895 results about "Adaptive algorithm" patented technology

The invention relates to a building design scheme multi-objective optimization comparison and selection method and device, equipment and a medium. The method comprises the steps of generating a multi-dimensional design parameter set by obtaining building information model data and parameterized design data; performing multi-dimensional target analysis and evaluation by using a multi-field joint simulation platform to generate a multi-dimensional evaluation index; a dynamic multi-objective optimization model is constructed through a dynamic weight adaptive algorithm in combination with project stage demands and user interaction data; carrying out iterative optimization by adopting an improved non-dominated sorting genetic algorithm to obtain an optimized design scheme gene sequence result, and introducing a spatial topology connectivity constraint to generate a Pareto optimal solution set; and according to the Pareto optimal solution set, generating an optimization scheme through user weight adjustment and scheme screening. According to the method, the optimal design scheme set meeting the project requirements can be quickly and efficiently generated and screened out, the project stage requirements and user preferences are met, and the design efficiency and the scheme quality are improved.

The invention relates to the technical field of crustal stress field data processing, in particular to a crustal stress field three-dimensional dynamic inversion method based on a multi-scale adaptive algorithm. The method comprises the following steps: acquiring a geological data set of a target area; constructing a crustal stress field three-dimensional initial model based on the geological data set, and performing geologic body space division and mesh generation to obtain crustal stress field three-dimensional mesh model data; performing multi-scale region division on the crustal stress field three-dimensional grid model data, and establishing a multi-scale weighting function to obtain multi-scale partition mapping information; and constructing a cross-scale boundary adaptive transmission mechanism, and establishing a stress tensor continuity constraint model at a multi-scale partition boundary to obtain cross-scale stress boundary coupling data. Through a multi-scale adaptive algorithm and dynamic closed-loop optimization, high-precision, dynamic and continuous inversion of a crustal stress field in a complex geologic structure is realized.

The invention provides a self-adaptive nonlinear image enhancement method and system for a low-light scene of a mobile terminal, and relates to the technical field of image enhancement, and the method comprises the steps: carrying out the image preprocessing and noise reduction, and carrying out the graying and noise suppression of an input color image through a local variance self-adaptive algorithm; adaptive down-sampling is carried out, and the down-sampling proportion is dynamically adjusted according to the image resolution and the content complexity, so that the processing efficiency is improved; brightness adaptive enhancement is carried out, and the overall brightness of the image is rapidly improved by adopting an Otsu method and a lookup table; contrast nonlinear enhancement: enhancing image details and contrast in combination with a Laplace operator and local mean adjustment; and color restoration: restoring the resolution through bilinear interpolation and performing weighted fusion to realize natural color reconstruction. And finally, a high-quality image of which the brightness, the contrast ratio and the color are remarkably improved is output. According to the invention, the recognition accuracy and processing efficiency of the low-illumination image are improved.

The invention provides an electric power and magnetic power dynamic joint detection method and system based on multivariable fusion, and relates to the technical field of power electronic equipment detection. According to the method, voltage, current and magnetic field sensors are arranged in a to-be-tested electromagnetic system, original physical quantity data are collected, and a multivariable time sequence data set is constructed; through preprocessing and feature extraction, a dynamic feature matrix related to electric power and magnetic power is generated; constructing a joint estimation model based on the matrix, and comparing the joint estimation model with a reference value to obtain a detection residual error; generating a state label according to the residual size and trend, and evaluating the estimation credibility in combination with a confidence interval; and a feedback mechanism is further constructed, model parameters and feature structures are optimized by adopting an adaptive algorithm, and online iteration and self-learning are realized. The method is high in detection precision, high in state recognition capability, good in adaptability and robustness, and suitable for energy flow monitoring and intelligent diagnosis of a complex electromagnetic system.

