1098 results about "Rapid identification" patented technology

The invention discloses a wide-area landslide rapid identification method based on an interpretable intelligent algorithm, and relates to the field of remote sensing science and technology, and the method comprises the steps: building a dual-channel feature extraction architecture through multi-source spatio-temporal data fusion and knowledge graph dynamic weighting: capturing image local textures through a lightweight CNN, modeling geological spatial correlation through a graph convolutional network, and carrying out the recognition of the landslide. Combining the SHAP value and causal reasoning to generate an interpretable contribution degree thermodynamic diagram and a rule chain; a knowledge graph bidirectional verification system is introduced, spatial logic contradictions are verified by using prior rules, and co-evolution of a model and a rule base is triggered based on misjudgment samples; outputting a multi-dimensional credibility report, quantifying uncertainty by Monte Carlo Dropout, and customizing interpretation granularity according to roles; a terrain-adaptive block-stream processing architecture is adopted, and edge lightweight deployment and federated learning are combined, so that wide-area real-time early warning and model dynamic updating are realized. According to the scheme, the limitation of a traditional black box model is broken through, and a disaster prevention closed loop with physical driving, transparent decision and second-level response is formed.

The invention discloses a typhoon long-distance rainstorm rapid identification method based on an AI meteorological model, a storage medium and a computer program product, and the method comprises the steps: obtaining historical weather data and / or forecast weather data, selecting an AI weather forecast model which is based on a deep learning technology and carries out the training through a large amount of historical weather data, representative and typical rainfall examples are screened, a control experiment and a sensitivity experiment are designed, and typhoon core vortexes are removed from an initial field. By comparing the difference between the results of the control experiment and the sensitivity experiment, the influence of the typhoon on the long-distance rainfall is quantified based on the whole-layer water vapor divergence change and rainfall change analysis, and the long-distance rainstorm influence range and intensity of the typhoon are identified and determined. According to the method, the precision of medium-term weather forecast and extreme weather event recognition is remarkably improved, efficient calculation and wide practical application value are achieved, a scientific basis is provided for disaster early warning and disaster reduction management, and a new method is provided for meteorological scientific research and teaching.

The invention discloses a stone cultural relic disease identification method based on a hyperspectral image, and aims to solve the problem that stone cultural relic disease investigation depends on artificial experience and is time-consuming and labor-consuming. The method comprises the steps that a hyperspectral image of a stone cultural relic to be explored is collected and preprocessed; extracting spectral features; selecting a central wave band of three wave bands of blue light, green light and red light to construct a true color image; extracting texture features and color features of the true color image; fusing the spectral features, the texture features and the color features to obtain combined features; establishing a typical disease feature database; training a disease classification model by using a random forest, a support vector machine, a gradient elevator and a convolutional neural network; carrying out typical disease identification by using the intelligent identification model and drawing a disease cloud picture; and calculating the typical disease area proportion of the to-be-detected area to carry out disease evaluation. According to the invention, in-situ, lossless and rapid identification of stone cultural relic diseases is realized, manpower and material resources are saved, and an intelligent identification method is provided for cultural relic protection exploration work.

The invention discloses an SSR (Simple Sequence Repeat) core primer combination and a kit for identifying a tea tree variety 'swan gooses 12' and application of the SSR core primer combination and the kit. According to the invention, three pairs of SSR core primers are screened by using an SSR fluorescence labeling detection technology, and the three pairs of SSR core primers can rapidly identify and detect the 'Yangou 12' from other tea tree varieties. The SSR core primer combination disclosed by the invention can be used for identifying the variety of the swan goose 12, the uncertainty of identification based on external morphological characteristics is overcome, the result is reliable and visual, the detection efficiency is high, the operation is simple and convenient, and meanwhile, the popularization, utilization and protection of the variety of the swan goose 12 are also facilitated.

The invention provides a primer group, a kit and an identification method for rapidly identifying genetic sex of siniperca chuatsi and application, and relates to the technical field of molecular biology. The primer group comprises an upstream primer and a downstream primer, the nucleotide sequence of the upstream primer is shown as SEQ ID NO. 1, and the nucleotide sequence of the downstream primer is shown as SEQ ID NO. 2. The primer group can accurately distinguish the male and female sex of siniperca chuatsi at the genome level, so that the physiological genetic sex of sex reversal individuals is rapidly identified. The establishment of the method provides an important scientific basis for the development of a subsequent siniperca chuatsi unisexual breeding technology and the research of a sex determination mechanism.

