411 results about "Computational Technique" patented technology

The invention discloses a power grid energy storage capacity demand determination method and system based on multiple time scales, relates to the technical field of power grid energy storage capacity demand calculation, and aims to solve the problem of inaccurate energy storage capacity demand calculation. By generating multiple scenes, quantitatively screening key scenes and incorporating various uncertain factors, energy storage capacity calculation focuses on high-influence scenes, the coping capacity of the scheme to actual risks is enhanced, decision scientificity is improved, ultra-short-term to long-term multi-time scales are divided, core contradictions of all the scales are captured in a targeted mode, limitation of a single scale is avoided, and energy storage capacity calculation efficiency is improved. According to the method, capacity requirements and equipment distribution are integrated, operation rules are defined, visual documents are generated, multi-scale energy storage cooperative operation is achieved, the stability of a power grid is guaranteed, meanwhile, cost is reduced, scheme landing performance and operation efficiency are improved, multi-time-scale power grid state prediction is carried out based on a dynamic database, and historical rules and real-time data are combined, so that the power grid state prediction efficiency is improved. And the prediction coordination is ensured through multi-dimensional verification.

The invention discloses a cross-platform virtual-real fusion scene construction method and system based on AI space calculation, and relates to the technical field of artificial intelligence and space calculation, and the method comprises the steps: carrying out the multi-scale feature fusion based on a received cross-modal conversion instruction set, and generating an initial image sequence; carrying out implicit field coding on a target object by combining a three-dimensional reconstruction algorithm to obtain an initial parameterized model; performing space-time alignment on the multi-view video stream, loading a digital scene asset package in combination with physical sensing data and a preset spatial index structure, and establishing a bidirectional data channel between a virtual scene and a physical sensor; performing rendering and illumination parameter adjustment on the initial parameterized model to obtain an optimized parameter model; performing differential coding processing on the optimization parameter model to obtain a target virtual-real scene fusion model; and distributing the target virtual-real scene fusion model to a preset terminal. The invention provides a virtual-real fusion construction method for end-to-end collaborative optimization, which is suitable for cross-platform live broadcast or dynamic interaction scenes.

The invention relates to the technical field of tunnel crack detection and artificial intelligence edge calculation, and provides a tunnel apparent disease detection method based on deep learning and knowledge distillation, which comprises the following steps: step 1, introducing spectral domain information enhancement to an original tunnel image, the edge texture features of the disease area in the image are enhanced through methods such as multi-scale wavelet transform and small-scale enhancement. Step 2, constructing a high-performance teacher model, introducing a flexible up-sampling structure to adapt to feature recovery requirements of different levels of semantic information, introducing an efficient visual coding module to enhance feature fusion capability of different scale channels, and designing a scale adaptive weighted loss function at the same time; by introducing a frequency spectrum enhancement mechanism, structural features of disease areas with low contrast, fuzzy edges and the like are remarkably enhanced in an image preprocessing stage, clearer information input is provided for a model, and the stable recognition capability of a system in environments of uneven illumination, complex background and the like is enhanced.

The invention discloses a heterogeneous computing power cooperative scheduling system and method for mixed precision training, and belongs to the technical field of artificial intelligence computing. The system comprises a computational graph analysis and operator portrait module which is used for analyzing and dividing a model computational graph and extracting operator features; the heterogeneous hardware capability sensing and matching module is used for managing performance files and real-time states of heterogeneous hardware in the cluster and matching optimal execution hardware for each calculation partition; and the data flow coordination and pipeline parallel controller is used for generating a global execution plan, managing cross-device data dependence and communication and calculating overlapping optimization execution efficiency through communication. According to the method, the problem of low scheduling efficiency of mixed precision training in a heterogeneous environment is solved, automatic and accurate mapping from a calculation task to heterogeneous hardware is realized, the training speed is remarkably improved, the training cost is reduced, and the overall resource utilization rate of a cluster is improved.