Systems and methods implement algorithms for initializing medical devices, including by programming initial medical device settings. Further, systems and methods implement adaptive algorithms for changing medical device settings. Settings initialization algorithms can calculate the settings most likely to be applicable for the patient at-issue using databases of positive clinical outcomes and respective therapy settings for existing users. Settings adaptation algorithm are configured to update medical device therapy settings at one or more intervals including based on databases of positive clinical outcomes and respective therapy settings for existing users.

The invention relates to the technical field of neural engineering and brain-computer interfaces, in particular to a closed-loop multi-modal nerve stimulation system and method based on time interference, and the system comprises a multi-modal stimulation module which is used for integrating electrical stimulation, magnetic stimulation and optical genetic stimulation, and generating a time interference field domain; the neural state sensing module is used for collecting real-time electroencephalogram signals, blood oxygen concentration and neural metabolite level data; the neural control center is used for fusing neural state data and stimulation parameters, and dynamically adjusting time interference frequency and stimulation intensity through an adaptive algorithm; the time sequence cooperation engine predicts a neural response time phase based on a deep learning model, and optimizes a stimulation time sequence and a mode switching strategy; and the visual interaction platform is used for rendering the nerve activation thermodynamic diagram and the stimulation parameter adjustment curve in real time. Therefore, the problems of adjustment strategy solidification, low adjustment precision, insufficient energy conversion efficiency and the like in the prior art are solved.

The invention relates to the technical field of roadway surrounding rock danger identification, and discloses a roadway surrounding rock danger identification model construction method, which comprises the steps of collecting multi-modal data, and generating preprocessed data through synchronous calibration and denoising; extracting a seismic wave frequency domain and image texture features, and generating a multi-modal feature matrix; in combination with a geological prior clustering mining abnormal mode, generating a labeled sample data set; generating a danger identification model based on a transfer learning and feature fusion training network; and the edge deployment model performs real-time reasoning, and generates an early warning result through an adaptive algorithm. According to the method, the frequency domain features of the seismic fluctuation signals and the depth texture features of the surrounding rock images are fused, the multi-modal feature matrix is constructed, abnormal mode mining is carried out in combination with geological prior knowledge, and early weak abnormal signals such as hidden fault slippage or asymmetric microfracture extension which are difficult to find by a single monitoring means can be effectively recognized.

The invention discloses a steam turbine safe operation health assessment method and system, and belongs to the technical field of steam turbine monitoring. The method comprises the following steps: acquiring running data of a steam turbine in real time through a multi-source sensor network deployed on key components of the steam turbine; the method comprises the following steps: preprocessing original operation data of a steam turbine by adopting edge computing equipment to generate a time domain waveform analysis diagram or a time sequence analysis diagram; transmitting the preprocessed structured feature data to an operation and maintenance management platform, inputting a fault diagnosis model for analysis, pre-training through a historical fault case library, and aligning feature distribution of the historical case library and current equipment data through a domain adaptive algorithm in transfer learning; and inputting a diagnosis result into a self-adaptive dynamic evaluation model based on a reinforcement learning engine, dynamically adjusting a model weight and an alarm threshold according to real-time working condition parameters, outputting a health score and a maintenance decision, and forming a closed-loop process of perception-edge processing-cloud diagnosis-dynamic optimization-decision output.

The invention discloses a transformer energy efficiency optimization cloud platform based on edge computing, and relates to the technical field of transformer energy efficiency management optimization. Aiming at the problems of data lag, protocol heterogeneity, response delay, security risk and the like in traditional energy efficiency management, an innovative architecture of edge intelligence, protocol standardization and cloud edge collaboration is provided; the platform deploys a lightweight AI model through edge nodes, processes data in real time in combination with a dynamic quantization and adaptive algorithm, and realizes anomaly detection and local decision; the cloud performs strategy simulation and verification based on a digital twin model, generates an optimization strategy through Monte Carlo tree search, pushes the optimization strategy to the edge for execution after signature encryption, and dynamically adjusts the priority of the strategy in combination with reinforcement learning; a closed-loop feedback system is constructed, and cloud side cooperation of energy efficiency analysis, load optimization and fault early warning is realized; according to the method, the operation efficiency and safety of the transformer are remarkably improved, the energy consumption and the operation and maintenance cost are reduced, and rapid access of heterogeneous equipment and sustainable expansion of the system are supported.

