Patents
Literature
Patsnap Eureka AI that helps you search prior art, draft patents, and assess FTO risks, powered by patent and scientific literature data.

456results about How to "Guaranteed Sensitivity" patented technology

Photovoltaic power station cluster power prediction method based on space-time correlation

The invention discloses a photovoltaic power station cluster power prediction method based on spatial-temporal correlation, and the method comprises the following steps: carrying out the correlation analysis of meteorological features in numerical weather forecast and the output power of photovoltaic power stations, and determining feature variables affecting the prediction of photovoltaic power; defining each photovoltaic station in the photovoltaic power station as a node, calculating the similarity between the nodes based on the characteristic variables and the output power of the photovoltaic power station, and constructing an adjacent matrix; and inputting the adjacent matrix into an optimized GAT-Transform model to obtain a power prediction result of the photovoltaic power station cluster. And the GAT-Transform model comprises a GAT module and a Transform module, and the GAT-Transform model comprises a GAT module and a Transform module. According to the method, high-precision and high-efficiency prediction of the cluster power of the photovoltaic power station is realized through comprehensive processing of spatial-temporal characteristics and combination of an efficient model structure and an optimization algorithm, meanwhile, the flexibility, expandability and certain interpretability of the model are kept, and powerful support is provided for photovoltaic energy management and scheduling.
Owner:NORTHEAST DIANLI UNIVERSITY +1

Debris flow monitoring and early warning system based on Internet of Things

The invention relates to the technical field of geological disaster monitoring, and discloses a debris flow monitoring and early warning system based on Internet of Things, comprising a multi-source data acquisition module used for acquiring environmental data related to debris flow; the data transmission module is used for transmitting the environment data to a data center or a cloud end from a collection point; the data preprocessing module is used for carrying out abnormal value cleaning, time synchronization and normalization processing on the environment data; the deep learning analysis module is used for analyzing the preprocessed data so as to predict the occurrence risk of debris flow; and the dynamic early warning and publishing module is used for generating a multi-stage early warning signal according to an output result of the deep learning analysis module and publishing early warning information to a user in multiple modes. The accuracy, the real-time performance and the adaptability of the debris flow early warning system are improved through multi-source data fusion, edge calculation and a dynamic learning mechanism so as to cope with debris flow disasters in a complex geological environment.
Owner:GUIZHOU PROVINCIAL NONFERROUS METAL & NUCLEAR IND GEOLOGICAL EXPLORATION BUREAU

Multi-modal artificial intelligence quality defect prediction method

The invention relates to the technical field of artificial intelligence manufacturing, and particularly discloses a multi-modal artificial intelligence quality defect prediction method, which comprises the following steps of: acquiring real-time multi-modal data from multiple sources, including text data, audio data, image data and sensor data, and comparing the real-time multi-modal data with historical data; performing feature extraction based on the accurate data; constructing a quality defect prediction model; realizing efficient quality defect prediction by using a deep learning technology; calculating a deviation between a predicted value and an actual value through a cross entropy loss function; according to the method, the prediction value is divided into accurate prediction and inaccurate prediction according to the preset threshold value, an early warning mechanism is established on the basis of accurate prediction, and early warning is carried out on the multi-modal data of the potential defects, so that correction measures are taken in time, and the product defect rate is reduced.
Owner:QINGDAO E LINGNET INTEGRATION CO LTD +1

Conserved neutralizing epitope QT7 of the major glycoprotein GP5 of the North American porcine reproductive and respiratory syndrome virus envelope, nucleic acid molecules, expression vectors, neutralizing antibodies and their applications

