60 results about "Three dimensionality" patented technology

The present invention provides an integrated hyperspectral water quality analysis method, which belongs to the field of hyperspectral water quality analysis. First, data preprocessing is conducted by water quality data collection and water quality image collection in early stage; second, three dimensionality reduction methods are adopted to conduct dimensionality reduction processing, and fused dimensionality reduction is conducted by parameter trade-off selection; third, machine learning algorithms are adopted to train and test hyperspectral water quality inversion models on spectral data after dimensionality reduction; finally, the hyperspectral water quality inversion models are selected and optimized. The present invention adopts an innovative fusion strategy in the aspect of data dimensionality reduction processing, which can achieve a better data dimensionality reduction effect, effectively remove noise and redundant information, and provide a more accurate and reliable data basis.

The invention discloses an activated carbon forming defect online detection method and device, and belongs to the technical field of material defect detection.The method comprises the steps that technological parameters are collected in real time and preprocessed, a parameter sequence is generated, self-sensing recognition of a new technology is conducted, lagged starting of a quality control vacuum period is avoided, and once the new technology is recognized, the quality control vacuum period is started. An unknown defect pre-recognition mechanism is triggered immediately, novel anomalies which do not conform to known defects are recognized, and an unknown defect candidate set is generated; constructing a multi-stage hierarchical detection network, performing adaptive detection on the unknown defect candidate set, establishing a reference model, screening samples, updating the reference model, and forming a defect identification model adaptive to a new process, so that the influence of insufficient samples on the detection accuracy is effectively relieved; and setting three-dimensional quantitative indexes, judging a transition period, automatically switching to a full-model detection mode when the indexes meet preset requirements, and objectively evaluating the recognition capability of the detection system on new process defects and the process suitability through the quantitative indexes.

The invention provides a safety training model, method and system based on a capability evaluation dynamic weight optimization model, and a computer medium, and the method specifically comprises the steps: building a capability-knowledge-course three-dimensional map model, building a three-dimensional mapping relation, and determining an initial weight; fusing, optimizing, adjusting and triggering subjective and objective weights dynamically based on student behavior data, generating a personalized curriculum recommendation sequence through a reinforcement learning algorithm, forming a state space by an ability evaluation vector and a knowledge mastery degree vector, and taking a curriculum resource set as an action space; the short-term learning efficiency and the long-term post assessment passing rate are used as reward functions to construct a recommendation strategy, meanwhile, the invention further provides a computer medium achieved based on the scheme, and the method can provide safety capability assessment means and training courses which are high in pertinence and good in accuracy for chemical enterprise employees.

The invention belongs to the technical field of meteorological radars, and particularly relates to an airship-borne meteorological radar three-dimensional wind field inversion method based on multi-angle asymmetric coplanar analysis, and the method specifically comprises the steps: enabling an airship to fly linearly in a certain direction, and carrying out the volume scanning, doppler radial velocity and radar scanning information of a target observed at each parking position of the airship-borne meteorological radar are acquired, and the position and slant distance of the target are calculated; dividing every two parking position observation results into one group, and selecting a combination meeting performance optimization conditions based on the performance optimization conditions of the MA-ASCAT model; on the basis of an inversion formula of an MA-ASCAT model, first two dimensions under a cylindrical coordinate system are obtained through inversion by utilizing the screened combination, and a third dimension is solved according to the wind speeds of the first two dimensions on the basis of an incompressible mass continuity equation; and converting the three-dimensional wind speed under the cylindrical coordinate system into a Cartesian coordinate system, so that the three-dimensional wind field inversion method is completed.

The invention discloses an automatic generation method for health science popularization articles, and belongs to the crossing field of computer technology and health science popularization. The method comprises the following steps that S1, health authority information is collected through a multi-source crawler, and an original material pool is obtained through Sension-BERT vectorization duplicate removal and BioBERT medical entity labeling; s2, classifying nine types of health themes by using a RoBERTa fine tuning model; s3, multiple AI model supplementary materials are made into a structured package; s4, predetermining optimal selection questions in combination with AI three-dimensional scoring and editing; s5, calling the multi-dimensional knowledge base to generate a professional knowledge packet; s6, generating an outline and performing three-dimensional five-score system auditing of knowledge point fullness and the like; s7, the LLM expands and writes the text and marks knowledge sources; s8, performing double-stage auditing; s9, anthropomorphic draft moistening; s10, intelligently illustrating the picture; s11, when the matching degree is smaller than a preset threshold value, automatically updating the knowledge base; and S12, integrating and outputting. According to the method, the health science popularization article is generated, the manual participation time is shortened to be within 30 minutes, the medical error rate is reduced to be below 5%, the daily output of a 10-person team is improved by 3-5 times, the annual human cost is reduced by 60% or above, and the large-scale and high-quality science popularization requirements are met.