The invention discloses a control method of an electric actuating mechanism. The electric actuating mechanism is used for driving a valve to be opened and closed. The control method comprises the steps that real-time operation state parameters of the electric actuating mechanism are collected; determining a parameter control target value of the electric actuating mechanism according to a preset control mode; extracting historical operation data of the electric actuating mechanism, and outputting a control signal on the basis of a set target value according to a preset adaptive algorithm in combination with the historical operation data and the current operation state parameter; and driving the electric actuating mechanism based on the control signal. According to the technical scheme, the control precision of the electric actuating mechanism on the valve can be effectively improved.

The invention provides a photovoltaic energy storage inverter power coordination control method based on an adaptive algorithm, and relates to the technical field of data processing, and the method comprises the steps: collecting the output power data of a photovoltaic module, the state data of an energy storage device, and the load demand data of a grid connection point, and carrying out the normalization processing of the data; calculating to obtain an initial power distribution parameter of the current period; respectively calculating a load change trend factor and a power fluctuation trend factor based on the load demand data and the output power data of the historical period, and coupling the load change trend factor and the power fluctuation trend factor to form a trend coupling factor; carrying out self-adaptive repair through a trend coupling factor; combining real-time grid-connected point load demand data with energy storage device state data to generate a target power distribution instruction and an energy storage charging and discharging adjustment instruction, and respectively issuing the instructions to an inverter control unit and an energy storage device control unit; according to the invention, the accuracy of the photovoltaic energy storage inverter power coordination control method is improved.

The invention relates to a network attack-oriented cluster area coverage collaborative search path planning method, which comprises the following steps of: regarding each unmanned aerial vehicle as a node, iteratively updating a state value, and modeling an unknown task area which needs to be subjected to coverage search; the node state values are sorted and compared, and normal neighbor node state values are reserved and updated; an adaptive MSR algorithm and an MPC-PSO-based path planning method are fused through a communication topology robustness constraint condition to form an anti-attack closed-loop control framework, and an optimal path decision at the current moment is made; according to the method, the relation between the number of neighbor nodes and communication topology robustness is found, and in multi-objective optimization of path planning, the lower limit of the number of the neighbor nodes serves as a hard constraint and is directly associated to a return function. The problems of how to defend network attacks and how to overcome path optimization and coverage search robustness insufficiency of the unmanned aerial vehicle cluster are solved, and it is ensured that state consistency and stable cooperative control can still be achieved when the unmanned aerial vehicle cluster is subjected to the network attacks.

The invention relates to the technical field of aerial video analysis, and discloses a large-scene multi-modal analysis system and method based on an unmanned aerial vehicle. According to the method, an original video frame sequence is obtained from aerial photography equipment, a scene change period and an object motion trend are extracted by calculating inter-frame motion feature vectors, and whether a scene is in a stable or dynamic state is judged by means of a state classification model. Calculating a tracking target position according to preset parameters in a dynamic state, and adjusting and analyzing a target area in combination with a motion trend; generating a camera parameter adjustment instruction according to the deviation value of the target area and the actual object position, and outputting the camera parameter adjustment instruction to an execution mechanism; acquiring a complexity value through the scene complexity feedback signal, updating an analysis strategy by adopting an adaptive algorithm when the complexity value exceeds a preset range, and outputting an adjustment signal; the output time sequence of control and strategy adjustment signals is optimized, meanwhile, the frame sequence is monitored in real time, the parameter change trend is analyzed, the preset parameters are dynamically updated to meet the user requirements, and efficient video analysis processing is achieved.

