winding, and after winding is completed, one group of silicon rubber cases is moved to the side of the other group of silicon rubber cases, so that the two groups of silicon rubber cases are attracted together through the two groups of fixed magnets and are hung on a hook composed of a fixed block and a limiting block, and a small waist shape is formed. The limiting block plays a role in preventing the auxiliary hole from being separated from the fixing block, and through the cable management structure, cables such as a power supply, a data line and the like in the cabinet can be arranged more clearly, neatly and orderly, so that daily management and maintenance are facilitated, entanglement and disorder of the cables can be reduced, and the risk of cable faults is reduced.

The present application relates to a system for profiling a tray, the system comprising: a turntable configured to rotate the tray; a light source positioned adjacent to the turntable, the light source configured to illuminate the tray when the tray is rotated on the turntable; a camera positioned relative to the light source, the camera configured to capture an image of the tray when the tray is rotated on the turntable; and a photographing compartment enclosing an area containing the turntable, the light source, and the camera, where the photographing compartment includes sides and a top to reduce an amount of ambient light entering the enclosed area.

The invention discloses an intelligent recyclable garbage sorting method based on AI identification, which comprises the following steps: S1, garbage is put into a sorting mechanism arranged in a garbage sorting device, and then the garbage on the sorting mechanism is shot through a camera; s2, the camera transmits a shot picture to an AI recognition module, and the AI recognition module recognizes a garbage type in the picture and transmits a recognition result to a Raspberry Pi control module; s3, the Raspberry Pi control module controls the sorting mechanism to throw the garbage into the classification garbage cans of the corresponding types according to the result transmitted by the AI recognition module, the type of the thrown garbage can can be automatically, rapidly and accurately recognized, the sorting mechanism is controlled to throw the garbage into the corresponding classification garbage cans, the process is highly intelligent, and the efficiency is high. The garbage sorting efficiency and accuracy are greatly improved, meanwhile, the structure is more simplified, the manufacturing cost is lower, the purchasing cost of a user is reduced, and the garbage sorting device is very suitable for being used in families.

The invention discloses a human body behavior recognition device and method based on an infrared image, and the method comprises the following steps: S1, collection of the infrared image: employing a single-chip microcomputer to drive an infrared sensor, enabling detected temperature data to form an 80 * 64 * n digital matrix, transmitting the 80 * 64 * n digital matrix to the single-chip microcomputer through an SPI, and storing the 80 * 64 * n digital matrix, then the temperature digital matrix is transmitted to a PC end in a UART communication mode; s2, preprocessing the infrared image; s3, making a human body behavior data set: designing and collecting infrared images of seven behavior actions of bending down, falling down, lying flat, standing, sitting still, standing-to-sitting and walking; s4, designing a behavior recognition algorithm; s5, behavior recognition algorithm training; and S6, identifying seven behavior actions. According to the method, the accuracy is higher, the recognition speed is higher, terminal deployment with limited computing resources is facilitated, and the actual application requirement is met.

The invention discloses a physical ability data acquisition method and system based on a BLE technology, relates to the technical field of intelligent data acquisition, and can acquire heart rate data of a monitored student object in real time through an intelligent wearable device, and transmit the heart rate data to a cloud platform by using the BLE technology. The heart rate data of the monitored student object can be subjected to time sequence feature analysis and multi-scale association by using a data processing and analysis algorithm based on artificial intelligence in a cloud platform; therefore, whether a heart rate abnormity early warning prompt is generated or not is intelligently judged according to the multi-scale heart rate mode characteristics of the heart rate data of the monitored student object. Through the mode, the real-time performance and accuracy of data can be ensured by utilizing the characteristics of low power consumption and efficient transmission of the BLE technology, and meanwhile, the system can quickly and accurately identify the abnormal heart rate mode and send out an early warning prompt when necessary, so that false alarms and missing alarms are reduced, and timely intervention and support are provided for health management of students.

