752 results about "Component analysis" patented technology

The invention discloses a charger shell screw hole rapid positioning method based on three-dimensional point cloud identification, and the method comprises the steps: collecting a polarization structured light three-dimensional point cloud, carrying out the preprocessing, building a workpiece coordinate system, and obtaining a steady point cloud; calculating a homology bar chart in the neighborhood of the point cloud, screening candidate areas according to a threshold value, and generating candidate masks and topology confidence; performing principal component analysis and polar coordinate projection on the candidate region, and outputting a local point cloud; projecting and correcting a local point cloud to obtain a complemented point cloud and implicit field features; solving a hole axis by adopting random sampling consistency, and carrying out Gaussian mixture fitting on an output center and a hole diameter; fusing topology and geometric scores as joint confidence, and judging whether to enter execution or not; and mapping the center and the axis to a robot coordinate system, and compensating and updating a parameter output result during assembly. According to the invention, through three-dimensional point cloud identification and multi-stage geometric topology analysis, rapid and accurate positioning and assembly adaptive correction of the charger shell screw holes are realized.

The invention relates to the technical field of retrieval enhancement generation, in particular to a method for automatically adjusting retrieval enhancement generation parameters based on content feature modeling. The method comprises the following steps: receiving an original query text of a user, performing component analysis, identifying terminologies, general vocabularies and question entities, and quantifying to form query fingerprints; acquiring a historical behavior sequence of the user, and constructing a score reflecting the level of the user by combining the query fingerprints and adopting a time decay weighting algorithm; the user level score is converted into specific retrieval parameter configuration, and a retrieval strategy blueprint is formed; guiding document library retrieval according to the retrieval strategy blueprint, and screening out a candidate knowledge set which is most matched with the professional level of the user; and according to the user level score, a preset instruction template is intelligently filled, and a situational generation instruction is constructed. According to the method, the problem of non-uniform cognitive load caused by a traditional system is solved through a retrieval enhancement generation technology, and the technical knowledge transmission efficiency and the user satisfaction are remarkably improved.

The invention belongs to the technical field of data processing, and particularly relates to a hidden ore body evaluation and positioning method based on multi-source data processing. The method mainly aims at the problems of incompleteness and isomerism of multi-source geological data in acquisition, fusion and modeling. Comprising the following steps: acquiring hyperspectral, geochemical and magnetic anomaly multi-source data of an evaluation area; intelligently complementing missing modal data by using a generative adversarial network based on geological constraints and modal outburst to form a complete multi-source data set; an unsupervised clustering algorithm combining geological correlation and entropy weight analysis is adopted to construct high-confidence-coefficient pseudo-label data, and knowledge mining of unlabeled samples is achieved; feature purification and dimension reduction are carried out through multi-modal feature fusion and hierarchical principal component analysis, and key feature vectors representing the existence of the ore body are extracted; and finally realizing space prediction of the concealed ore body by utilizing the classification model. According to the method, a high-quality data basis and a unified processing framework are provided for intelligent recognition of the hidden ore body, and efficient and accurate positioning of the hidden ore body is achieved.

The invention discloses a Mars transverse wind ridging few-sample remote sensing interpretation method and system based on an SAM model. According to the method, contrast stretching preprocessing is carried out on an original Mars image, a dual-branch feature fusion framework based on VIT and CNN is constructed to extract and fuse image features, a position coding generator is introduced to support input of any size, a fine-tuning SAM strategy of a selective freezing encoder and a trainable mask decoder is adopted, and LoRA low-rank adaptation optimization calculation is combined, so that the Mars image is obtained. And a double-branch prompt generation module is used for fusing labeled and unlabeled data to generate a high-precision prompt, and finally, an interpretation result is output through a mask decoder and connected component analysis is carried out, so that instance-level labeling is realized. The system comprises an image preprocessing module, a double-branch feature extraction and fusion module, an image embedding generation module, a model fine tuning module, a prompt embedding generation module and an interpretation module. According to the method, the recognition precision and robustness of the mars transverse wind ridge formation under the condition of few samples are effectively improved.