The invention belongs to the technical field of three-dimensional modeling, and particularly relates to an intelligent operation and inspection method and system for transmission, transformation and distribution equipment based on a large model. Comprising the steps of region division, path optimization, three-dimensional modeling, risk prediction and the like. According to the method, geographic information of transmission, transformation and distribution equipment is acquired, and an adaptive algorithm is utilized to divide inspection areas and optimize an unmanned aerial vehicle inspection path; a high-precision three-dimensional point cloud model is constructed through a laser radar and image data, and geometric features and appearance defect features of equipment are extracted; performing multi-modal feature fusion based on the large model, and outputting defect severity and risk confidence; and in combination with semantic analysis of the construction file, space matching and early warning of potential risks are realized. According to the method, intelligentization, visualization and predication of the operation and maintenance process of the transmission, transformation and distribution equipment can be realized, and the safety of equipment operation and maintenance and the scientificity of management decision are improved.

The invention provides a new energy light truck energy distribution method and system, and the method comprises the steps: obtaining the position and track of a vehicle to obtain the road section data of a current driving road section, carrying out the data preprocessing of the road section data, and carrying out the fusion of the preprocessed data, so as to obtain the current load and driving state of the vehicle; inputting the preprocessed data into an initial energy consumption model, introducing a deep learning algorithm to dynamically adjust model parameters of the initial energy consumption model to obtain an iterated target energy consumption model, and obtaining energy consumption coefficient data according to the target energy consumption model; and a motor torque demand curve is generated in real time according to the current load, the road slope and the predicted endurance mileage, torque distribution data are adjusted in combination with an adaptive algorithm, the adjusted torque distribution data are transmitted into a vehicle power control system to obtain real-time operation data of the vehicle, and then energy distribution is conducted on the new energy light truck. According to the invention, the endurance prediction precision is improved, and the real-time adaptive optimization of torque distribution control is realized.

According to the mass center self-balancing chassis system for U-rib weld defect detection and the control method, the problem that when detection equipment moves on a U-rib curved surface, the precision is reduced due to mass center deviation and view field inclination can be solved. The system integrates a Mecanum wheel omni-directional moving unit, four groups of electric telescopic supporting legs, a two-axis holder and an ECU (electronic control unit), and has the innovation point that dynamic mass center balance and posture self-correction are realized through rigid-flexible combined supporting leg design and multi-modal sensing fusion. The ECU adopts an improved Canny algorithm to process the image, extracts a U rib edge point set and fits a reference center; in combination with pressure and tilt angle sensor data, a real-time mass center is solved through a moment balance formula, and then an improved self-adaptive PID algorithm is used for driving the supporting legs to stretch out and draw back, so that the mass center returns to a reference center; the two-axis holder compensates the view field angle according to the inclination angle data to ensure that the view field angle is perpendicular to the weld joint. Through cooperation of a mechanical structure and an algorithm, the system can control centroid offset to be smaller than or equal to 5 mm and view field inclination to be smaller than or equal to 1 degree, the detection stability and precision are remarkably improved, and the algorithm is easy to deploy.

The invention relates to an unmanned aerial vehicle autonomous path navigation system based on multi-modal fusion and adaptive optimization and an algorithm thereof. The system comprises a multi-modal sensor module, a data fusion processing module, an adaptive algorithm processing module and a flight control module. The method comprises the following steps: step 1, data acquisition; step 2, data fusion; step 3, algorithm processing; and 4, controlling the flight of the unmanned aerial vehicle. The method has the advantages that multiple sensors such as a visual camera, a laser radar and a millimeter wave radar are integrated through deep fusion of multi-modal sensors, a fusion algorithm based on deep learning and a Transform architecture is adopted, weights are automatically distributed through an attention mechanism, deep data fusion is achieved, and the accuracy and accuracy of data fusion are improved. Compared with a single sensor or a simple data fusion mode, the problem that the single sensor is easily interfered by the environment is effectively solved, and the obstacle recognition accuracy and the environment modeling precision are greatly improved under the complex environments of low illumination, strong smoke, electromagnetic interference and the like.