ActiveCN118496323Bguaranteed specificityGuaranteed sensitivitySsRNA viruses positive-senseVirus peptidesAgglutinating antibodyTGE VACCINE
The present invention discloses a conserved neutralizing epitope QT7 of the main glycoprotein GP5 of the envelope of North American porcine reproductive and respiratory syndrome virus (PRRSV-2), a nucleic acid molecule, an expression vector, a neutralizing antibody and its application. The present invention inserts the neutralizing epitope QT7 into the Peg pili of Salmonella pullorum, introduces the inert carrier bacteria S9H to obtain a recombinant bacterium, and the surface of the recombinant bacterium can functionally express and display the neutralizing epitope QT7, mixes with the pig-derived immune serum on a glass plate, and determines the positive and negative nature of the reaction by observing whether agglutinated particles visible to the naked eye are produced within 2 minutes. The agglutination test can qualitatively and quantitatively determine the agglutination antibody, and the agglutination antibody level measured by the agglutination test directly mediated by the neutralizing epitope antigen QT7-antibody is strongly positively correlated with the PRRSV-2 neutralizing antibody level tested by the classical cell neutralization test, so that the present invention can be used to detect whether there is anti-PRRSV-2 neutralizing antibody and the neutralizing antibody level in the serum sample, and has potential application prospects in the immune effect of PRRSV-2 vaccine, the evaluation and prevention and control of anti-PRRSV-2 neutralizing antibodies.
Owner:YANGZHOU UNIV

Distribution network planning index prediction method based on artificial intelligence

The invention relates to the technical field of distribution network planning, and provides a distribution network planning index prediction method based on artificial intelligence, which combines three models to establish a multi-level model, and significantly improves the accuracy and efficiency of distribution network planning index prediction. The main trend and influence factors of distribution network planning are identified by using a multi-layer sensor (MLP), and the nonlinear modeling capability is provided. The time dependence and periodic change of data are captured through a time sequence model ARIMA, and the sensitivity to short-term fluctuation is ensured. A long short-term memory (LSTM) network captures a complex non-linear relationship and long-term dependence, and the prediction precision is further improved. And finally, through a Blending technology, weighted summation is carried out on prediction results of the models, so that maximum output of model advantages is realized, and the accuracy of a fusion result is optimized. Model parameters are dynamically adjusted through a real-time data flow mechanism and an adaptive learning algorithm. Therefore, the prediction is real-time, the method can adapt to continuously changing market and environment conditions, and high prediction precision is kept.
Owner:ZHAOTONG POWER SUPPLYING BUREAU OF YUNNAN POWER GRID

Hundred-kA impact large-current sensing device based on multi-composite magnetic core material

The invention belongs to the technical field of electrical measurement, and particularly discloses a hundred-kA impact large-current sensing device based on a multi-composite magnetic core material. The device comprises a multi-composite magnetic core framework, a current winding coil, a parallel resistor and a shielding shell, the multi-composite magnetic core framework is formed by compounding at least two magnetic materials, the multi-turn current winding coil is distributed on the multi-composite magnetic core framework, and the two ends of the parallel resistor are connected with the two ends of the current winding coil respectively; the shielding shell is located on the outer side of the composite magnetic core framework. A current conductor to be measured penetrates through the composite magnetic core framework in the axial direction, and the measurement output end of the current winding coil is connected with an output circuit. The technical problem that an existing impact large-current sensing device cannot meet the technical requirements of high-amplitude and high-accuracy impact large-current measurement is solved.
Owner:CHINA ELECTRIC POWER RES INST WUHAN BRANCH +2

Water leakage monitoring method for digital twinning system of coal mine shaft

A water leakage monitoring method for a digital twinning system of a coal mine shaft comprises the steps that an RFID antenna and a Twinin-Tag sensor are installed on the wall of the shaft, a monitoring terminal is arranged on the ground, and an RFID reader is arranged at an outlet of the shaft; a shaft monitoring signal is collected in real time through a Twindow-Tag sensor, the shaft monitoring signal is read through an RFID reader, and then shaft monitoring data are obtained by analyzing and processing the shaft monitoring signal; collecting RSSI reading and phase data by using a short time window, and performing state identification by using a decision tree; carrying out coarse segmentation and fine segmentation on the leakage signal, and determining the location of a water leakage point in combination with a change point detection algorithm; and the leakage volume and the leakage rate are estimated. According to the method, the water leakage point can be accurately positioned, the volume and flow velocity information of the leaked water in the shaft can be obtained, so that a reliable basis is provided for subsequent maintenance decision making, and the method is suitable for long-term monitoring operation of the coal mine shaft.
Owner:CHINA UNIV OF MINING & TECH +1