The invention relates to the technical field of component analysis, in particular to an intermediate infrared spectrometer sensor verification system which comprises the following steps: acquiring spatial distribution information of a target sample through an automatic sampling module, and generating a corresponding first feature group based on the spatial distribution information; adjusting an emission wave band and a modulation strategy corresponding to a mid-infrared light source, converting mid-infrared photons into visible light signals, and generating a corresponding second feature group; performing down-sampling processing on the original spectral data to generate a corresponding third feature group; inputting the sparse spectral coefficient into a pre-trained deep learning reconstruction model, and dynamically correcting the reconstruction process in combination with an environment temperature compensation parameter to generate reconstructed spectral data corresponding to high resolution; and analyzing the reconstructed spectrum data, matching a preset substance spectrum database, extracting a characteristic absorption peak position and an intensity ratio of the target substance, and generating a corresponding final analysis result. According to the invention, the intelligence of the sensor verification system can be improved.

The invention provides a battery pack fault combined diagnosis method and system based on improved independent component analysis, and relates to the technical field of battery pack fault diagnosis. Comprising the following steps: acquiring voltage difference variation between adjacent single batteries in a specific time window as a diagnosis feature; constructing a diagnosis model by applying an independent component analysis method based on the diagnosis characteristics of the battery pack under the normal working condition; based on the diagnosis model, the SPE and the statistics are jointly used as detection indexes, and the voltage difference variation of the target battery pack is diagnosed and evaluated; when the fault of the target battery pack is diagnosed, calculating the improvement contribution of the diagnosis characteristics of the target battery pack to the SPE or the statistics, and obtaining an improvement contribution value; and the improved contribution value is compared with an average contribution value reference, abnormal variables are screened out, and fault type judgment and positioning are carried out in combination with the topological relation of the abnormal variables. According to the invention, rapid and accurate detection, positioning and isolation of connection faults and short-circuit faults are realized.

The invention discloses a data processing method for geological disasters, and particularly relates to the technical field of geological disaster monitoring and data processing. Collecting and normalizing multi-source monitoring data, and constructing a unified time sequence fusion vector; identifying a weak abnormal signal based on the disturbance change rate and a sensitivity model; high-disturbance feature points are extracted through sliding window clustering, and a dynamic risk factor matrix is constructed in combination with a remote sensing image; extracting dominant risk factor feature vectors by using principal component analysis, inputting the dominant risk factor feature vectors into a disaster evolution simulation model, and predicting a future high-risk area in combination with a historical path library; and finally, potential micro-disaster trigger points and risk levels are output, and graded early warning processing is realized. The system has the advantages of high disturbance identification precision, strong simulation prediction capability, automatic early warning response and the like, and is suitable for an intelligent early warning system for sudden geological disasters such as landslide and debris flow.

The invention discloses a power industry electroencephalogram cap-oriented key channel screening method and system, and the method comprises the following steps: firstly, collecting electroencephalogram signals of a power worker in typical states of waking, fatigue and the like by adopting a 32 or 64 channel electroencephalogram cap, and carrying out the preprocessing operations of band-pass filtering, ICA (Independent Component Analysis), signal-to-noise ratio evaluation and the like; artifacts and inferior channels are removed, and then multi-dimensional features such as the frequency domain, the time domain, the entropy value and statistics of each channel are extracted to comprehensively score the importance of the channels. Constructing a plurality of channel subsets, verifying the recognition performance of the channel subsets through a classification model, and screening out a group of key channels with the highest information expression capability in cognitive state discrimination; according to the method, the number of electroencephalogram channels is reduced, the complexity and calculation overhead of acquisition equipment are reduced, the feasibility of real-time deployment of the model in an electric power field is improved, and the method has relatively high industry adaptability and engineering application value and is suitable for various key scenes with relatively high requirements on personnel state perception.