The invention relates to a complex manufacturing system optimization decision support method, which comprises the following steps of: collecting multiple types of data through distributed sensing nodes, fusing the data into a unified data model through semantic mapping, and constructing a dynamic data graph; key features are mined by using an improved deep learning model, and a multivariable prediction model is constructed to realize advanced prediction of equipment faults and the like; constructing a multi-objective optimization model, and generating and dynamically sorting a Pareto optimal scheme by adopting an adaptive algorithm; and issuing the scheme to an execution system, tracking execution and quantifying an evaluation result, and feeding back the evaluation result to a platform to optimize model parameters and feature weights to form closed-loop iteration. The invention aims to solve the problems of low data integration level, insufficient prediction precision, poor decision scheme adaptability and difficulty in continuous optimization in a complex manufacturing system.

The invention discloses a rolled piece temperature control system based on a steel rolling process, which relates to the technical field of steel rolling and comprises a detection module, a control module, an execution module and a cooling module, the detection module comprises infrared thermometer arrays, an ultrasonic thickness gauge, an electromagnetic flowmeter, a pressure transmitter and a thermal resistance thermometer, and the infrared thermometer arrays are used for being arranged in front of and behind a cooling section and respectively collecting initial temperatures before entering a cooling system. Through the self-adaptive PID algorithm, the finite difference heat conduction model and the Smith pre-estimation compensation strategy are combined, the problem of control delay caused by the large lag characteristic in the cooling process is solved, control parameters can be dynamically adjusted according to the rolling speed, the steel grade characteristic and the like, temperature adjustment better fits the actual working condition, and the control precision is improved. The phenomenon of supercooling or insufficient cooling in a traditional control mode is fundamentally avoided, and the uniformity and stability of the performance of rolled pieces are guaranteed.

The invention relates to the technical field of intelligent control of injection molding machines, in particular to an injection molding machine closed-loop mold adjustment control system and method based on multi-sensor fusion and a self-adaptive algorithm, and the system comprises a sensing layer which adopts a multi-sensor fusion closed-loop feedback architecture and comprises a pressure sensing array, a displacement sensing unit and a temperature sensing array. The data acquisition module is used for synchronously acquiring mold clamping force, position and temperature data of a mold, constructing a three-dimensional error compensation model and realizing dynamic correction of the parallelism of the mold through space force field analysis; the execution layer comprises a servo motor, a belt wheel and a ball screw and is used for adjusting the position of the template; a double-closed-loop control module is arranged in the control layer, the double-closed-loop control module comprises outer ring mold clamping force control and inner ring position control, mold clamping force deviation serves as input of an outer ring, position correction is generated through a self-adaptive PID algorithm, the inner ring converts the position correction into a servo motor rotation angle instruction, and therefore precise control over the mold clamping force is achieved.

The invention relates to the technical field of biological breeding, in particular to a rice hybrid combination optimization method based on an adaptive algorithm, and aims to solve the problems that existing rice breeding is poor in adaptability, multi-objective optimization is difficult and artificial experience is limited. The method is characterized in that a data integration module, a dynamic target module, a parent library module, a genetic prediction module, a self-adaptive optimization module (including a genetic algorithm and reinforcement learning), a simulation evaluation module and a recommendation feedback module are constructed, and precise prediction and dynamic optimization of hybrid combination are achieved. By adopting the scheme, the limitation of artificial experience is effectively overcome, the breeding decision is changed from static experience to dynamic data drive, and the breeding efficiency, accuracy and adaptability are remarkably improved.

The invention discloses a liquid crystal display filtering and light transmission performance monitoring system and a monitoring method thereof, and relates to the technical field of liquid crystal displays. An ambient light real-time monitoring module; a dual-optical-path differential measurement module; a signal processing and intelligent compensation module; a polarizer dynamic adjustment module; a data storage and analysis module; light intensity fluctuation data are collected in real time through an ambient light real-time monitoring module, reference light and transmission light are synchronously collected through a double-light-path differential measurement module, a signal processing and intelligent compensation module calls a self-adaptive algorithm library, compensation coefficients are calculated in combination with ambient light spectrum data, and a polaroid adjusting instruction is generated. And then an adjusting instruction is sent to the polaroid dynamic adjusting module through the CAN bus, the polaroid dynamic adjusting module feeds back an actual thickness value in real time after adjustment is completed, closed-loop control is formed, and through real-time compensation of ambient light and dual-light-path differential measurement, the influence of a station LED lamp and natural light is eliminated, and the monitoring precision of light filtering and light transmission performance is improved.