The invention, which relates to the technical field of orchard picking management, discloses an orchard picking adaptive management system comprising an environment monitoring module, a data analysis and prediction module, an automatic control module, an intelligent picking module, a pest and disease monitoring module and a pest and disease control module. The environment monitoring module is used for collecting orchard environment data and transmitting the orchard environment data to the data analysis and prediction module, and the data analysis and prediction module is used for preparing and preprocessing the data, determining a model order and establishing, checking and optimizing an ARIMA model for yield prediction. According to the invention, through the integrated ARIMA model, the orchard picking adaptive management system can carry out accurate yield prediction based on the historical time sequence data of the orchard; the problems of inaccurate orchard yield prediction, complex data processing and model establishment, difficult model optimization and selection, lack of basis for picking plan making and low orchard management efficiency can be solved.

This invention discloses a leak detection device equipped with a leak detection block (750), comprising: a main fuel line connection hole (762a) penetrating the center of the body; a first annular spatial connection groove (782) formed at a predetermined depth from one side surface of the body, located radially spaced from the center of the main fuel line connection hole (762a); a second annular spatial connection groove (792) formed at a predetermined depth from the other side surface of the body, located radially spaced from the center of the main fuel line connection hole (762a); and a first separator insertion hole (810) drilled from the outer peripheral surface of the body, spanning the first annular spatial connection groove (782a). 2) The second annular space connecting groove (792) is connected to each other in a communicative manner, wherein when the separator (850) is inserted into the first separator insertion hole (810), the head (852) of the separator (850) closes the portion of the first separator insertion hole (810) between the first annular space connecting groove (782) and the second annular space connecting groove (792), thereby closing the communicative connection between the first annular space connecting groove (782) and the second annular space connecting groove (792), while communicatively connecting the sensor insertion hole (854) leading to the end surface of the head (852) to the first annular space connecting groove (782) or the second annular space connecting groove (792).

The invention belongs to the technical field of contract auditing, and discloses a contract comparison and recognition method based on deep learning, and the method comprises the specific steps: 1, importing a recognition system; step 2, training based on a deep learning OCR model; step 3, performing OCR operation on the picture; 4, identifying and uploading the content; 5, contract output or modification is judged; and step 6, risk prompting. By automatically comparing contract texts, the system can quickly identify key information, the complexity of manual operation is reduced, the system has good adaptability, contracts of different fields and types can be processed without manually adjusting parameters, the manual intervention requirement is greatly reduced due to the high automation degree, contract auditing is more efficient, and the contract auditing efficiency is improved. According to the method, multi-language texts can be easily processed, the automatic data extraction function provides great convenience for subsequent data analysis, and compared with manual text transcription, the OCR technology has better cost benefits in large-scale document processing.

The application provides a kind of intelligent construction site safety monitoring and abnormal behavior detection method, it is related to behavior identification technical field, including: based on video frame data, through target detection algorithm, personnel and cigarette in the construction site multi-modal data are positioned, to output first detection result;Based on the first detection result, video frame data and construction site multi-modal data in the preset time period, utilize 3D CNN to combine attention mechanism, to output second detection result;Based on the second detection result and construction area security level information, by isolated forest and self-encoder algorithm, get comprehensive abnormal score.

The present invention relates to a fiber for fabric authenticity identification and a fabric authenticity identification system using the same, the fiber for fabric authenticity identification comprising an infrared light-emitting body, the fiber being included in a yarn or a fabric, the fiber having a function of fabric authenticity identification by mixing a special pigment with the fiber, and the fabric authenticity identification system using the same. Therefore, the effect of authenticity identification or counterfeiting prevention can be achieved. For example, for a specific functional product, analysis after fabric preparation becomes complex and infeasible, but rapid and accurate identification can be achieved through the fiber for fabric authenticity identification of the present invention.

The invention discloses an official document automatic formatting method and system, which relates to the field of artificial intelligence technology, and comprises the following steps: inputting an official document for identification and typesetting; preprocessing the input official document data; identifying the elements of the input official document data; preprocessing the data in the training library and inputting it into the word segmentation model of HMM for word segmentation, and outputting the word segmentation label; converting the word segmentation label, character and word of each word segmentation in the training library into vectors and inputting them into a language big model, and iteratively training the vectors; inputting the BERT model output vectors into the language big model to obtain semantically complete word segmentation vectors, and segmenting official document paragraphs of different elements based on the semantically complete word segmentation vectors and appropriate segmentation rules; dividing the segmented official document data into lines according to the upper limit of the number of words in each line, and then obtaining the semantically complete official document typesetting, which is conducive to improving work efficiency and smooth government orders.