The invention discloses a water quality monitoring probe fault prediction and life evaluation method and system based on time sequence deep learning, and belongs to the technical field of water quality monitoring. The objective of the invention is to solve the problems of passive maintenance, no performance attenuation quantification, inaccurate service life evaluation and high early warning false alarm of the existing probe. The system obtains and preprocesses multi-dimensional time sequence data including response time, measurement precision and the like, inputs the data to a time sequence feature deep learning module, and extracts depth features by using an LSTM-CNN hybrid network in combination with an attention mechanism; the probe performance attenuation quantification module establishes a baseline according to the initial data, calculates a three-dimensional attenuation index and weights the three-dimensional attenuation index to obtain a comprehensive attenuation index; the Weibull life evaluation module performs modeling according to the data and calculates the residual life, the reliability and the confidence interval; the fault risk scoring module fuses multiple parameters and generates 0-100 score risk scores and four-level early warning and maintenance suggestions; through core algorithm innovation, predictive maintenance is realized, cost is reduced, monitoring continuity is improved, and the method is suitable for full-life-cycle management of the probe.

The invention discloses an anti-unmanned aerial vehicle interference effect evaluation method and system, and belongs to the technical field of anti-unmanned aerial vehicle interference effect evaluation, and the method comprises the following steps: collecting two types of anti-unmanned aerial vehicle scheme cases, carrying out the standardization and classification processing of cases which are good in evaluation but failing in off-site simulation, and building a failure case library; cases which are both good in evaluation and off-site simulation are classified, then the economic benefit range of each type of evaluation method is determined, and a successful case library is established. According to the method, a failure case library containing structured data and a success case library with a clear economic benefit range are established, a to-be-evaluated scheme is evaluated, then three-dimensional data are extracted from the failure case library to construct a multi-dimensional association model, and an evaluation blind area is positioned; and finally, three evaluation methods are generated in combination with a success case library, economic benefits are calculated and sorted, the difference between evaluation and off-site simulation can be effectively reduced, the evaluation accuracy is improved, and off-site simulation failure conditions are reduced.

The invention relates to the technical field of human-computer interaction, in particular to a B2C customer service response system based on sentiment analysis, which comprises the following steps: acquiring an input statement of a user in a B2C customer service dialogue and constructing a semantic sequence, then deconstructing semantic conflicts to judge the context polarity, and correcting a progressive link path to determine the final sentiment deviation of the user. And matching and adjusting the mood of the customer service response corpus according to the emotion deviation, screening the corpus of which the emotion and the mood are matched, and combining to generate a customer service response output result. According to the method, context linkage between entries is realized by constructing a semantic flow, a directional screening mechanism is introduced at a conflict position to unify emotion judgment, final sentence correction is combined in an emotion link to avoid deviation, and dynamic replacement is performed by comparing user expression amplitude in a mood level, so that generated corpora are kept consistent in strength and style; corpora are organized through continuous units, so that connection is natural, and synchronous optimization of dialogue content in three dimensions of emotional polarity, mood coordination and semantic coherence is achieved.

The invention relates to the technical field of medical image processing, in particular to a semi-supervised surgical instrument segmentation method and system based on a teacher-student model, and the method comprises the following steps: obtaining a three-dimensional ultrasonic image for surgical instrument segmentation; constructing and training a network structure formed by combining a single teacher network and three student networks to obtain a semi-supervised surgical instrument segmentation model for segmenting surgical instruments from the three-dimensional ultrasonic image; and outputting a surgical instrument segmentation result on the three-dimensional ultrasonic image by using the semi-supervised surgical instrument segmentation model. According to the method, the semi-supervised surgical instrument segmentation model is constructed through knowledge distillation of the single teacher network and the three student networks, so that the three student networks in the semi-supervised surgical instrument segmentation model are matched with three dimensions of three-dimensional ultrasound respectively, parallel processing of extraction and segmentation of three-dimensional image features is achieved, and the accuracy of three-dimensional image segmentation is improved. The information processing amount is small, the structural complexity is low, the efficiency is improved, and the timeliness requirement in the operation process can be met.