The invention discloses a drainage pipeline multi-parameter sensing trenchless repair dynamic regulation and control method and system. Through a fusion technology path of multi-source comprehensive data acquisition, edge calculation real-time cross validation of abnormity, pipeline state simulation analysis of defect levels, intelligent algorithm dynamic adjustment of repair parameters and geographic information visualization platform full-process monitoring, closed-loop management from abnormity identification to repair regulation and control is realized. A comprehensive data set is formed through multi-source data acquisition, an edge computing technology is utilized to quickly identify abnormities, a simulation technology is combined to evaluate the severity of defects, then process parameters are optimized and repaired based on defect levels, and a regulation and control instruction is generated through linkage of a visual platform and a pump station dispatching system. And finally, a pipeline repair and operation regulation and control scheme is formed through integration, and the repair effect and the system stability are ensured. According to the method, the pipeline abnormity processing accuracy and the repairing efficiency are remarkably improved, and a technical guarantee is provided for safe and stable operation of an urban drainage system.

A facial expression recognition method based on grid attention and pyramid segmentation attention belongs to the technical field of deep learning image processing and emotion calculation, and comprises the following steps: introducing a grid attention mechanism and a pyramid segmentation attention mechanism on a ResNet101 large model, capturing local expression detail features through a grid attention module, establishing multi-scale global feature association by using a pyramid segmentation attention mechanism; hierarchical segmentation is carried out on the backbone network, and information from multiple hierarchies is effectively fused; and dynamically balancing the contribution degree of each level of features through learnable parameters, and integrating the extracted features to judge the expression category. Experiments show that the model achieves the recognition accuracy superior to that of a traditional model on a public data set in natural scenes such as complex illumination and posture change. According to the method, a solution with high robustness is provided for facial expression recognition in a complex environment, and the method has important application value in the fields of intelligent human-computer interaction, mental health assessment and the like.

The invention provides a low-cost visual field large model for visual multi-modal information processing, and the model comprises an image encoder module which is used for converting an input image into low-dimensional feature representation; the feature extraction module is used for extracting multi-scale visual features based on a hierarchical multi-task learning strategy and reducing redundant calculation through a cross-modal parameter sharing mechanism; the task specifying module is used for designing a lightweight sub-model for image classification, target detection and image generation tasks, and integrating pruning and quantification technologies to optimize calculation efficiency; the reasoning optimization module is used for reducing model reasoning complexity and energy consumption by adopting a low-rank decomposition LoRA and mixed precision calculation technology; and the multi-modal fusion module is used for integrating vision, text and sensor data through a cross-modal attention mechanism to generate cross-modal joint feature representation so as to improve task robustness in a complex scene. According to the method, through the design of model architecture, parameter quantity and calculation optimization, the requirement for hardware resources is remarkably reduced.

The invention relates to the technical field of computer vision and emotion calculation, in particular to an emotion real-time recognition and long-term monitoring method based on facial expressions. According to the method, an emotion recognition result is obtained by recognizing a high-definition facial image, and the emotion recognition result, environment information and physiological state data are fused to obtain time-space aligned multi-modal data; performing emotional causal analysis based on the multi-modal data, and judging emotional causes by combining a rule engine and a machine learning model: outputting a real-time emotional state recognition result and a periodic emotional report according to the emotional causes, and performing differentiated feedback according to the emotional causes. According to the method, through multi-source data fusion and a causal inference mechanism, the accuracy and interpretability of emotion recognition are effectively improved, and the technical span from passive recognition to personalized active intervention is realized.

The invention relates to the technical field of endoscope video processing and intelligent edge computing, in particular to an endoscope video enhancement processing intelligent edge computing system. Comprising a data acquisition module which is used for acquiring an original video frame sequence of edge endoscope equipment in real time; the feature extraction module is used for determining an instantaneous feature vector representing the dynamic change of the operation scene; the criticality quantification module is used for quantizing and generating a surgical event criticality score; the tuning logic module is used for generating a discrete and stable calculation normal form switching instruction; the assembly line switching module is used for responding to the calculation normal form switching instruction and executing an asynchronous weight preheating strategy; the utility evaluation module is used for constructing a dynamic utility function to evaluate system performance; and the threshold value correction module is used for performing closed-loop correction on the high-criticality threshold value by adopting a gradient rising strategy. According to the method, the stability of system decision making is enhanced, the smooth transition of the video processing flow among different calculation paradigms is ensured, and the robustness of the system is improved.

The invention relates to the technical field of multi-camera image processing and computational photography, and discloses an image processing method and system based on four-camera cross-focal-length continuous zooming fusion. The image processing method comprises the steps of system initialization, image acquisition, zoom routing, automatic ROI extraction and tracking, field-of-view cutting and geometric alignment, image fusion and binocular depth recognition. Through systematized multi-camera collaborative design, an innovative mechanism is introduced in key links such as zoom routing, geometric alignment, image fusion and depth recognition, smooth zoom, space consistency, detail fidelity, power consumption optimization and high-quality 3D perception are realized, the imaging quality is improved through the effects, the application scene is expanded, and the application prospect is wide. And a comprehensive solution is provided for mobile photography, AR and intelligent visual systems.