The invention discloses a multi-dimensional binary software component analysis system and method, and the system comprises a static enhanced feature extraction module which is used for extracting multi-dimensional features of a target binary system through a disassembling engine, carrying out the fusion of the multi-dimensional features to form an enhanced feature representation model, and carrying out the preliminary matching through the combination of a feature hash library; the component verification module is used for recording runtime behaviors and analyzing the behavior log to verify a static analysis result; the intelligent knowledge base linkage module is used for realizing automatic traceability and risk assessment; the false alarm filtering and confidence evaluation module is used for optimizing credibility grading of a final result, constructing a false alarm filtering mechanism based on Bayesian reasoning, integrating results of multiple analysis stages, calculating the confidence of an identification result of each component, outputting a grading report and supporting a user to screen high-credibility results as required; and comprehensiveness, accuracy and intelligence of binary software component identification are realized.

The invention belongs to the technical field of electrical digital data processing, and particularly relates to a Hash cache-based software component analysis optimization method, which comprises the following steps of: S1, scanning multi-dimensional features for the first time, and calculating storage component Hash by fusing dynamic behavior features, the method comprises the following steps: acquiring dynamic behavior characteristics of components during operation through a lightweight sandbox environment, and extracting basic characteristics, content characteristics and dependency characteristics of a software package and each component; hash values are introduced to serve as unique identifiers of software packages and components, and the Hash values of the components and corresponding detection results are calculated and stored during first-time scanning; according to the method, the component hash value is verified preferentially during subsequent scanning, if the hash value exists in the cache library, the historical detection result is directly reused, the full-amount analysis process does not need to be repeatedly executed, and on the premise that analysis accuracy is guaranteed, repeated calculation is reduced through a cache reuse mechanism, SCA analysis efficiency is improved, and resource consumption is reduced.

The invention relates to the technical field of livestock breeding environment monitoring, regulation and control, and discloses a livestock breeding environment gas dynamic monitoring system. The system comprises a data acquisition module for acquiring data by using a multi-source gas sensor; the gas component analysis module is used for analyzing the data based on a mixed spectrum decomposition algorithm to generate a gas component characteristic matrix; the environment regulation and control decision module inputs the matrix into a multi-mode collaborative decision model to generate a regulation and control instruction; the optimization model construction module is used for constructing a multi-constraint dynamic optimization model to optimize ventilation equipment parameters; and the hierarchical control execution module is used for executing the regulation and control instruction through a hierarchical architecture. The system can comprehensively monitor breeding environment gas, accurately regulate and control ventilation equipment, achieve gas concentration equalization and energy consumption minimization, improve the breeding environment quality, reduce the breeding cost and push intelligent development of the livestock breeding industry.

The invention relates to a method, a system, equipment and a medium for identifying a broken generation of ancient ceramics. The method comprises the following steps: acquiring multi-dimensional digital image data of ancient ceramics and carrying out classification processing to generate an original image set; extracting and fusing macroscopic features and microscopic features of the original image set to generate a multi-modal feature matrix; on the basis of the multi-modal feature matrix, a sample set with the similarity higher than a preset threshold value in a historical database is matched through a deep hash algorithm, and a first cutoff interval is determined in combination with sample age information; obtaining a glaze component analysis result of the ancient ceramic, and confirming a second cutoff interval according to the glaze component analysis result based on the first cutoff interval; and obtaining an artificial generation interruption result, and generating a generation interruption identification result according to a preset comprehensive decision rule in combination with the artificial generation interruption result and the second generation interruption interval. Through fusion of image multi-dimensional features, science and technology detection data and expert experience, efficient, objective and accurate determination of the ancient ceramic age interval is realized.

A power transformation equipment temperature anomaly identification and diagnosis method comprises the following steps: collecting state variables, performing cleaning, interpolation complementation and abnormal point elimination on multi-source data through a time synchronization mechanism, and constructing a unified data matrix; extracting statistical features and time sequence dynamic features in the time sequence based on the data matrix, and performing dimensionality reduction on redundant information in combination with a principal component analysis method to form a multi-dimensional fusion feature vector; an unsupervised learning model based on LSTM-AE is constructed, a normal working condition data learning feature reconstruction mode is utilized, and a reconstruction error is taken as a criterion to identify potential temperature anomaly; and calling a preset expert rule base and a knowledge graph, automatically analyzing dominant factors causing anomalies, and identifying typical anomaly types. According to the invention, automatic identification and classification diagnosis of the temperature abnormity of the power transformation equipment under an unsupervised condition are realized, the accuracy and response speed of fault identification are obviously improved, and the intelligence and practicability of equipment operation state monitoring are enhanced.