The invention relates to the technical field of engineering BIM (Building Information Modeling), and discloses a BIM digital-analog driving-based quantity calculation, price combination and carbon calculation integrated method and system, and the method comprises the steps: creating each professional three-dimensional model in three-dimensional modeling software; component attribute definition is carried out on each component in the three-dimensional model, and ID information is given to the components; all professional three-dimensional models are integrated in three-dimensional modeling software, the BIM model is imported into a BIM measurement platform, and a bill of quantity is generated; the engineering quantity list is imported into a cloud pricing platform and a cloud carbon metering platform for rapid pricing and carbon metering; the system automatically checks price and carbon calculation results, and quickly locates and checks the position of the BIM model with excess cost and excess carbon emission; and generating an engineering quantity list, a project cost file and a carbon emission report. The system comprises a three-dimensional model construction module, a model calculation integration module and a report generation module. According to the invention, three-dimensional management of quantity-carbon-cost in a design and construction stage of a municipal infrastructure project is realized.

This invention discloses an automatic statistical method for three-dimensional product structure analysis across multiple production lines based on stored procedures. It encapsulates statistical logic through database stored procedures and utilizes GROUP BY and CASE WHEN interval judgments to achieve automatic aggregation by steel type, thickness range, and width range. This method supports unified processing of data from 14 or more production lines and allows for flexible simplification of statistical dimensions. It solves the problems of complexity and inefficiency in traditional multi-production-line product structure analysis, and features high efficiency, stability, and ease of maintenance. This invention has been practically implemented in state-owned steel plants for many years.

The invention relates to the technical field of computers, and discloses a large model reasoning service dynamic connection number limiting method based on service attributes, which comprises the following steps of: 1, carrying out initialization setting on service attributes, resource parameters and a weight model; 2, collecting multi-dimensional data, wherein the collected multi-dimensional data comprises load data, service data and resource data; step 3, carrying out data preprocessing and data fusion on the collected multi-dimensional data to form a three-dimensional fusion data set; step 4, calculating a three-dimensional dynamic weight based on the three-dimensional fusion data set; 5, generating a connection number threshold value according to the three-dimensional dynamic weight, and carrying out connection request distribution and over-limit processing; and step 6, forming closed-loop feedback through real-time monitoring, historical backtracking and algorithm tuning. By utilizing the scheme of the invention, the refined management and control of the connection number in the large model reasoning service are realized, and the resource cost is optimized while the high-priority service quality is ensured.

Surface determination systems, threat detection systems and medical treatment systems are described. According to one aspect, a surface determination system includes processing circuitry configured to access a three-dimensional complex-valued image volume of a target, define image locations of the image volume using first and second dimensions of the image volume, for each of the first locations, identify voxels along a third dimension of the image volume that correspond to the respective first location, for each of the first locations, select one voxel that corresponds to the first location as having an increased amplitude compared with other voxels that correspond to the first location, and for each selected voxel, use phase information to identify a second location in the third dimension for the selected voxel that corresponds to a location of a surface of the target and that is different than a third location of the respective voxel.

This application discloses a method, apparatus, device, and medium for item recommendation across multiple user groups, scenarios, and tasks, relating to the field of artificial intelligence technology. The method involves acquiring user-side information, item-side information, and corresponding scene environment information in different scenarios; calculating weight dimensions, shared feature vectors, and environment dimensions to obtain a first dimension, a dense vector, and a second dimension; performing attention mechanism feature cross-calculation to obtain a third dimension; calculating item recommendation scores using the first dimension, dense vector, second dimension, and third dimension to obtain an item recommendation value; determining the target item recommendation value; identifying the target item information corresponding to the item recommendation value; and recommending the target item information to the user. This application can construct a three-dimensional perspective model encompassing multiple scenarios, user groups, and tasks, thereby improving the ability to represent differentiated user groups, meeting diverse user needs, and achieving comprehensive item recommendation across multiple user groups, scenarios, and tasks.

The invention discloses a three-dimensional space material grabbing method and device and a vehicle assembling system, and relates to the technical field of process control. Comprising the steps of determining two-dimensional coordinates of a material according to an image of the material; the sensing arm extends to the material from the first position until the sensing arm makes contact with the material so as to obtain a second position when the sensing arm makes contact with the material; according to the first position and the second position, coordinates of the material in a third dimension perpendicular to a two-dimensional plane corresponding to the two-dimensional coordinates are determined; and the materials are grabbed according to the two-dimensional coordinates of the materials and the coordinates of the materials in the third dimension. According to the technical scheme, compared with an existing 3D visual recognition technical scheme, the same grabbing function can be achieved, and the equipment cost is greatly reduced. Through measurement and calculation, the 2D vision cost of photographing to obtain images is reduced by one order of magnitude compared with the 3D vision cost, so that the method has a wide application prospect.