Decentralized federated learning method and system oriented to computing power network

The invention belongs to the technical field of federated learning, and particularly relates to a decentralized federated learning method and system oriented to a computing power network. The method comprises the steps that a model demand side in a computing power network uploads a federated learning task to a block chain network, screening computing power nodes according to a reputation score and a data score, and taking the screened computing power nodes as computing nodes to participate in training; the trained local model is verified through the block chain network and then is sent to a model demand side; and the model demander uses a weighted aggregation algorithm based on data fairness to aggregate the local models received in the current round and updates the global model by using an aggregation result. According to the method, the block chain is combined with federated learning under the background of the computing power network, and the universality of the model and the fairness of data are improved while the availability of the federated learning model and the node training enthusiasm are ensured.
Owner:SHANDONG COMP SCI CENTNAT SUPERCOMP CENT IN JINAN +1

A fluxgate current sensor and its use method

The present invention relates to the field of sensor technology, and in particular to a fluxgate current sensor and a method for using the same, comprising a base and a fluxgate sensor, wherein the fluxgate sensor is mounted on the top of the base, the inner cavity of the fluxgate sensor is filled with argon, and a positioning mechanism and a sealing monitoring mechanism are respectively mounted on the front and the top of the side wall of the fluxgate sensor. The present invention drives the worm to rotate by a knob, and the first turntable rotates clockwise under the cooperation of the worm and the worm gear teeth, and the guide groove squeezes the pin inward, prompting the slider to move inward, and the roller clamps the conductor to limit the position, and the worm can be braked by rotating the fastening nut to contact the base, thereby realizing the conductor limit, allowing the conductor to always be in the center of the fluxgate sensor, and reducing magnetic field distortion; achieving the purpose of fluxgate sealing monitoring, having a good heat dissipation effect, preventing the fluxgate sensor from reducing its responsiveness to magnetic field changes, ensuring sensitivity, and avoiding failures of the fluxgate sensor when working in extreme environments, and having high reliability.
Owner:SHANDONG HUANBANG ELECTRONIC TECH CO LTD

Smoke detection system and method

The invention relates to the field of machine vision and image recognition, and particularly discloses a smoke detection system, which comprises a data preprocessing module used for carrying out normalization, data enhancement and clipping and zooming on an input image to obtain preprocessed image information; the feature extraction module is used for performing feature extraction by using a convolutional neural network structure, generating multi-scale feature representation, obtaining a feature extraction image, and obtaining a multi-scale feature fusion image through a multi-scale feature fusion technology in combination with feature maps of different levels; and the CBAM attention mechanism module comprises a space attention sub-module and a channel attention sub-module, and the space attention sub-module is used for enhancing features of an important space region in the multi-scale feature fusion image according to a space attention mechanism. By adopting the technical scheme of the invention, real-time detection can be realized on resource-limited equipment in a lightweight and efficient manner, and the method has high recognition accuracy for small-target and dynamically changing smoke in a complex scene.
Owner:XINJIANG LEON TELECOM TECH

A method and system for automatically detecting surface defects of a flexible touch screen

The present invention discloses a method and system for automatically detecting surface defects of a flexible touch screen, comprising the following steps: S1: using the Retinex method to compensate for the illumination of the surface image of the flexible touch screen, and converting it to the HSV color space to obtain the converted image; S2: calculating the directional feature map based on the converted image; S3: constructing a Gaussian mixture model based on the directional feature map, and clustering the features using the expectation maximization algorithm to obtain the initial defect area; S4: optimizing the initial defect area using morphological processing to obtain the optimized defect area; S5: splicing the converted image and the optimized defect area to obtain an input image, and performing defect detection on the input image based on a multi-scale feature fusion network of deep learning. The present invention can realize fast and accurate detection of surface defects of a flexible touch screen.
Owner:LENGSHUIJIANG JINGKE ELECTRONIC TECH CO LTD