The invention discloses a brain-controlled rehabilitation equipment-oriented control strategy generation method, device and equipment and a storage medium, and relates to the technical field of signal processing, and the method comprises the following steps: acquiring a multi-channel electroencephalogram signal, and preprocessing the multi-channel electroencephalogram signal to obtain a target electroencephalogram signal comprising a steady-state visual evoked potential signal and a motor imagery signal, the target electroencephalogram signal corresponds to a preset action category; decoding the steady-state visual evoked potential signal by adopting filter group task related component analysis to obtain a correlation score vector; decoding the motor imagery signal by adopting a Mangban dynamic routing space-time network model to obtain a classification score vector; performing posterior probability distribution conversion and weighted fusion on the correlation score vector and the classification score vector to obtain fusion probability distribution; and generating a target control strategy according to the action category corresponding to the highest probability value in the fusion probability distribution. The control strategy obtained by the invention can consider both intention recognition precision and rehabilitation nerve activation effect.

The invention relates to the technical field of skin care product detection, and discloses a multi-dimensional rapid detection and comprehensive analysis system for a skin care product formula. The method comprises the following steps: acquiring infrared spectrum data of a skin care product through a component spectrum acquisition device, and extracting characteristic peak information to generate an original spectrum characteristic vector; the formula component analysis engine matches the standard spectral characteristics, identifies active components and concentration proportions and outputs a component identification list; the component interaction relation modeling module constructs a component collaboration or antagonism relation network and generates a graph; the stability prediction unit analyzes component stability change trends in different environments and outputs a prediction report; the safety evaluation module calculates a safety threshold value and a risk level and generates an evaluation matrix; the formula optimization suggestion generator screens replaceable components or adjusts the concentration, and outputs optimization suggestions; and the detection report integration unit carries out structured integration on the result to form a final report. According to the system, multi-link integrated detection is realized, and the detection requirements of skin care products are met.

The invention discloses a graphite equipment weld defect intelligent detection and positioning system based on multispectral imaging, and particularly relates to the field of welding, and the system is characterized in that a multiband image acquisition unit acquires spectral image data of multiple bands and transmits the spectral image data to a multispectral data storage library; the spectral parameter analysis unit extracts characteristic parameters of the weld joint area and establishes a spectral characteristic parameter library; a spectrum abnormity identification unit locates a suspected defect area of spectrum abnormity; the weld defect positioning unit marks defect boundaries and ranges; the defect identification and grading unit calculates a defect severity index; the element component analysis unit is used for detecting the suspected defect position sample and analyzing the types and contents of trace elements in the suspected defect position sample; the defect and component data are integrated through the weld quality comprehensive evaluation unit, the quality grade and the rectification suggestion are generated, the problem that the defect cause cannot be positioned in the prior art is solved, data support is provided for follow-up maintenance and process optimization, and potential safety hazards of equipment operation and maintenance are reduced.

The invention discloses a boiler energy efficiency prediction method and system based on sampling flue gas component analysis. The method comprises the following steps: acquiring heat flux density data of a plurality of radial positions, and inverting a temperature distribution profile through a radiation heat flux prediction model; constructing a multi-component concentration layering graph based on the diffusion equation and the component transport equation; constructing a sampling path optimization model by using temperature gradient and concentration fluctuation intensity information, driving a dynamic telescopic sampling probe to arrange a sampling position, and collecting a flue gas sample; the analysis model determines original concentration data, thermal-mass offset correction is executed in combination with temperature distribution, and a concentration correction function is constructed to obtain a corrected concentration value; and fusing boiler operation condition data, and inputting into the energy efficiency prediction model to output a heat efficiency index. According to the method, temperature field inversion driven by heat flux density is constructed, a multi-component concentration layering graph is generated in combination with a diffusion and transport model, and the spatial representativeness of flue gas sampling, the physical authenticity of component data and the accuracy of energy efficiency prediction are remarkably improved.