The invention provides a vehicle end picture dimension reduction method, device and system based on principal component analysis, and a medium. The method comprises the following steps: acquiring an image in an RGB format acquired by an intelligent vehicle camera; the image in the RGB format is converted into a first three-dimensional array, the first dimension of the first three-dimensional array is the height of the image, the second dimension of the first three-dimensional array is the width of the image, and the third dimension of the first three-dimensional array is an R color channel, a G color channel and a B color channel; separating pixel data of R, G and B color channels in the first three-dimensional array to obtain R, G and B pixel matrixes; carrying out dimensionality reduction on the R, G and B pixel matrixes by adopting a principal component analysis method to obtain R, G and B pixel matrixes after dimensionality reduction; and combining the R, G and B pixel matrixes after dimension reduction into a second three-dimensional array, and generating a dimension reduction image based on the second three-dimensional array. According to the invention, dimension reduction based on the principal component analysis algorithm can be independently carried out on each color channel in the RGB image, cross-channel interference is avoided, and independent optimization of color information is ensured.

The invention relates to a method for testing a fracture track of a metal material under a multi-field coupling effect. The method comprises the steps of sample preparation, high-temperature speckle preparation, sample loading, strain field acquisition, fracture morphology observation, data processing and fracture model establishment. By constructing a new fracture test method under the action of multi-field coupling, the problems of lack of physical field coupling factors, non-uniform test systems, limitation of fracture models and the like at present are effectively solved. A coupling loading test system capable of synchronously applying complex stress, high strain rate and transient high temperature is established, and real simulation of the multi-physical field coupling effect under the actual disaster working condition is achieved; a metal material fracture test system with universality and comparability is formed; key data support is provided for building a fracture model comprehensively considering the stress three-dimensional degree, the strain rate effect and temperature softening, the accuracy and reliability of fracture track prediction of the steel structure under the extreme disaster are remarkably improved, and an important basis is provided for engineering disaster resistance design and safety evaluation.

The invention provides a personnel positioning method and device, electronic equipment, a storage medium and a program product, relates to the technical field of data processing, combines the machine room characteristics of a data center, constructs a positioning tracking model, promotes a three-dimensional model to a three-dimensional space, realizes three-dimensional acceleration triple consistent distribution, and is beneficial to solving the blind spot problem in UWB positioning. Through the motion data and the acceleration at the current moment, the motion data at the next moment is calculated, and the accurate estimation of the motion state of the person at the next moment is realized. According to the invention, through positioning the tracking model, calculating the acceleration, calculating the motion data of the personnel at the next moment according to the motion data and the acceleration at the current moment, and correcting the motion data at the next moment, accurate positioning of the personnel in the machine room is realized.

The present invention discloses a head half-tube forming machine, comprising a discharge part, a three-dimensional forming part connected to the discharge part, and an arc adjustment part connected to the three-dimensional forming part; the discharge part is provided with a discharge cavity M3 for leading out the formed half-tube, the three-dimensional forming part comprises a three-dimensional discharge head, and the three-dimensional discharge head is provided with a three-dimensional discharge cavity M4 for leading out the three-dimensional spiral half-tube; it also comprises a first rotating part; the three-dimensional forming part comprises a second rotating part; the arc adjustment part is connected to the three-dimensional discharge head and drives the three-dimensional discharge head to reciprocate along the Z axis, so that the three-dimensional discharge cavity M4 reciprocates along the Z axis while rotating around the X axis, so as to adjust the three dimensions of the three-dimensional spiral half-tube. The present invention reduces the material production cost, the labor cost, improves the assembly efficiency, improves the welding quality and the appearance quality after assembly, and reduces the frequency of the common problem of leaks in pressure vessel products.