Diaphragm, preparation method thereof and battery

The invention provides a diaphragm, a preparation method thereof and a battery, the diaphragm comprises a porous film, a sensor is embedded in the porous film, and two sides of the porous film are respectively provided with a coating layer; the coating layer is one or both of an oxide and a polymer. The sensor is integrated into the battery diaphragm, so that the detection accuracy of the battery state is higher, the problem can be found at the initial stage of a battery fault, and a hysteresis phenomenon is avoided; the material / structure and the like of the battery cell are not influenced, and the battery cell is not damaged; irreversible damage to the sensor is avoided, and the sensitivity, the accuracy and the stability of the sensor are guaranteed.
Owner:SHENZHEN INST OF ADVANCED TECH CHINESE ACAD OF SCI

Minimum port interdigital sensor array based on screen printing and preparation method thereof

The invention discloses a minimum port interdigital sensor array based on screen printing and a preparation method thereof, and belongs to the field of flexible wearable electronics and micro-nano sensing devices. The preparation method of the interdigital sensor array comprises the following steps: S1, preparing an interdigital electrode layer; s2, coating a first insulating layer; s3, coating an electrode layer wire; s4, coating a second insulating layer; s5, coating a spacing layer; s6, preparing a sensitive layer; s7, packaging the sensor array; and aligning and bonding the electrode layer, the first insulating layer, the electrode layer wire, the second insulating layer and the spacer layer thin film prepared in the step S5 with the sensitive layer prepared in the step S6, and packaging to obtain the minimum port interdigital sensor array based on screen printing. The number of output ports of the interdigital sensor array can be reduced as much as possible, excellent flexibility is kept, the overall mechanical property is improved, and high sensitivity and high stability are achieved.
Owner:OCEAN UNIV OF CHINA +2

Construction method and system of prediction model, storage medium and equipment

The invention belongs to the technical field of natural language processing, and provides a prediction model construction method and system, a storage medium and equipment, and the method comprises the steps: carrying out the preprocessing of data at an input layer of a prediction model, carrying out the splicing processing of a target code and a target word vector, and obtaining a corresponding target comprehensive feature representation, taking the target comprehensive feature representation as the input of a prediction model network layer; a network layer of the prediction model is constructed, the network layer of the prediction model comprises three layers, the first layer and the third layer of the network layer of the prediction model are both double-layer bidirectional long-short-term memory networks, and the second layer of the network layer of the prediction model is a self-attention mechanism layer; and in an output layer of the prediction model, parameters of the prediction model are trained through a cross entropy loss function and cosine similarity, and a Nadam optimizer is used to adjust and optimize the parameters until the prediction model is obtained through adjustment and optimization.
Owner:CHINA THREE GORGES CORPORATION

Dry-type air-core reactor turn-to-turn short circuit fault detection method based on resonance wave peak average offset

The invention discloses a dry-type air-core reactor turn-to-turn short circuit fault detection method based on the average offset of resonance wave crests, and the method comprises the steps: constructing a circuit model of a dry-type air-core reactor to simulate normal operation and turn-to-turn short circuits at different fault degrees and fault positions, and obtaining a corresponding impedance logarithmic amplitude spectrum; respectively calculating the resonance wave crest offset and the characteristic value of the impedance logarithmic amplitude spectrum corresponding to each turn-to-turn short circuit; obtaining an impedance logarithmic amplitude spectrum of the dry-type air-core reactor to be detected, and calculating a resonance wave crest offset and a characteristic value; and if the resonance wave crest offset is greater than the preset threshold value, matching the resonance wave crest offset and the characteristic value with the resonance wave crest offset and the characteristic value of the impedance logarithmic amplitude spectrum corresponding to each turn-to-turn short circuit, and determining the turn-to-turn short circuit fault position and fault degree of the to-be-detected dry-type air-core reactor according to the matching result. According to the invention, the accuracy of turn-to-turn short circuit fault detection of the dry-type reactor is improved, and the safe operation of the dry-type air-core reactor is ensured.
Owner:STATE GRID HUNAN ELECTRIC POWER COMPANY LIMITED +2