The invention discloses a rapid marine ship target identification method based on YOLOv8, and is applied to the technical field of target identification. Comprising the following steps of collecting an image sample of a ship navigation water area in a digital vehicle, cleaning and marking the image sample, performing data enhancement on the image sample to obtain a data set which can be used for training and verification, constructing a target identification network model based on a YOLOv8 deep neural network, and obtaining a target identification result by setting hyper-parameters such as a learning rate and a batch processing size. And carrying out transfer learning by using the pre-training model, removing a repeated detection frame by using a non-maximum suppression algorithm and carrying out result filtering by using a confidence coefficient threshold to carry out post-processing on a reasoning result of the model so as to complete prediction of the ship target, and carrying out sparse pruning training on the trained model neural network so as to obtain the ship target. And carrying out calibration and quantification on the pruned model. According to the method, on the premise that the model detection precision is reserved to the maximum, memory consumption during equipment operation is reduced, and meanwhile the model detection speed is increased.

The invention discloses a system and method for visually detecting the quality of a pipe pile. The system comprises a bearing module used for bearing the pipe pile, an end detection module and a movable pipe pile body scanning module. The end detection module is used for collecting end face data of the end of the pipe pile and recognizing the defects of the end of the pipe pile according to the end face data of the end of the pipe pile. The movable pipe pile body scanning module moves along a pipe pile and is used for collecting pipe pile body data, and pipe pile body defects are recognized through the pipe pile body data. And summarizing the pipe pile end defects and the pipe pile body defects to generate a defect report. Rapid identification, quantitative analysis and data archiving of pipe pile defects are achieved, and the detection precision and the production efficiency are remarkably improved.

The invention relates to the technical field of environment monitoring, in particular to an intelligent water quality monitoring method and system for collaborative sampling of a bionic fish school and an unmanned aerial vehicle and computer equipment. The method comprises the following steps: dividing monitoring grid units by using GIS data, establishing a layered communication architecture, dynamically deploying a bionic fish school to carry out real-time water quality monitoring, and scheduling and optimizing resource allocation through a fuzzy rule; historical pollution event data and real-time sensing data are integrated to construct a water quality evaluation index system, and a Nash equilibrium strategy is adopted to determine an optimal weight to generate a dynamic pollution situation cloud chart; constructing an inferior fingerprint database to realize rapid identification of pollution events and trigger a two-stage response mechanism; and finally, combining the bionic fish school sampling data and the unmanned aerial vehicle pre-diffusion coordinate to dynamically simulate the effect of the treatment measure. Air-water integrated efficient monitoring is achieved, the accuracy and timeliness of water quality anomaly detection are improved, and comprehensive technical support is provided for water area environment protection.

The invention discloses a rapid identification method applied to defibrillation requirements of an AED device, and belongs to the technical field of medical equipment. According to the method, multiple links including acquisition, preprocessing, feature extraction, mode classification, real-time classification and defibrillation triggering work cooperatively, and advanced signal classification and mode recognition technologies are introduced, so that efficient and accurate electrocardiosignal analysis is realized in the AED device, the response speed and accuracy in the emergency treatment process are remarkably improved, and the emergency treatment efficiency is improved. Multi-dimensional features of electrocardiosignals are extracted and input into a pattern classification model, automatic recognition of arrhythmia is achieved, whether defibrillation operation needs to be carried out or not can be accurately judged, the possibility of misjudgment is reduced, self-optimization can be carried out according to new electrocardiosignal data through a dynamic learning and retraining mechanism, and the accuracy of defibrillation is improved. The adaptive capacity to a complex heart rhythm mode is improved, the first-aid response time is effectively shortened, the safety of a patient is guaranteed, and the first-aid efficiency and the treatment effect are improved.

The invention relates to the technical field of mark point identification and positioning, in particular to a polygon mark point rapid identification and positioning method and system and a storage medium. Then, the gradient direction information and the coordinate information of the reserved strong edge are used for rapidly estimating the rotation angle and the center of the polygon mark point to be detected, the method is small in calculation amount and easy to operate, and rough pose information of the polygon mark point to be detected can be rapidly obtained; the estimated angle is further corrected in a small range in a rotary projection mode, the deviation between an estimated value and a real value is reduced, and a more accurate initial value is provided for subsequent registration; finally, the corrected coarse positioning result is used as an initial value for registration, the convergence speed can be increased, the number of iterations can be reduced, meanwhile, the probability that iteration falls into a local extreme value can be reduced, the operation speed is increased, and the accuracy rate is increased.