The invention discloses a transmission conductor galloping displacement abnormity monitoring method based on Beidou RTK, relates to the technical field of conductor displacement monitoring, solves the technical problems of low reliability of galloping parameter calculation and abnormity judgment and lack of continuous over-limit trend prediction capability, and provides a transmission conductor galloping displacement abnormity monitoring method based on Beidou RTK by realizing accurate coordinate system conversion and eliminating projection deformation influence. Two form reconstruction methods are provided, flexible selection can be performed according to requirements, meanwhile, environmental parameters such as real-time icing and wind speed are fused into a three-dimensional kinetic equation, the goodness of fit between a model and an actual galloping state is improved, an LSTM-ARIMA mixed time sequence model is constructed for a continuous overrun scene, and the abnormal trend is predicted in combination with meteorological data in the future 72 hours, so that the prediction accuracy is improved. And meanwhile, a three-dimensional composite risk index is introduced, three types of trends of slow remission, stable maintenance and rapid deterioration can be pre-judged in advance, and operation and maintenance personnel can deploy resources in advance according to the trends, so that the conversion from passive alarm to active prevention and control is realized, and the fault response time is shortened.

The invention discloses an urban and rural fusion development evaluation method based on a combined weighting and coupling coordination model, and relates to the technical field of data processing. The method comprises the following steps: S1, constructing an urban and rural fusion index evaluation system; s2, three-level combination weighting is carried out; s3, constructing a space-time two-dimensional improved coupling coordination model; step S4, performing evaluation based on ESDA and an obstacle degree model; and S5, multi-dimensional verification is carried out. According to the method, a high-quality data basis is provided for dynamic empowerment through accurate indexes screened in a three-dimensional mode, a scientific basis is provided for space-time coupling coordination calculation through reasonable weights of the dynamic empowerment, reliability is guaranteed through multi-scene verification of accurate results of space-time optimization, and finally, an intelligent evaluation report and a policy suggestion library are output. The method directly serves urban and rural fusion development decision, improves the overall evaluation efficiency, and improves the conversion efficiency from the evaluation result to the policy decision.

To achieve an automatic inspection of a biological tissue capable of highly accurate determination.SOLUTION: In the inspection image Tim, a feature amount is associated with a combination of the first coordinates, the second coordinates, and the specific physical quantity. The first coordinate and the second coordinate specify a position on the focal plane when the biological tissue is imaged. The specific physical quantity includes a physical quantity related to light from a substance of a biological tissue present at a position specified by the first coordinates and the second coordinates. The feature quantity includes the intensity of light having a specific physical quantity. The inference model Md extracts the information on the abnormality from the inspection image Tim by performing a three dimensional convolution operation on the feature amount with the first coordinate, the second coordinate, and the specific physical quantity as a first dimension, a second dimension, and a third dimension, respectively, on the inspection image.SELECTED DRAWING: Figure 6

This invention relates to a visualization method and application system based on multidimensional state space, within the field of visualization processing. It acquires and standardizes multidimensional signals and historical data from detection points. In the first visualization region, the radius of concentric circles in a Cartesian coordinate system is dynamically determined based on the historical data distribution density. The first and second dimensions are mapped to coordinates to obtain data points. Directional features are generated using quadrant positions, and deviation degree and level labels are generated using Euclidean distance. In the second visualization region, the third dimension is mapped using a bar chart array. The output is a comprehensive interface including dual regions and tables, presenting the evolutionary state using directional features. Finally, by using real-time statistical frequency analysis, when a structural drift of the expected benchmark occurs, the radius or level is corrected. This invention improves the accuracy and adaptability of monitoring.

The invention discloses a virtual-real interaction method and system for a naked-eye 3D large screen, and relates to the technical field of large screen interaction. According to the method, a crosstalk distance value, a crosstalk discrete value and a crosstalk deviation value three-dimensional evaluation system are introduced, a crosstalk general merit value is calculated after three-dimensional indexes are subjected to normalization processing, and a screening inclination angle with a relatively high crosstalk general merit value is selected as an optimal lenticular grating inclination angle of a corresponding grid unit; the crosstalk rate level, the fluctuation stability and the high-interference occurrence frequency are considered, and the problem that in the prior art, parameters are screened only based on the average crosstalk rate single index, and the stability and the reliability of the crosstalk rate are not considered is solved.

The application relates to the technical field of blast furnace ironmaking control, in particular to a metallurgical electric appliance data-oriented query analysis method and system. Silicon content in molten iron, slag byproduct alkalinity multiple and coal gas conveying capacity of key nodes are collected in real time through sensors. When any parameter exceeds a preset threshold value, it is determined that there is high-emission tail gas abnormality. Based on real-time data and historical records, corresponding device node sets are determined respectively. Sample data sets are constructed by using monitoring records in the high-emission abnormal state. A similarity calculation model is trained to convergence, and a deep learning model is generated. The three device node sets are input into the model, and potential leakage nodes are output. Visual warning information of target nodes is generated. The application integrates three-dimensional data, improves the high-emission determination accuracy, realizes accurate leakage point tracing by combining the device node space topological relationship and the graph neural network, identifies CO2 high-emission risk by the alkalinity change duration, and establishes a hierarchical warning system.