Grounding device of flexible low-frequency power transmission system and type selection method thereof

The invention discloses a grounding device of a flexible low-frequency power transmission system and a model selection method of the grounding device. The grounding device comprises an alternating current converter used for realizing conversion between a first frequency and a second frequency; two sides of the first frequency transformer are respectively connected with the alternating current converter and a first frequency alternating current power grid; the second frequency transformer comprises a second frequency grounding device and is used for connecting the alternating current converter and a second frequency alternating current power grid; the second frequency grounding device comprises a grounding resistor, one end of the grounding resistor is connected to the neutral point of the second frequency transformer, and the other end of the grounding resistor is grounded to provide a zero potential reference point for the alternating current converter; and the zero-sequence current collection winding is used as a third winding of the second frequency transformer and provides a main path of the zero-sequence current. Through cooperation of the grounding resistor and the zero-sequence current collection winding impedance, zero-sequence currents on the two sides of the second frequency transformer do not cross each other or are maintained at an allowable zero-sequence current low cross-over level.
Owner:ELECTRIC POWER RES INST OF STATE GRID ZHEJIANG ELECTRIC POWER COMAPNY +1

IGH gene clone type detection multiplex PCR primer group, design method and application

The invention discloses a multiplex PCR (Polymerase Chain Reaction) primer group for IGH (Insulin Growth Hormone) gene clone type detection, a design method and application. According to the primer group design method, an IGH candidate primer database is established by setting a weighted scoring mechanism, primers are screened by setting a score penalty mechanism, and when a new clone type or an amplification reagent changes, the primers can be rapidly replaced; the length range of an amplification product corresponding to the F-end primer and the length range of an amplification product corresponding to the R-end primer are strictly controlled, so that various clone-type sequences can be well amplified, sequenced and detected, and the requirement of clinically monitoring whether tumors relapse or not according to clone-type characteristics can be met.
Owner:SHANGHAI CINOPATH MEDICAL TESTING CO LTD

Application-level medical large model training method, application method, equipment, medium and product

The invention discloses an application-level medical large model training and application method, device, medium and product, and relates to the technical field of artificial intelligence, the method comprises the following steps: segmenting a medical scene to obtain a plurality of small worlds; for each small world, acquiring doctor-patient dialogue data in the small world; performing anonymization processing on the doctor-patient dialogue data to obtain anonymized data; refining the anonymized data to obtain training data suitable for the base large language model; and training the base large language model by using the training data to obtain an application-level medical large model suitable for the small world. According to the method, a data source of a universal base large language model is divided into a plurality of small worlds, panoramic full-amount data acquisition is performed in a specific small world, a small world enhancement model is established, finite rules and possibilities under specific time and space are learned, and compared with the universal base large language model, the method has the advantages that the method is simple and convenient to implement. And the specific performance aiming at the clinical environment is higher.
Owner:INSTITUTE OF BASIC MEDICAL SCIENCES CHINESE ACADEMY OF MEDICAL SCIENCES

Immunodetection method based on holographic imaging mode conversion and deep learning

The invention discloses an immunodetection method based on holographic imaging mode conversion and deep learning, and belongs to the field of immunoassay detection. The deep learning holographic mode conversion system constructed based on the non-paired data set reduces the influence of diffraction interference on target quantification to the greatest extent, so that holographic super-resolution imaging is realized, and challenges in paired data set generation are overcome; besides, a functional glass slide is designed, and rapid and sensitive holographic imaging detection is realized in combination with a click reaction signal amplification strategy; the functionalized glass slide is combined with the portable holographic imaging equipment, so that the detection efficiency of the system is greatly improved while accurate signal amplification is carried out, the application scene is expanded, and antibiotic residues can be rapidly and sensitively detected. The invention opens up a new research direction for high-sensitivity on-site convenient detection of trace target objects in portable equipment and complex samples.
Owner:DALIAN POLYTECHNIC UNIVERSITY +1