The invention discloses a ship water gauge reading semantic segmentation method, relates to the technical field of ship water gauge scale calculation, and aims to solve the problems of inaccurate reading and low efficiency of traditional manual or conventional image processing in a complex environment. According to the method, firstly, a water gauge area image sequence is obtained and preprocessed, and then the water gauge area image sequence is input into a deep learning semantic segmentation model; when the model is trained, a composite loss function combining weighted cross entropy loss, Dice loss and a wave region prediction confidence penalty term is adopted so as to enhance segmentation robustness under a complex background. An initial segmentation image output by the model is further optimized through a prior geometric constraint-based connection algorithm, a water gauge number and a scale line are accurately positioned based on an optimized segmentation result, and a draft reading is smoothly calculated through data. And the accuracy, efficiency and reliability of water gauge reading automation are obviously improved.

The embodiment of the invention provides a spectrum red shift measurement method, device and equipment and a medium, which can be applied to the fields of astronomical information technology and high-performance computing technology, and the method comprises the following steps: receiving N observation spectrum data to be measured; determining a processing mode of the observed spectrum data; in response to the processing mode being a first processing mode, processing the observation spectral data through a first calculation path, the first calculation path performing red shift measurement calculation based on a first numerical calculation library by matching the observation spectral data with a set of template spectral data to obtain a calculation result, the calculation result of the first numerical calculation library is consistent with the calculation result of a reference algorithm within a preset precision range; in response to the processing mode being a second processing mode, the observed spectral data is processed through a second computational path that reconstructs the redshift measurement calculations into a tensor parallel model based on a second library of numerical calculations and performs on parallel computing hardware, the second library of numerical calculations supporting tensor calculations.

The invention discloses a ship water gauge scale line fitting algorithm, and relates to the technical field of ship water gauge scale calculation, and the method comprises the steps: obtaining a ship water gauge region image through an image collection device to extract a measurement point coordinate, and building a mapping relation between a pixel position and a physical draft; a weighted least square method is adopted to perform nonlinear curve fitting, a weight coefficient is calculated based on a local neighborhood standard deviation, a polynomial order p is dynamically adjusted according to an adjusted decision coefficient, and the decision coefficient is iteratively optimized to improve the balance between amplitude and model complexity. And the system repeatedly corrects model parameters through a dynamic verification mechanism of the residual sum of squares and a preset threshold value, and finally an accurate water gauge scale fitting curve is generated. The method effectively solves the problems that traditional manual observation is prone to environmental interference and large in measurement error, high robustness is kept in a complex water area environment, and reliable data support can be provided for the fields of ship load evaluation, channel safety management and the like.

The invention discloses an electric ship remote intelligent control method and system based on cloud-side cooperation, and relates to the technical field of cloud-side cooperation computing, and the method comprises the steps: carrying out the feature extraction of images around a ship and time-space aligned multi-modal data, obtaining visual features, motion features and environment features, carrying out the feature fusion, and generating an environment situation vector; inputting the environment situation vector into a channel boundary detection model, analyzing the emergency degree of the current situation, generating a collision avoidance instruction, and generating a ship control decision through a three-level collaborative decision; combining a ship control decision with ship control characteristics, simulating a traffic development situation in a navigation area, comparing with historical ship navigation conflict data, and identifying potential conflict data; through collaborative fusion of the ship-end decision-making unit, the shore-end decision-making unit and the cloud-end decision-making unit, an optimal control strategy is generated in combination with respective advantages, intelligent decision-making of global optimization in a complex environment is realized, and adaptability and safety of ship control are enhanced.