The application belongs to the technical field of electric power storage, and discloses a power warehouse operation simulation and bottleneck diagnosis system based on digital twinning. A multi-source perception architecture is constructed through an acquisition and adaptation module, covering full-factor data such as equipment operation and personnel operation. Combined with a five-dimension fusion model of a twinning modeling module and a laser scanning modeling technology, the physical space and operation logic of the power warehouse are accurately restored. A simulation and deduction module adopts a fusion simulation algorithm and a double-scene template library to realize linkage simulation of the whole process and control instructions of warehousing, storage, picking, and delivery, can simulate the operation state under different scenes, and avoid efficiency problems such as equipment waiting and path redundancy in advance. A bottleneck diagnosis module constructs a three-dimension evaluation system of operation state, control parameters, and equipment performance, integrates real-time working condition data and simulation analysis results, scientifically distributes index weights through an analytic hierarchy process, and realizes quantitative division of bottleneck grades.

Embodiments of the present disclosure relate to a method, a device, and a computer program product for rendering. The method comprises modeling a three-dimensional (3D) scene as a four-dimensional (4D) tensor, wherein a first dimension, a second dimension, and a third dimension in the 4D tensor correspond to an X-Y-Z coordinate axis in the 3D scene, a fourth dimension represents a channel dimension corresponding to an encoded feature, and the encoded feature is obtained by encoding the 3D scene with an encoder corresponding to a decoder. The method further comprises performing interpolation on the 4D tensor to obtain streaming scene information associated with the 3D scene. The method further comprises rendering the 3D scene on the basis of the streaming scene information. According to the method, a unified and flexible architecture can be provided for edge-cloud 3D collaboration, and a 3D scene representation can be processed effectively and efficiently.

The invention discloses a resistor pin welding defect detection device based on machine vision, and relates to the technical field of welding defect detection devices. According to the invention, through cooperation of the servo motor, the screw rod and the guide rod, the convenience and precision of detection position adjustment are improved, and then the function of multi-angle detection of the resistor pin can be realized; the analysis module is used for filtering anomalies in the image data, and compared with a traditional method, the efficiency of filtering aperiodic noise such as sudden strong light and equipment vibration is improved; when gray scale deviation is detected for the first time, single sampling misjudgment is avoided through recursive verification of n groups of gray scale blocks; accurate welding spot edge positioning is realized by using the gray difference between the pins and the welding spots and adjacent block logic; the area, length and width of a welding spot are decomposed into independent dimension parameters, three-dimensional comparison is performed in combination with a preset standard, the defect type is accurately positioned, and the situation that subsequent process parameters are adjusted in disorder due to positioning errors of defect roots is prevented.

The embodiment of the invention provides an image semantic alignment re-recognition method and device, equipment and a storage medium, and relates to the technical field of image processing. The method comprises the following steps: obtaining a global semantic matching degree according to an identification query feature and an identification candidate feature, obtaining a target semantic matching degree according to an identification mask query feature and an identification mask candidate feature, and obtaining a matching pair according to an alignment mask query feature and an alignment mask candidate feature; and obtaining an alignment semantic matching degree of the identification mask query feature and the identification mask candidate feature based on the matching pair, and obtaining a re-identification result of the query image and the candidate image according to the global semantic matching degree, the target semantic matching degree and the alignment semantic matching degree. The unsupervised alignment capability of the hidden space diffusion model is utilized to reduce the labeling cost, the accuracy of feature association is enhanced through global, target and alignment three-dimensional semantic matching, and finally dual optimization of the labeling cost and the recognition accuracy in a semantic re-recognition task is realized.

The present invention provides an integrated hyperspectral water quality analysis method, which belongs to the field of hyperspectral water quality analysis. First, data preprocessing is conducted by water quality data collection and water quality image collection in early stage; second, three dimensionality reduction methods are adopted to conduct dimensionality reduction processing, and fused dimensionality reduction is conducted by parameter trade-off selection; third, machine learning algorithms are adopted to train and test hyperspectral water quality inversion models on spectral data after dimensionality reduction; finally, the hyperspectral water quality inversion models are selected and optimized. The present invention adopts an innovative fusion strategy in the aspect of data dimensionality reduction processing, which can achieve a better data dimensionality reduction effect, effectively remove noise and redundant information, and provide a more accurate and reliable data basis.