Keyboard assembling jig

The invention discloses a keyboard assembling jig, and relates to the technical field of jigs, the keyboard assembling jig comprises a jig seat, a keyboard assembling main body is placed at the top end of the jig seat, a clamping mechanism is arranged between the outer wall of the keyboard assembling main body and the keyboard assembling main body, a plurality of assembling keycaps are arranged at the top of the keyboard assembling main body, and a mounting plate is arranged at the top of the jig seat; an adsorption mechanism is arranged between the top end of the mounting plate and the multiple assembling keycaps, a pair of side connecting plates is fixedly connected to the two ends of the jig seat, and a movable cleaning mechanism is arranged between the outer walls of the side connecting plates and the top of the keyboard assembling body. The movable cleaning mechanism and the vibration dust falling mechanism are matched with the pair of scrap collecting half discs, dust adhering to the keyboard assembly main body and the multiple assembly keycaps in the conveying process can be cleaned, the keyboard assembly jig has the cleaning function, the functionality of the keyboard assembly jig is improved, and therefore in the assembly process, interference of dust impurities is reduced, and the assembly efficiency of the keyboard assembly jig is improved. The keyboard use sensitivity is ensured.
Owner:彭妍霓

Method for determining high-performance non-preheating welding process of maritime work steel

The invention belongs to the technical field of welding, and discloses and provides a maritime work steel high-performance non-preheating welding process determination method which comprises the following steps: acquiring a first cold crack sensitivity index required by non-preheating welding at a specific temperature and a critical constraint distance and a critical constraint degree of a base metal sample; obtaining the minimum diffusible hydrogen content of a welding material required in non-preheating welding, and selecting an actual welding material based on the minimum diffusible hydrogen content; the cold crack sensitivity coefficient of the base metal sample is used for optimizing the carbon equivalent of the base metal sample to obtain a new base metal; a second cold crack sensitive index smaller than the first cold crack sensitive index is obtained, and the new base metal corresponding to the second cold crack sensitive index serves as actual base metal; and the time t100 when the peak temperature of the actual welding material and the actual base metal after welding is cooled to 100 DEG C under different welding heat inputs and the critical cooling time tcr of the welding joint are obtained, and the welding heat input when the difference value between t100 and tcr is the maximum serves as the actual welding heat input. A quantitative analysis theory system is provided for developing the welding process, orthogonal tests are avoided, and the method is high in universality.
Owner:TIANJIN UNIV

Nucleic acid gel dye based on aggregation-induced emission material and application thereof

The invention discloses nucleic acid gel dye based on an aggregation-induced emission material and application of the nucleic acid gel dye. The nucleic acid gel dye based on the aggregation-induced emission material comprises AIE molecules of any one structure in a formula (1) and a formula (2). The nucleic acid gel dye based on the aggregation-induced emission material is used for common gel imaging dyeing of nucleic acid electrophoresis, can be simultaneously suitable for agarose gel electrophoresis and polyacrylamide gel electrophoresis, and is especially suitable for blue light imaging. The nucleic acid gel dye based on the aggregation-induced emission material is good in dyeing effect, high in sensitivity and short in dyeing time, adapts to all existing dye excitation channels, and becomes an ideal choice matched with a laser scanner for use. The defect that the sensitivity is low when the existing nucleic acid dye is applied to polyacrylamide gel electrophoresis due to poor dyeing effect is effectively overcome, and particularly the problems that the existing dye excitation channel is relatively single and the like are solved.
Owner:AIE INSTITUTE

Muscle elasticity modulus detection system and method based on flexible wearable sensor