The invention relates to the technical field of digital twinning and computer simulation, in particular to a power transmission and transformation project optical cable connection information operation and maintenance system based on a logic model, which comprises a model fusion module for constructing a digital twinning anchor point set by fusing a three-dimensional physical model and a secondary system logic model of a power transmission and transformation project; when the simulation analysis module receives a logic loop alarm, the physical path is reversely traced based on anchor point data, physical field simulation operation is executed, and a physical degradation risk point is output; the augmented reality module realizes spatial registration of a three-dimensional physical model and a live scene through a spatial computing technology, and visually presents logic loop and risk point information. According to the method, the space-time twin anchor points are constructed, the physical field simulation operation is driven, and the diagnosis result and the operation and maintenance instruction are superposed on the live scene through the augmented reality technology, so that the intelligent operation and maintenance closed loop of the optical cable connection information of the power transmission and transformation project is realized, and the troubleshooting efficiency and accuracy are improved.

The invention discloses a method for detecting the state of an earthing knife circuit breaker based on deep learning and a key point detection calculation technology, and relates to the field of power equipment detection, and the method comprises the steps: collecting a visible light image; performing data enhancement processing; building a deep learning model and improving the model; an improved EDC-YOLO model is introduced and a GeoKeyNet method for key point geometric feature calculation is combined, accurate state judgment can be carried out on obvious and easily-recognized key points on a grounding knife switch, and particularly under the condition that some grounding knives have small displacement, key point detection can effectively capture subtle changes, so that the opening and closing state of the grounding knife switch is accurately recognized, and the detection accuracy of the grounding knife switch is improved. The method not only makes up the limitation of traditional sensor monitoring, but also provides an efficient, accurate and automatic monitoring scheme especially under the condition that subtle changes cannot be fully captured and comprehensive monitoring cannot be achieved, and the monitoring precision and the working efficiency of the grounding knife-switch circuit breaker can be remarkably improved.

The invention discloses a night enhanced imaging and dynamic denoising method for an automobile data recorder, relates to the technical field of vehicle-mounted image processing and computational photography, and is used for solving the problems of insufficient brightness and too strong noise of night videos of the automobile data recorder under a low-illumination condition. Estimating the zero-mean Gaussian noise intensity of a read link and the Poisson noise intensity increasing along with the brightness in the same frame, converting the zero-mean Gaussian noise intensity and the Poisson noise intensity into brightness equivalent parameters, and introducing a unified control function G as the scale reference of all enhancement and noise reduction links; defining a sharing priori H and carrying out unified constraint in dense optical flow, depth time sequence denoising and local tone mapping; and in combination with the motion mask and the background model, performing differential space-time denoising and proportional write-back, and in cooperation with motion blur detection, spatial variation motion kernel and constrained deconvolution and multi-frame detail recharge, obtaining target brightness and outputting an RGB frame after proportional write-back.

The invention relates to the technical field of earthwork calculation, in particular to an earthwork calculation method based on an unmanned aerial vehicle aerial survey technology. According to the technical scheme, the method comprises the steps of image control point layout and data acquisition, unmanned aerial vehicle image data acquisition, aerial triangulation processing and model generation. Based on an unmanned aerial vehicle aerial survey technology, image control points are scientifically arranged to collect high-precision coordinates, high-quality image data are obtained, a high-precision model is generated through aerial triangulation processing, the precision is strictly verified, data are extracted through professional software to construct an optimized DTM model, and precise earth-rock volume calculation is achieved; a calculation result is encrypted and stored by means of a block chain and is in butt joint with a construction management system, and a GIS platform and a BIM model are combined to generate a visual progress diagram, simulate slag yard capacity early warning and the like, so that the precision and efficiency of earth-rock calculation are comprehensively improved, powerful support is provided for engineering construction management, auditing tracing and ecological restoration decision, and the method is suitable for popularization and application. And the scientificity and informatization level of project management are obviously improved.

The invention discloses a video image generation method and system driven by edge computing, and relates to the technical field of computer vision and edge computing, and the method comprises the steps: a terminal device initiates a generation task and perceives resources; dynamically dividing the generated task into a plurality of sub-tasks through a decision model based on the perception information, and deciding the execution position of each sub-task at the terminal, the edge node or the cloud; the nodes cooperatively execute the distributed subtasks and exchange intermediate data; and a final synthesis result is fed back to the terminal. The system comprises a terminal module, an edge node module, a cloud data center module and a collaborative management module. According to the invention, through terminal-edge-cloud cooperative computing, delay of a generation task and cloud bandwidth consumption are effectively reduced, data privacy protection is enhanced, system resource utilization efficiency and expandability are improved, and the method is especially suitable for video image generation application scenes with high real-time performance requirements.