The invention relates to the technical field of measurement and detection, in particular to a test sieve calibration method based on machine vision. Comprising the following steps: acquiring a test screen image by a microscope, firstly performing graying processing on an original image, and converting a color image into a grayscale image; the method comprises the following steps: carrying out binarization processing on a grey-scale image, carrying out binarization on the image, setting a grey-scale value of a pixel point on the image to be 0 or 255, enabling the whole image to present an obvious visual effect which is only black and white, and better analyzing the shape and the contour of an object through binarization; and performing connected domain analysis on the binary image, finding a pixel point to which each sieve hole belongs, endowing each pixel point with a label through the connected domain analysis, and forming a connected domain by the pixel points with the same label value so as to realize segmentation of the region of interest. The method is applied to the measurement calibration work of the test sieve, the working efficiency of verification and calibration personnel can be greatly improved by using the method, and human resources are saved.

The invention discloses a Beidou-based road slope disaster early warning method, and relates to the technical field of road slopes. The method comprises the following steps: determining and arranging Beidou receiving nodes in key subareas of a road slope for data acquisition to obtain road slope node data; performing displacement difference calculation on the three-dimensional coordinates of the road slope node data, and dividing key sub-region categories; performing feature extraction on the road slope node data to obtain a road slope node feature vector, constructing a road slope node similarity matrix, and based on the road slope node similarity matrix, obtaining a high-consistency track cluster through a density clustering method and contour coefficient screening; carrying out dimensionality reduction on road slope node feature vectors of high-consistency track clusters in the key sub-region categories by utilizing a principal component analysis method, and fitting track trends to obtain evolution indexes; euclidean distance is calculated through historical landslide event trajectory library indexes, a risk value is generated, and when the risk value exceeds a threshold value, early warning is performed and a risk level is established.

The invention relates to the technical field of engineering surveying and mapping, in particular to a structural point cloud principal axis identification method based on adaptive weighted principal component analysis, which comprises the following steps: acquiring a point cloud data set, initializing a weight set, and performing weighted centralization on the point cloud data set after point cloud downsampling; calculating a weighted covariance matrix according to the point cloud data set after weighted centralization, and obtaining a feature value and a normalized feature vector through the weighted covariance matrix; obtaining a distance measurement set of all point clouds; calculating an iteration weight set; and calculating the eigenvalue of the weighted covariance matrix and outputting an iterative spindle rotation angle. In the application of the method, along with the increase of the complexity of the point cloud, the advantage of adaptive weight estimation is more remarkable than that of common scale estimation, so that the high-precision rapid convergence performance when the common scale estimation is used for processing the uniform and symmetrical point cloud is reserved; and interference of outliers and environmental noise points on main shaft identification can be effectively prevented when highly non-uniform distribution point clouds are processed.

The invention relates to a TL-AEAT-BIGRU post-compression yield prediction method based on data joint driving under knowledge constraint, which adopts a CWGAN-GP model based on conditional constraint to perform data enhancement on a small amount of multi-source data, adopts a Pearson + mRmR correlation analysis algorithm to determine main control factors influencing post-compression yield, and performs prediction on the post-compression yield by using a TL-AEAT-BIGRU model. And classifying reservoir categories through a main component analysis method based on the reservoir classification standard established by the former. Correlation experience knowledge between the post-compression yield and the input main control factors is combined with the TL-AEAT-BiGRU post-compression yield prediction model, and the screened main control factors are used as input for post-compression yield prediction. The result shows that compared with other yield prediction models, the model has the best performance in the post-pressure yield prediction of the research area. An ablation experiment shows that each module of the model contributes to improvement of the yield prediction effect, so that the yield prediction accuracy of the model is effectively improved, the problem that the yield prediction precision of the reservoir after pressure is not high under the condition that the sample size is small is effectively solved, and the method has guiding significance on yield analysis of an oil and gas well.