The invention relates to the technical field of pathogen detection, relates to a POCT (point-of-care testing) rapid nucleic acid detection system, and in particular relates to a double-stranded probe-assisted loop-mediated isothermal amplification system for detecting brucella. According to the system, an amplification primer and a detection double-stranded probe are designed aiming at a Brucella specific gene bcsp31, and loop-mediated isothermal amplification (LAMP) and a double-stranded probe (TP) based on foot-mediated strand displacement are integrated to form a TP-LAMP unit of the scheme. The amplification result can be monitored in real time through a PCR luminoscope, and can also be visually identified through a nano biosensor (AuNPs-LFB). The method is rapid, sensitive, specific and easy to use, can solve the technical problems of non-ideal detection speed, specificity and sensitivity of a brucella POCT detection technology in the prior art, can be used as a valuable brucella rapid identification tool, and has ideal application and popularization prospects.

The invention discloses an SNP (Single Nucleotide Polymorphism) molecular marker combination related to corn kernel traits and application thereof, and belongs to the technical field of molecular genetics. The SNP molecular marker related to the corn hundred-grain weight is developed by researching the relation between the Zm00001d014723 gene and the corn grain weight, early prediction of the yield of the selfing line is further achieved through PCR amplification and agarose gel electrophoresis and through direct and rapid identification of seedlings, and a new thought is provided for molecular breeding.

The invention provides a portable detection device and method for rapidly identifying peritonitis, and relates to the technical field of medical instruments, the device comprises: an illumination module for irradiating a peritoneal dialysis liquid sample; the sensor module is used for capturing a sensing data set of the state change of the peritoneal dialysis liquid sample in the irradiation process; the detection chip downloads the trained peritonitis risk detection model, performs data detection on the sensing data set and outputs a risk detection result; and the display module is used for displaying on a display screen. The peritonitis risk detection device can solve the technical problem that a user is difficult to independently complete rapid and accurate detection in a family or community environment due to poor portability and complex operation of a traditional peritonitis detection device, and achieves the purposes of rapidly analyzing a peritonitis sample through optical characteristics and a machine learning algorithm, outputting a peritonitis risk detection result in real time and improving the detection efficiency. And the accuracy and the efficiency of early peritonitis recognition of the user or the medical personnel are improved.

The invention belongs to the technical field of precise and intelligent breeding, and provides a sheep apparent state recognition method and system based on an improved supervised learning network, and the method comprises the steps: apparent state category determination, data collection, dynamic trajectory prediction and filtering noise reduction, superposition windowing processing and feature extraction, data merging, data set grouping and model training. According to the method, dynamic trajectory prediction and filtering noise reduction are carried out on the collected data, and superposition windowing processing and feature extraction are carried out on the preprocessed data, so that the noise of the training data is reduced, the identification precision of the model is improved, scientific, accurate and rapid identification of the sheep apparent state is realized, and the identification efficiency of the sheep apparent state is improved. And breeding personnel can be helped to know the physiological and health conditions of the sheep in time.

The invention discloses a rapid identification and comprehensive evaluation method for a polygonatum cyrtonema provenance, and belongs to the field of improved variety breeding and quality control of traditional Chinese medicinal materials. The method systematically measures a plurality of specific biological traits covering three dimensions of growth, physiology and medicine. And carrying out standardization processing on the measured multi-dimensional data, carrying out dimensionality reduction by using a principal component analysis model, constructing an evaluation model capable of comprehensively reflecting the comprehensive quality of the provenance, and calculating a quantitative comprehensive evaluation score. The score can objectively and scientifically divide different provenances into four different quality grades including excellent, good, common and to-be-improved quality grades. The method has the beneficial effects that the strong positive correlation between the easily measured growth traits of the overground part and the medicinal component content of the underground rhizome for determining the final quality of the medicinal material is established, so that the rapid, predictive and non-destructive early screening of the provenance quality is realized.