The invention provides a proton dose distribution determination method and device based on a graphics processor, which can be applied to the technical field of medical radiation dose calculation. The method comprises the following steps: calculating pencil beam tracks of a plurality of pencil beams; obtaining respective dose nuclear parameters of the pencil beam at a plurality of intersection points; according to the radiation field dose truncation radius, voxels of interest are determined from tissue voxels in the tissue voxel grid, and the radiation field dose truncation radius represents the upper limit value of the acting distance of all pencil beams in one radiation field to generate remarkable contribution to dose distribution in the transverse direction; calling a plurality of threads to execute dose calculation tasks of the plurality of voxels of interest in parallel based on the respective dose kernel parameters of the plurality of intersection points to obtain voxel deposition doses of the voxels of interest; and determining a dose distribution result related to the specified tissue according to the voxel deposition doses of the plurality of voxels of interest.

The invention discloses a mode-missing-oriented multi-mode small sample emotion recognition method and device, and relates to the technical field of emotion computation.The method comprises the steps that S1, electroencephalogram data and eye movement data are obtained; s2, constructing a multi-modal small sample emotion recognition model comprising a feature extraction module, a missing modal reconstruction module, an uncertainty perception adaptive fusion module and a small sample classification module; s3, training the multi-modal small sample emotion recognition model by using the electroencephalogram data and the eye movement data to obtain a trained multi-modal small sample emotion recognition model; and S4, performing emotion recognition by using the trained multi-modal small sample emotion recognition model. According to the method, the multi-modal small sample emotion recognition model is constructed, and the technical effect of remarkably improving the accuracy, robustness and generalization ability of multi-modal emotion recognition in complex reality scenes such as modal deficiency, labeled sample scarcity and cross-subject distribution offset is achieved.

The invention relates to the technical field of interface thermal resistance calculation, and discloses a method for measuring and calculating interface thermal resistance based on an infrared imaging technology, which comprises the following steps: fixing and aligning paired upper and lower test pieces to form a to-be-measured contact interface, respectively preheating to a preset initial temperature difference and stabilizing, and monitoring and recording temperature change and test piece thermophysical parameters; enabling the test piece to be in contact, starting infrared imaging equipment to record full-field transient temperature data in a temperature equalization process, and forming a three-dimensional data matrix; building a thermal resistance correlation model, constructing a heat conduction equation in combination with the parameters to solve a simulated temperature field, and inverting related heat conduction coefficients by using a conjugate gradient method; and calculating microscopic and total contact thermal resistance according to the inversion coefficient and the geometric parameters, and deducing macroscopic thermal resistance. According to the method, non-intrusive transient infrared temperature measurement is combined with a reverse heat transfer algorithm, so that the microscopic thermal resistance, the macroscopic thermal resistance and the total contact thermal resistance are accurately distinguished and quantified.

The invention discloses a security and protection monitoring data sharing and analysis platform and method based on cloud computing, and relates to the technical field of cloud computing. The security and protection monitoring data sharing and analysis platform and method based on cloud computing comprises the following steps: S1, collecting monitoring data in real time, performing data preprocessing, and optimizing visual images and video streams; s2, a distributed computing architecture is designed, uploading and synchronous data sharing of monitoring data are carried out, and edge device computing is optimized; s3, target vehicle detection and state initialization are carried out, multi-target vehicle tracking is realized, the vehicle position and state are predicted, target matching is carried out, error correction is carried out, and real-time updating and feedback are carried out; and S4, identifying a traffic event, analyzing and predicting a traffic state, evaluating a prediction result in real time, and carrying out event feedback and optimization decision response. The problem that accurate real-time detection and tracking cannot be achieved due to shielding, misrecognition and loss of targets when a vehicle moves at a complex urban intersection is solved.

The invention relates to the technical field of ion beam physical calculation, in particular to an ion beam optical property evaluation method based on a CPU-GPU hybrid parallel stable double-gradient conjugate algorithm. According to the technical scheme, the method comprises the following steps that a CPU-GPU hybrid parallel architecture is initialized, and CPU end and GPU end memory allocation, CUDA related object creation and GPU equipment initialization and parameter setting are completed; constructing a linear equation set, a sparse coefficient matrix A and a right-end vector b, and setting an initial solution vector, convergence precision epsilon and related parameters of the number of iterations; and analyzing the sparseness of the coefficient matrix A. Through combination of three core mechanisms of CPU-GPU cooperative calculation, dynamic load balancing and intelligent preprocessor selection, efficient, stable and universal evaluation of the ion beam optical characteristics is successfully realized, and an excellent solution is provided for solving the solving problem of a large-scale sparse linear equation set in the field of high-performance calculation.