The invention provides a muscle elasticity modulus detection system and method based on a flexible wearable sensor. The muscle elasticity modulus detection system comprises a flexible wearable ultrasonic sensor array, an ultrasonic pulse generation receiver and a data acquisition system / PC terminal, receiving a pulse signal generated by the ultrasonic pulse generation receiver through the flexible wearable ultrasonic sensor, converting the pulse signal into an ultrasonic signal, transmitting the ultrasonic signal to a detected object, and feeding back an ultrasonic echo signal to the ultrasonic pulse generation receiver; and the ultrasonic pulse generation receiver sends the ultrasonic echo signal to the data acquisition system / PC terminal for analysis and processing to obtain the corresponding muscle elasticity modulus. The flexible wearable ultrasonic sensor has certain flexibility, and the sensors are connected through a snakelike circuit, so that the stretchable and bending performance of the flexible wearable ultrasonic sensor array is ensured, the sensor array is more closely connected with the human body, and the wearable purpose is achieved.
Owner:SHANGHAI UNIV

High performance liquid detection method for letemovir content

The invention discloses a high performance liquid detection method for letemovir content, and belongs to the technical field of pharmaceutical chemicals. The method comprises the following steps: (1) preparing a sample; (2) detection: carrying out isocratic elution by adopting a chromatographic column taking octadecyl bonded silica gel as a filler, taking an ammonium dihydrogen phosphate solution with the concentration of 20mmol / L as a mobile phase A and taking acetonitrile-methanol (80: 20) as a mobile phase B; and (3) analysis: recording a chromatogram, and calculating the content of the letemovir component by a peak area according to an external standard method. According to the method, an isocratic elution detection method is selected, a base line is smooth, the requirement on an instrument is low, the detection time is short, the detection efficiency can be improved, and consumables and reagents are saved.
Owner:SHANDONG JIKUN BIO-PHARM CO LTD

Open domain image recognition method based on combinable text prompt framework

The invention discloses an open domain image recognition method based on a combinable text prompt framework, and the method comprises the steps: 1, carrying out the preprocessing and tokenization of an image x, and obtaining a tokenization feature Fv through a clip visual model; 2, initializing a task related text prompt ptt and a visual related text prompt pvt; step 3, pvt learns a prompt # imgabs0 # after visual fusion by using a cross-modal attention mechanism, step 4, ptt and # imgabs1 # adopt a self-attention mechanism to fuse a learning combination prompt # imgabs2 #, step 5, a clip text model is adopted to obtain # imgabs3 # features, a loss optimization prompt and a prompt combination module are calculated, the method learns visual related prompts and visual unrelated prompts respectively, and the visual related prompts, the visual unrelated prompts, the visual unrelated prompts, the visual unrelated prompts and the visual unrelated prompts are subjected to visual fusion. The visual related prompt captures visual information by using cross-modal attention, and the visual irrelevant prompt realizes mutual optimization through combination of self-attention and the visual related prompt.
Owner:BEIJING NORMAL UNIVERSITY

Power transmission line disconnection fault protection method, device and equipment and storage medium

The invention relates to a power transmission line disconnection fault protection method, device and equipment and a storage medium, and the method comprises the steps: collecting a phase voltage value, a zero sequence voltage value and two groups of phase current values of a power transmission line in real time, carrying out the real-time monitoring of a circuit breaker, and obtaining a position state signal; according to the phase current value and the position state signal, whether the current transformer has a disconnection fault is judged in real time, and when the current transformer has the disconnection fault, a CT disconnection signal is output; obtaining a current break variable of the transmission line, calculating a self-produced zero-sequence current according to the phase current value, judging whether the self-produced zero-sequence current, the zero-sequence voltage value or the current break variable meets a preset condition or not, and if yes, outputting a disconnection protection starting signal; and after disconnection protection is started, judging whether the power transmission line has a disconnection fault or not according to the phase voltage value, the phase current value, the position state signal and the CT disconnection signal, and if the judgment result is yes, outputting a circuit breaker tripping outlet and a starting failure outlet.
Owner:HUADIAN ELECTRIC POWER SCI INST CO LTD

Face recognition method and device based on adaptive learning, equipment and medium