The invention provides a D-A type organic fluorescent small molecule, a preparation method and application of the D-A type organic fluorescent small molecule in detection of nitrofuran antibiotics, and belongs to the technical field of fluorescent sensing. The D-A type organic fluorescent micromolecule is composed of 2-(2-aminophenyl) benzothiazole and an electron-donating group, the excitation spectrum of the D-A type organic fluorescent micromolecule is obviously overlapped with the absorption spectrum of nitrofuran antibiotics, and fluorescence quenching response can be shown through the inner filter effect; the trace nitrofuran antibiotics can be detected in a liquid phase in an ultrafast, high-sensitivity, excellent-selectivity and strong-anti-interference manner. According to the invention, a quantitative evaluation system can be constructed in combination with a smart phone, the RGB value of the image of the liquid-phase fluorescence sensor is output in real time, and rapid, accurate and visual quantitative detection of nitrofuran antibiotics in a real food sample or a water source is realized. In addition, the filter paper fluorescence sensor has the advantages of being portable and easy to prepare, and nitrofuran antibiotics can be rapidly identified on site in an actual scene.

The invention belongs to the technical field of molecular identification of insects, and particularly relates to multiple PCR (Polymerase Chain Reaction) primers and a method for rapidly identifying thrips insect varieties. The invention discloses a multiple PCR primer and a method for rapidly identifying the variety of thrips, the identification primer and a multiple PCR system can rapidly, effectively and accurately identify the three types of thrips, and the sizes of PCR product strips of tobacco thrips, comb flower thrips and common large thrips are 104 bp, 237 bp and 439 bp respectively. The detection limit of the sensitive tobacco thrips is 500 copies, and the detection limits of the sensitive tobacco thrips and the sensitive common large thrips are both 100 copies. Compared with common PCR which can only identify one type of thrips at a time, the multiplex PCR system can complete identification of various types of thrips at a time, has the advantages of being easy to operate, low in cost, capable of saving time and the like, has very high value in the aspect of field thrips detection, and can be widely applied to the fields of plant protection and the like.

The invention provides a multi-mode underwater biological intrusion detection method and device, and the method comprises the steps: carrying out the sampling of each sampling grid of a target water area through acoustic equipment and image equipment, and obtaining the acoustic data and image data of each sampling grid; extracting and fusing the acoustic data and the image data of each sampling grid to obtain multi-modal features of each sampling grid, taking the multi-modal features as input of a discriminator in the generative adversarial network, and obtaining a discrimination result output by the discriminator as existence of invasive organisms or absence of the invasive organisms in the sampling grids; under the condition that the invasive organism exists in the sampling grid, sampling and detecting the eDNA environment sample on the sampling grid, and determining whether the invasive organism exists in the sampling grid or not based on a detection result. According to the method, rapid identification is carried out by utilizing multi-modal information and acousto-optic characteristics, the identification accuracy is improved by utilizing eDNA, the defects of a single method are overcome, and the efficiency of monitoring invasive species is improved.

The invention belongs to the technical field of biological detection, and discloses an SNP (Single Nucleotide Polymorphism) site combination for rapidly identifying Chinese merino sheep. The invention provides the SNP locus combination for rapid identification of the Chinese merino sheep, and the combination consists of 10 SNP loci and has high specificity and relatively high prediction accuracy, so that variety identification of the Chinese merino sheep is realized. By integrating a molecular marker technology and bioinformatics analysis, the identification process is simplified, the identification efficiency and reliability are improved, and a solid foundation is laid for effective protection, scientific management and reasonable utilization of genetic resources of the Chinese merino sheep.

The invention belongs to the technical field of agricultural biology, and relates to a rice high-yield salt-tolerant molecular marker and application thereof. The rice high-yield salt-tolerant molecular marker is located at the 615th and 560th sites of chromosome 6 of a rice reference genome IRGSP1.0, and the polymorphic site is G / A. The invention further discloses a preparation method of the rice high-yield salt-tolerant molecular marker. The high-yield and salt-tolerant characteristics of the germplasm carrying the A nucleotide allelic variation at the site are obviously lower than those of the germplasm carrying the G nucleotide allelic variation, specifically, the yield of the GG genotype is the highest, the yield of the GA genotype is the middle, and the yield of the AA genotype is the lowest. By utilizing the molecular marker and the corresponding identification method, the genotype can be accurately and rapidly identified in the accurate design of the high-yield salt-tolerant rice, the phenotype identification cost of a field test in the later period is reduced, and the application of an accurate design technology of the rice in breeding is promoted.