The invention discloses an OpenMC-based BNCT dose full-process automatic calculation output method, and particularly relates to the technical field of boron neutron capture therapy dose calculation, and the method comprises the steps: reading a DICOM-CT image, automatically completing HU value segmentation, material assignment and voxel grid modeling, and generating an OpenMC geometry and material input card; setting a neutron / photon source, an operation mode and a counting grid, and executing Monte Carlo transport calculation; an ICRP flux-dose conversion coefficient library is built in, and dose calculation is automatically completed; the dose field is written into an Rt dose file, and the Rt dose file can be directly imported into mature medical image processing software for visual evaluation and plan optimization. According to the method, end-to-end automation from the CT image to the clinical available dose file is achieved, manual intervention is not needed, the BNCT dose calculation efficiency, precision and standardization level are remarkably improved, and a reliable tool is provided for BNCT dose simulation research.

The invention relates to the field of engineering management, and particularly discloses a quality management system and method for engineering construction, and the method comprises the steps: firstly carrying out the preprocessing of a BIM model, and constructing a reference frame integrating visual features and spatial geometric information; when a problem image shot by a mobile terminal at a construction site is received, a computer vision and space calculation technology is introduced, a two-dimensional field image is analyzed into an accurate pose of a shooting camera in a three-dimensional BIM space, and sight line tracing is automatically performed in a BIM digital model based on the accurate pose. And intelligently inferring and retrieving candidate components which are highly matched with the perspective of the problem. And finally, on-site unstructured problem data and the unique identifier of the BIM component are firmly bound through lightweight interaction confirmation with the user, and a standardized structured problem record is obtained. Therefore, the problems of low manual positioning efficiency and unstable association can be solved, and the whole-process intelligent closed loop of quality problem data from acquisition to structured storage is realized.

The application discloses a three-dimensional medical image visual language model pre-training method of reconstruction mixed strategy, belongs to the technical field of medical image computation, and comprises the following steps: constructing a medical image text pair data set, a language text mask reconstruction strategy, a visual image mask reconstruction strategy, a semantic perception fusion strategy and multi-task joint learning; the large language model is fine-tuned, the fine-tuned large language model is used to extract diagnosis and attribute information in a medical report and generate efficient prompts, the large language model has strong generalization ability, and the cost of manual labeling is greatly saved; the semantic perception fusion strategy is used for combining text features obtained by a text encoder and image features obtained by an image encoder to obtain new text features, so that the text perceives diagnosis and attribute information of the image in advance, the alignment of the image and the text in an embedding space is further optimized, and the pre-training efficiency is improved.

The invention discloses an optical flow calculation method, device and equipment based on comparative learning and a readable storage medium, and relates to the technical field of optical flow calculation. Comprising the following steps: firstly, performing feature extraction on an original image and a target image through a convolutional network to obtain an original feature map of the original image, background features of the original image and a target feature map of the target image; constructing a related body based on the original feature map and the target feature map; the related body is used for recording the pixel matching degree; then attention calculation is carried out on the background features to obtain background attention, and a 2D motion vector is extracted based on a correlation body; and finally, inputting the background attention and the 2D motion vector into an online network to obtain an optical flow calculation result. According to the method, the accuracy of an optical flow calculation result is improved.

The invention discloses a video face irreversible privacy processing method, edge computing equipment and a video face irreversible privacy processing system, and belongs to the technical field of video data processing and privacy computing. The method comprises the following steps: receiving YUV naked stream data at an edge side; performing face detection and behavior recognition, and determining a target face region and a behavior category; extracting a corresponding face feature vector; after the face feature vector is extracted, performing unrecoverable elimination on the original pixel data of the target face area immediately; performing irreversible hash operation on the face feature vector to generate an anonymous hash value; immediate unrecoverable removal is carried out on the face feature vector; and packaging the anonymous hash value and the behavior category information into a micro data packet, and uploading the micro data packet to a cloud server. According to the invention, through the irreversible hash and instant memory zero clearing technology, face pixels and feature information are destroyed at the edge side, and the cloud cannot be recovered, so that the compliance requirement of'unrecoverable 'is completely met.