The embodiment of the invention provides a face recognition method and device based on adaptive learning, equipment and a medium, and belongs to the technical field of financial science and technology. The method comprises the following steps: acquiring a current face image of a target user; comparing the current face feature vector of the target user with a historical face feature vector of the target user, and determining a feature vector variable quantity; adjusting a first target model parameter of the face recognition model through an incremental learning module under the condition that the feature vector variable quantity exceeds a first preset change threshold value; under the condition that the feature vector variable quantity exceeds a second preset change threshold value, adjusting a second target model parameter of the face recognition model through an online learning module; and optimizing the face recognition model according to the adjusted first target model parameter or the second target model parameter. According to the embodiment of the invention, self-adaptive learning can be carried out on the face change of the user, and the long-time accuracy of face recognition can be ensured in a financial scene.
Owner:PING AN BANK CO LTD

Cycloid hydraulic motor brake device mounting structure and brake device

The utility model discloses a cycloid hydraulic motor braking device installation structure and braking device, the installation structure comprises a shell of the braking device and a braking piston, the inner wall of the shell is expanded in sequence to form a first annular step and a second annular step which are used for abutting against a friction pair and the braking piston, the braking piston is in a step shape, and the braking piston is connected with the first annular step and the second annular step. The diameter of the head of the brake piston is expanded, and the head of the brake piston and the second annular step define a containing cavity space used for containing hydraulic oil. The head structure of the brake piston is changed, the brake piston and the shell are directly used for forming the containing cavity space used for braking or releasing braking, the installation precision is optimized, the number of parts is reduced, maintenance is easy, and the situation that the brake sensitivity is reduced due to long-time use is completely avoided.
Owner:JIANG SU GUO RUI JI XIE ZHI ZAO YOU XIAN GONG SI

Pressure sensor and preparation method thereof

The invention provides a pressure sensor and a preparation method thereof, the pressure sensor comprises a first wafer, a top semiconductor layer, a lead structure, a lead hole and a conductive column, the upper surface layer of the first wafer is provided with an absolute pressure cavity and an isolation groove; the top semiconductor layer is stacked on the upper surface of the first wafer, the lower surface layer of the top semiconductor layer is provided with a sensor functional area, the edge area of the sensor functional area is provided with four varistors which are uniformly and symmetrically distributed, each varistor comprises at least one resistance area, and the edge and the peripheral area of the sensor functional area are provided with a plurality of connecting areas connected with the two ends of each varistor; the plurality of lead structures are located on the lower surface of the top semiconductor layer and located in the isolation groove, and one end of each lead structure is electrically connected with the connection area; the lead hole is opened from the bottom surface of the first wafer and the lead structure is exposed from the bottom surface; the conductive pillar fills the lead hole. The sensor structure is improved, and the functional area of the sensor covers the opening of the absolute pressure cavity, so that the device is prevented from being influenced by the external environment, and the sensitivity of the device is ensured.
Owner:CRM ICBG (WUXI) CO LTD

Intelligent follow-up visit management system for skin disease patients

The invention discloses an intelligent follow-up visit management system for skin disease patients, and relates to the technical field of Internet. During operation of the system, multi-source data is acquired from mobile equipment, medical equipment and environment acquisition equipment worn by the patients, a comprehensive original data set D covering illness conditions and environment factors is formed, summarizing calculation is performed, and a follow-up visit result is obtained; according to the method, the dynamic follow-up scoring index DP is generated, so that the follow-up frequency adjustment identifier SF is generated, accurate regulation and control of the follow-up frequency are realized, and the defect that a fixed follow-up period cannot respond to the change of an illness state in real time is overcome. The evaluation feedback and adjustment module calculates a follow-up visit frequency adjustment identifier SF for the second time based on follow-up visit data fed back by the patient, and substitutes the follow-up visit frequency adjustment identifier SF into the data processing and feature extraction module for iterative optimization; the system can continuously adapt to dynamic changes of illness conditions, and precision and individuation of follow-up visit plans are achieved.
Owner:SINGLE INFORMATION TECHNOLOGY GROUP CO LTD