The invention relates to the technical field of artificial intelligence, in particular to a software component analysis system based on artificial intelligence, which comprises a data acquisition and preprocessing unit, an AI-driven component identification and analysis unit, a risk detection and evaluation unit, a knowledge base and dynamic updating unit and a result visualization and disposal suggestion unit. According to the method, through a collaborative mechanism of multi-modal semantic fingerprint generation, cross-language component matching and dynamic and static feature dual verification, the bottleneck that traditional software component analysis is insufficient in recognition precision of high-confusion, cross-language and secondary packaging components is effectively broken through, and precise recognition and analysis of the components and key attributes thereof are achieved; therefore, a more reliable technical support which better meets actual business requirements is provided for software component security analysis.

The invention relates to the technical field of three-dimensional point cloud segmentation and terrain feature extraction, and discloses a step terrain segmentation method for enhancing L0 gradient minimization, which comprises the following steps of: calculating point cloud features based on iterative principal component analysis: constructing a dynamic weight distribution mechanism by introducing a robust M estimator; direction self-adaptive neighborhood optimization is carried out on the slope and panel transition area of the original point cloud, the continuity of the normal vector of the slope area is enhanced, and the geometric difference of the transition area is highlighted; multi-scale super voxel segmentation: adopting a normalized spatial metric segmentation method, generating super voxel units by using a distance, normal vector and perpendicularity weighted normalization mechanism, performing planar and non-planar classification on super voxels in combination with a multi-scale strategy, and performing recursive segmentation on non-planar super voxels at a smaller resolution; performing curved surface segmentation based on L0 gradient minimization and global energy optimization; according to the method, the definition of the segmentation boundary and the completeness of slope structure extraction can be remarkably improved, and the slope and the disc table can be accurately distinguished by using the vertical characteristics.

The invention relates to the field of power system clean low-carbon evaluation, and discloses an RAM-based power system clean low-carbon evaluation method and system. The method comprises the following steps: collecting multi-source heterogeneous data in real time and carrying out exception cleaning and structured processing to construct a standardized data set; dynamically adjusting the weight based on an entropy weight method and expert experience, calculating an efficiency value by using an RAM model, and optimizing a constraint condition; recognizing input redundancy and output insufficiency by decomposing slack variables, and generating a high-value improvement list in combination with principal component analysis and analytic hierarchy process; and finally, matching the optimization strategy library to form a decision report and realizing closed-loop updating of the index library. The system can dynamically evaluate the clean low-carbon level of the electric power system, accurately identify energy efficiency improvement key points, realize evaluation-optimization-feedback full-process automatic management, and significantly improve the accuracy of an evaluation result and the effectiveness of decision support.

The invention discloses an artificial leather quality rapid detection method based on spectral analysis, and particularly relates to the field of quality detection.The method comprises the steps that component composition characteristics, surface and microcosmic characteristics and dynamic response characteristics of artificial leather are synchronously obtained through multi-dimensional original data acquisition, and a time-space correlation data set is formed through timestamp alignment and space calibration; constructing component analysis, surface association and dynamic response models, and eliminating space-time migration through a synchronous optimization model; after multi-dimensional features are extracted, optimized features are generated through cross-direction tensor fusion, attention enhancement and gating cycle unit time sequence fusion; and finally, quantizing the quality index based on a multi-task model, and outputting a quality grade and a traceable report in combination with a dynamic threshold value through hierarchical coding, graph neural network association enhancement and evidence fusion. The single sample collection duration is less than 3 minutes, the response time is less than 1 second, and full-dimension, rapid and accurate detection and quality traceability are realized.