The invention provides an insulating oil gas mass spectrum analysis method and system. The method comprises the following steps: firstly, circulating vacuum degassing for one time, then deoiling and dehydrating the removed gas, introducing the deoiled and dehydrated gas into a mass spectrometry detection unit, carrying out primary detection, calling a self-adaptive algorithm and selecting proper circulating vacuum degassing times based on a primary detection result in combination with degassing temperature and degassing pressure of circulating vacuum degassing, and entering second-stage vacuum degassing; and deoiling and dehydrating the deoiled gas, and introducing the deoiled gas into a mass spectrometric detection unit for secondary detection to obtain a final detection result. According to the method, effective fusion of efficient degassing and rapid detection is realized, and reliable technical support is provided for rapid and accurate online monitoring and real-time fault diagnosis of the insulating oil dissolved gas.

The invention relates to the technical field of virtual impedance self-adaptive control of energy storage converters, in particular to a virtual impedance self-adaptive control method of a grid-forming type energy storage converter, which comprises the following steps of: acquiring three-phase voltage and current of a grid-connected point in real time; performing phase-locked loop processing and coordinate transformation to obtain a d-axis voltage, a q-axis voltage and a current under a synchronous rotating coordinate system; calculating instantaneous power, and obtaining active power and reactive power; virtual resistance and virtual inductance are calculated according to power fluctuation through an adaptive algorithm; the virtual impedance equation generates a voltage drop compensation signal; and superposing the compensation signal and the reference voltage to generate a new modulation wave, and inputting the new modulation wave into the pulse width modulation module to realize power oscillation suppression and stable operation. According to the method, the problems between damping enhancement and loss reduction in network construction type control are solved, the dynamic and steady-state performance of the converter is improved, and a key technical support is provided for large-scale and high-reliability application of the network construction type energy storage converter in a novel power system.

The invention discloses an AI development optimization system based on dynamic learning and an adaptive algorithm, and relates to the technical field of artificial intelligence, and the system comprises a collection module which collects GPU occupation data and transmission channel state data, and generates a separation operation signal; the identification module is used for triggering operation task analysis according to the separation operation signal and distributing a transmission channel identifier for the subtask; the path module queries the idle state of each computing unit of the GPU and generates an exclusive communication path; the detection module is used for running the subtasks in the exclusive communication path and outputting channel signal mixing degree; the triggering module is used for setting a mixing degree threshold value and triggering a path replacement signal and a merging operation signal; and the merging module suspends the sub-task processing process based on the merging operation signal, performs merging according to the operation sequence of the sub-tasks, and outputs a complete task. According to the method, efficient resource scheduling and task optimization of AI development are realized through coordination of dynamic task topology analysis and self-adaptive communication path binding.

The invention relates to the field of computing power scheduling, and discloses an artificial intelligence-based computing power scheduling method, which comprises the following steps of S1, full-dimensional data acquisition: acquiring hardware resource data, task operation data, user demand data and heterogeneous computing power characteristic data through distributed sensing nodes; s2, data processing and feature construction: preprocessing the data collected in the S1, extracting multi-dimensional features through feature engineering, and constructing a feature matrix adapted to an AI model; s3, computing power demand intelligent prediction: learning the feature data preprocessed in the S2 by using a federal deep learning model, and outputting time-phased and type-divided computing power demand prediction results; and S4, scheduling strategy optimization generation. Hardware resources, task operation, user requirements and heterogeneous computing power characteristic data are comprehensively collected by relying on edge-core two-level distributed sensing nodes, the collection frequency can be dynamically adjusted between 50ms and 1mi n through a self-adaptive algorithm, and real-time performance and resource economy are both considered.