The invention relates to a DRZXConv-based rapid identification method for percocephalus latus, which comprises the following steps of: collecting pictures, dividing the percocephalus latus into different morphological stages such as larva, adult, overlapping, curling and hiding, and marking the different morphological stages; an improved YOLO11 network model is constructed, all C3K2 modules in a Backbone network and a Neck network are replaced by C3k2DRZX modules, each C3k2DRZX module is a convolution structure obtained by introducing dynamic convolution Dynamic Conv and receptive field attention convolution RFAConv on the basis of an original C3k2 module, each C3k2DRZX module comprises a plurality of BottleneckDRZX submodules, each BottleneckDRZX submodule replaces each corresponding Bottleneck submodule in a universal YOLO11 network model, and each BottleneckDRZX submodule comprises a plurality of BottleneckDRZX submodule and a plurality of BottleneckDRZX submodule. Each BottleneckDRZX sub-module is used for realizing multi-scale feature acquisition through DRZXConv convolution, and the DRZXConv convolution is used for realizing multi-scale feature acquisition; and inputting a to-be-detected image into the trained model to obtain the position, the morphological category and the confidence coefficient of each percocele in the image. According to the method and the device, the percheilophila hemsleyi with different forms and characteristics can be more accurately captured, the recognition efficiency of the percheilophila hemsleyi in overlapped and hidden states is enhanced, and the recognition accuracy and reliability are improved.

The invention discloses a rainfall intensity rapid identification method and system based on multi-sensor data fusion, and relates to the technical field of hydro-meteorological monitoring and urban drainage scheduling, and the method comprises the steps: obtaining a historical rainfall data set, drawing a peak curve, and extracting an evolution sequence feature before a peak value; correlation analysis is carried out on the data and complete rainfall process characteristics to establish a historical rule model; rainfall intensity data are collected in real time through multiple sensors, and space-time dimension features are fused; matching the fusion features with historical features, and screening target historical events; constructing a rainfall reproduction prediction model to reconstruct a current rainfall evolution process; and finally, calculating an inflow load based on a reconstruction result, and generating a dynamic drainage scheduling scheme. According to the invention, accurate prediction of the rainfall process and intelligent scheduling of the drainage system are realized, and the urban waterlogging prevention capability is improved.

The invention relates to the technical field of intelligent camera shooting, and provides a camera with an intelligent mobile tracking function, which comprises a mounting seat for mounting and fixing, the bottom side of the mounting seat is rotatably connected with a rotating seat, the rotating seat is driven by an internal motor I, the outer wall of the rotating seat is connected with a base through a rod body, and the rod body is connected with the base. A spherical shell is arranged on the bottom side of the base, and a camera lens is installed on the bottom side of the spherical shell. Intelligent route prediction and rapid tracking can be carried out, a simulation route tracking module predicts a moving route of a target through a motion trend simulation unit to generate a simulation monitoring camera shooting path, so that a camera can make a tracking action in advance, and tracking data of a previous camera can be transmitted to a next camera to assist rapid identification; the risk of target loss is reduced, and the tracking continuity and accuracy are improved.

The invention relates to the technical field of dynamic monitoring, in particular to a chemical production safety risk dynamic monitoring system based on the Internet of Things, and the system comprises a data collection module which receives pressure monitoring data, temperature fluctuation data and pipeline flow velocity data measured by a chemical plant sensor in real time, records the time of each data point, and transmits the data to a server; and sorting and organizing the data according to a time sequence. According to the method, by receiving and sorting the chemical plant sensor data and implementing comprehensive analysis of the multi-scale time window, the capture and analysis precision of key parameter fluctuation is remarkably improved, the scientificity of rapid identification and risk prediction of variable fluctuation in the chemical process is promoted, potential risks can be found in advance, and the risk prediction efficiency is improved. Therefore, production strategies are adjusted in real time, production safety is ensured, pre-judgment of future risks is optimized by analyzing long-term trends and periodic changes, the accuracy of risk assessment and the safety of operation are improved, the probability of accidents is reduced, and the safety of personnel and equipment is protected.

The invention discloses a fault identification processing method and device based on an internet of things digital weighing system, and the method comprises the steps: a fault detection step: tracking and recording the weighing data and state of each sensor of the internet of things digital weighing system, and judging whether the sensors have faults or not; and an output replacement step: if a sensor fails, generating predicted weighing data according to the historical weighing data of the sensor and the weighing barycentric coordinates of the weighing platform, taking the predicted weighing data as the output of the failed sensor, and outputting the weighing value of the whole scale in combination with the weighing data of other normal sensors. According to the invention, rapid identification and early warning can be realized, and the influence of abnormal data of the sensor can be effectively eliminated or reduced by learning and utilizing historical weighing data and a weighing gravity center change rule of the weighing platform, so that the reliability of the internet-of-things digital weighing system can be effectively improved; and production losses such as shutdown and production halt caused by the abnormal problem of the sensor of the weighing system of the field user are reduced.