The invention discloses a de-signed binary code reverse analysis method based on library function matching. The method comprises the following steps of: obtaining de-signed binary codes and a candidate library file list thereof in the embedded equipment, and extracting a library function and information of each object file of each candidate library file; performing matching operation on each object file in sequence to establish a library function matching table; and acquiring addresses for calling other library functions in each library function, performing consistency comparison with the addresses of the successfully matched library functions, and confirming the calling relationship of the library functions, thereby determining the function of the de-signed binary code in the embedded equipment. According to the method, the library function in the de-signed binary code can be accurately identified, the calling relation is checked, and the method is suitable for component analysis and traceability of scenes such as computer programs, embedded firmware and DSP programs so as to accurately identify the function of the de-signed binary code in embedded equipment.

The invention provides a multi-cavity mold PET bottle preform production consistency optimization control method, which comprises the following steps of: arranging high-response infrared thermal imaging and piezoelectric pressure sensors to realize high-precision real-time synchronous acquisition of temperature and pressure data of each cavity, and establishing a cavity exclusive thermal history fingerprint model by using time sequence feature extraction and a deep neural network; combining the thermal historical fingerprints and the real-time filling data, dynamically calculating the thermal interference intensity between the cavities, and generating a multivariable coupling relation matrix; decomposing an overall consistency target by using principal component analysis and a multi-target optimization algorithm, generating a personalized compensation instruction adapted to thermal inertia and sensitivity of each cavity, and realizing multivariable decoupling cooperative control through adaptive PID parameter adjustment and virtual decoupling variables; wall thickness detection quality feedback is introduced, a model closed-loop calibration mechanism is constructed, adaptive iterative optimization of a control strategy and forming quality is guaranteed, and the consistency of multi-cavity injection molding products and the robustness of the system are remarkably improved.

The invention provides a food classification component identification method and system based on multi-scale feature attention synergy, and the method comprises the steps: constructing a multi-task neural network model, sequentially inputting a food image into a global feature module and a progressive local feature module, and respectively extracting global semantic information and refined local features of different scales; feature interaction between food category identification and component analysis tasks is promoted through a cross attention cooperation module, and key areas and important component information in images are dynamically selected; gradually activating different depth levels of the network by adopting a staged end-to-end training strategy, and introducing a KL divergence constraint to enhance the difference between feature scales; and simultaneously generating prediction results of food categories and component labels through a multi-task classifier. The technical problems of food identification and component analysis task splitting, insufficient local feature expression and insufficient cross-task information complementation are solved.

The invention provides a self-adaptive control method and system for equipment operation parameters, and relates to the technical field of self-adaptive control of oil field operation parameters, and the method comprises the steps: obtaining multi-source data such as structure displacement, process state parameters, component analysis data and fluid performance parameters to evaluate a current process stable state, and outputting a state quantized value; then, in combination with the component analysis data, the state quantized value and the current value of the fluid performance parameter, a pre-trained decision model is used for analysis and decision making, and an expected target value which the fluid performance parameter should reach is generated; and finally, the expected value serves as a leading target, feedback of process state parameters is sensed in real time, operation parameters of key process execution mechanisms such as a drilling machine and a slurry pump are dynamically adjusted through a self-adaptive control algorithm, conversion from passive response to active optimization in the drilling process can be achieved, and the well wall stability and the drilling efficiency are effectively improved.

The invention discloses a leakage monitoring method based on an LNG fuel gas system. The leakage monitoring method comprises the steps that a self-adaptive frequency modulation continuous wave active acoustic scanning network is established; blind source separation is carried out on mixed signals in acoustic scanning network abnormal events by adopting self-adaptive kernel independent component analysis; constructing a leakage feature mapping model of the physical information neural network; leakage source accurate positioning and quantification based on acoustic tomography and Bayesian reasoning are carried out; multi-modal decision fusion is carried out based on the multi-dimensional data sources received in parallel, a false alarm suppression mechanism is set, and time continuity verification and space consistency verification are carried out; establishing a reinforcement learning model for autonomously optimizing a monitoring strategy according to environment change and system state, and realizing adaptive optimization; the strategy network after self-adaptive optimization is deployed at the cloud, actions are generated regularly according to the current state, the actions are issued to the regional gateway and the edge node for execution, and iterative updating is carried out, so that the monitoring accuracy in a complex environment is improved, and the false alarm rate is reduced.