The invention provides a vehicle path planning method based on a multi-head adaptive Actor-Critic algorithm, and the method comprises the steps: obtaining urban road network data comprising at least one warehouse center and more than two demand client nodes through a simulation generation technology, obtaining a training and testing data set, carrying out the coding processing of the urban road network node training data, and carrying out the coding processing of the urban road network node training data, obtaining a multi-dimensional vehicle path coding sequence on a two-dimensional plane space [0, 1] * [0, 1]; a deep reinforcement learning framework is built based on the multi-dimensional vehicle path coding sequence, a multi-head adaptive Actor-Critic algorithm is integrated to build an MHAAC model, and a vehicle path planning system is formed; and the constructed MHAAC model is adopted to carry out path solving on urban road network data, and a global optimal path scheme is generated under the condition that all customer demand constraint conditions are met. According to the scheme, the problems of poor adaptivity, prolonged solving time, reduced solution quality and the like caused by limitations of sparse environmental information features, fixed feature embedding, static parameter generation, single decoding strategy and the like are successfully solved.

The invention discloses a multi-stage graded crushing and particle size homogenization control method for ferrosilicon crushing. The method comprises the following steps: S1, raw material characteristic perception and pretreatment; s2, self-adaptive multi-stage crushing; s3, performing dynamic coupling classification screening; s4, performing intelligent closed-loop optimization; according to the method, the problem of granularity fluctuation caused by parameter adjustment lag in a traditional control method is avoided through the self-adaptive PID algorithm, the pertinence and the real-time performance of crushing parameter adjustment are remarkably improved, and an accurate parameter basis is provided for granularity gradient control; the grading and crushing parameter coupling model realizes dynamic matching of the grading efficiency and the crushing load, so that the process redundancy caused by invalid material returning is reduced while the granularity screening precision is ensured; the equipment wear compensation machine learning model can accurately quantify granularity drift caused by equipment characteristic difference in different crushing stages, and the long-term stability of granularity prediction is improved.

The invention discloses a spatial deformation analysis method and system based on unmanned aerial vehicle surveying and mapping and a set unmanned aerial vehicle surveying and mapping system, and the system comprises a data acquisition module S1, a preprocessing module S2, a deformation analysis engine S3, a visual early warning module S4, a report generation module S5 and a system management module S6. Space reference precise transmission, error space distribution modeling, multi-source data space fusion and cross-scale three-dimensional field construction are realized, and'dynamic environment anti-interference 'and'error space-time propagation control' are broken through, so that all-weather credible perception of geologic body millimeter-level deformation is realized. According to the spatial deformation analysis method and system based on unmanned aerial vehicle surveying and mapping, through triple optimization of multi-source sensing fusion, error spatial modeling and a self-adaptive algorithm, the defects of traditional unmanned aerial vehicle deformation monitoring in the aspects of complex environment adaptability, precision reliability and automation degree are overcome, a long-time-sequence error propagation model and edge calculation acceleration are broken through, and the method and the system have good application prospects. Therefore, millimeter-level near-real-time early warning of geological safety risks is realized.

The invention discloses a fuzzy video real-time jitter removing method and system, and the method comprises the steps: distributing feature points through multi-scale FAST detection, lightweight MLP detection and dynamic weight, combining a MobileNet-FPN network and the lightweight MLP dynamic weight to extract the robust features of the feature points, modeling the global matching relation of the feature points of two adjacent frames through Transform, and carrying out the real-time jitter removing of a fuzzy video. And a Bayesian adaptive RANSAC algorithm is adopted to realize high-efficiency mismatching pair elimination, and finally an affine transformation matrix is solved and reverse mapping is performed to compensate jitter. According to the method, while the real-time performance of the mobile terminal is ensured, the image stabilization precision in a violent shaking scene is remarkably improved.