The invention discloses an Indel molecular marker for identifying ricefield eel strains in two Guangdong regions. The nucleotide sequence of the Indel molecular marker A is shown as SEQ ID NO: 5 or SEQ ID NO: 7; the nucleotide sequence of the Indel marker B is as shown in SEQ ID NO: 6 or SEQ ID NO: 8. The invention also provides a primer pair for amplifying the Indel molecular marker. By adopting the Indel molecular marker primer pair provided by the invention to amplify genome DNA of ricefield eel strains in different regions, the ricefield eel strains in the two regions can be quickly screened out according to amplified fragments. The Indel marker provided by the invention can be applied to rapid identification of germplasm resources of a specific geographical strain so as to obtain excellent characters of the strain and finally realize precise breeding.

The invention discloses a geothermal water layer quality rapid identification method and system based on an algorithm model, belongs to the technical field of geological exploration, and aims to correct logging curve data of a thermal reservoir research well area, effectively avoid possible system errors in a logging data acquisition process, including improper operation modes or instrument self-measurement errors and the like. Standardized logging data can more remarkably reflect regional geologic features compared with original data, principal component analysis is conducted on sample data through a PCA method, an original sample space is reconstructed according to the contribution rate of principal components, the dimensionality of an input space is effectively reduced, meanwhile, the correlation between variables of the input space is eliminated, and the accuracy of the input space is improved. The identification precision is improved; according to the method, the recognition accuracy is greatly improved, a more scientific and reliable method is provided for well logging data thermal reservoir recognition, and the method has practical significance in exploring geothermal resource development.

The invention discloses a rapid diagnosis method for blueberry continuous cropping soil obstacle factors, and belongs to the technical field of soil environment analysis. Based on combined analysis of soil physical and chemical factors, metagenome and metabolome, pathogenic bacteria types and key metabolites significantly related to the blueberry continuous cropping obstacle are extracted, and a characteristic diagnostic index system is formed. The index system comprises main pathogenic bacteria types, typical harmful metabolites, key soil physical and chemical factors and appropriate or risk threshold ranges thereof, and provides a scientific basis for rapid identification of soil continuous cropping obstacles. Comparative analysis is carried out by establishing a standardized index library, rapid comparison of related indexes and standard indexes in a newly sampled soil sample is realized, and whether the soil has continuous cropping obstacle risks or not is judged in an auxiliary manner by integrating the deviation degree of physical and chemical factors and biological factors.

The invention provides a tunnel back break intelligent rapid identification technology based on machine vision. According to the method, firstly, the actual space distance of a reference point on a tunnel face is obtained through a laser range finder, and the pixel and physical size conversion proportion is calculated in combination with a shot tunnel face image; and then, screening high-quality guide information by using a visual language model CLIP, and driving a segmentation model SAM for fine adjustment of a tunnel scene to realize accurate segmentation of the tunnel face contour. The contour pixel area obtained through segmentation is converted into the actual physical area in combination with the conversion proportion, the actual physical area is compared with the preset tunnel design contour area, and therefore the over-excavation or under-excavation state of the tunnel face is automatically judged, and the size of the over-excavation or under-excavation area is determined. According to the method, the automation degree, precision and real-time performance of back break detection in tunnel engineering are remarkably improved, the limitation of traditional manual measurement is overcome, and key technical support is provided for intelligent tunneling of a tunnel.

The invention provides an unmanned aerial vehicle photovoltaic real-time inspection system and method based on an end-edge-cloud architecture, and belongs to the field of photovoltaic power station operation and maintenance. Comprising an end device which directly interacts with an environment or a user and collects data and comprises an unmanned aerial vehicle end, a flight controller, a pan-tilt camera end and a mobile phone end; the edge computing platform is realized by adopting an embedded platform carried above the unmanned aerial vehicle end, and is used for reading end equipment data, executing fuzzy judgment based on a visible light image and two-stage defect detection based on an infrared image, and uploading an image detection result to the cloud server; and the cloud server is responsible for data storage and message pushing, receiving the image result uploaded by the embedded platform and pushing the image result to the mobile phone terminal. According to the invention, an end-edge-cloud collaborative architecture facing a power station is provided, the diagnosis and decision delay is greatly shortened, the rapid identification and positioning of the defects of the photovoltaic equipment are realized, and the high-precision identification of the defects is realized by adopting a two-stage detection algorithm based on target identification and defect classification.