1794results about How to "Improve environmental adaptability" patented technology

The invention relates to the technical field of server scheduling, and discloses a large model reasoning scheduling method based on an off-network computing power server, which comprises the following steps of: collecting hardware information of the off-network computing power server, and constructing a server resource vector; decomposing a large model reasoning task into a plurality of sub-tasks, and constructing task modeling; according to the hardware information of the off-network server and the task dependency relationship, performing task division by using an optimization method; performing dynamic task scheduling on the sub-tasks based on the current load state of the server; according to the real-time calculation load of the server, dynamically migrating part of tasks to other servers to ensure the load balance among the servers; and executing the distributed task, monitoring the running state of the server and the task completion condition, and dynamically adjusting a task distribution strategy according to feedback data. Through modeling task requirements and server resources, in combination with real-time monitoring and optimal scheduling, efficient calculation and resource utilization maximization of large model reasoning in an off-network environment are realized.

The invention discloses a fatigue driving monitoring and early warning method based on multi-dimensional feature fusion, and relates to the field of intelligent traffic and safe driving, and the method comprises the steps: carrying out the feature extraction of a collected face image of a vehicle driver through a feature extraction model based on a convolutional neural network, and obtaining a face key point; then calculating the eye closing rate, the mouth opening rate, the head posture data and the facial muscle data to finally obtain multi-dimensional features, and fusing the multi-dimensional features by using a multi-layer perceptron to obtain fused facial features; and then, a fatigue state detection model obtained based on recurrent neural network training is utilized to detect the facial feature time sequence at a plurality of continuous sampling moments so as to detect whether a fatigue driving state occurs and trigger alarm feedback. According to the method, the fatigue state of the driver can be identified and judged more accurately through combination of multi-dimensional data fusion and a time sequence analysis algorithm, and the monitoring and early warning accuracy and reliability are improved. And the system can keep high robustness even in an illumination change or complex driving environment.

The invention provides a hybrid representation learning-based omni-directional motion control method for a quadruped robot in a three-dimensional environment, aiming at solving the problem that an existing motion control method for the quadruped robot is limited by a narrow forward view angle and external perception noise of the robot, so that the omni-directional motion capability in the three-dimensional environment is difficult to obtain. Comprising the steps of designing various input and output contents of a first-stage end-to-end neural network learning framework, designing a training process including mixed representation learning based on a reinforcement learning framework, constructing a specific neural network model of each module, and constructing a quadruped robot simulation motion control training environment. The method has the advantages that network input and output content beneficial to efficiently learning the relation between actual observation of the robot and the environment truth value is provided, and a mixed representation learning method including supervised learning and comparative learning is introduced, so that the robot can effectively deduce environment information beyond the visual field range of the robot in the training process.

The invention provides a task allocation and optimal scheduling method and system for a multi-robot collaborative industrial assembly line, and relates to the technical field of robots, comprising the steps of constructing a task allocation strategy network through deep reinforcement learning, encoding workpiece types and attribute information into a state vector input network, a multi-objective optimization function including total assembly time minimization, load balance degree maximization and energy consumption minimization is designed, a near-end strategy optimization algorithm is adopted to train network parameters, a robot task mapping matrix is output as a distribution scheme, and the network parameters are dynamically optimized and updated in an online learning mode. The task allocation efficiency of the multi-robot collaborative assembly production line can be remarkably improved, the dynamic optimization scheduling of the production line is realized, the system energy consumption is reduced, and the assembly efficiency is improved.

The invention relates to the field of power system safety monitoring, and discloses a power transmission line vibration monitoring and early warning method for dealing with extreme weather, which comprises the following steps: arranging a multi-level sensor network to monitor vibration and environmental conditions of a power transmission line; monitoring data are processed and transmitted in real time through edge calculation and a 5G network; performing trend prediction on the vibration data by using a multi-modal deep learning model; an early warning threshold value is adjusted through a distributed self-adaptive dynamic threshold value control method; and when abnormal vibration or a high-risk trend is detected, generating and issuing early warning information. Through a multi-level sensor network, an intelligent early warning algorithm and dynamic threshold control, efficient monitoring of power transmission line vibration and accurate early warning in extreme weather are realized, the real-time performance of data transmission and processing is effectively improved, the accuracy of vibration trend prediction is enhanced, the false alarm rate and the missing report rate are reduced, and the method is suitable for popularization and application. Therefore, the method has higher adaptability and stability in a complex environment.

The invention provides a low-altitude industry chain information intelligent cooperation and airspace management platform based on federal learning. Through cluster ad hoc network and decentralized federated learning, the method is divided into three stages of data acquisition, data preprocessing and analysis application, low-altitude scene data communication of dynamic topology and static topology is realized, and the efficient communication requirement in a low-altitude operation scene is met. According to the system, differential data transmission strategies are adopted, optimization design is carried out on a single machine, a co-task cluster, a cooperative cluster and a non-cooperative cluster, and the communication efficiency and the resource utilization rate are improved. In the data preprocessing stage, federated learning aggregation of multi-source data fusion, data anonymization and decentralization is combined, and data privacy protection and safe sharing of three parties of industry, university and research are ensured. In the analysis and application stage, low-altitude aircraft task scheduling, airspace flow control and environmental condition evaluation are realized through aircraft analysis, airspace analysis, environmental analysis and digital twinborn technologies.

The invention relates to a remote-control system of a stratospheric airship and a control method thereof. The control system is divided to four levels: a decision-making level for executing the selection and switch of a remote-control flight mode according to the actual working state data, and sending a corresponding mode command; a planning level for dynamically planning and generating an airship motor drive command according to the mode command sent by the decision-making level; an executing level for realizing the drive control to an airship motor according to the motor drive command of the planning level; and a perceiving level for finishing the real-time monitoring and data collection of the airship working state, and providing the original data of the airship working state to the decision-making level. The four-level control system architecture design of the control system is capable of simplifying the control device, improving the control efficiency, and fundamentally solving the problems of the function, efficiency and reliability of the near space airship ground and airborne control system.

The invention discloses an unmanned aerial vehicle target intelligent identification and perception system and method based on deep learning, and relates to the technical field of artificial intelligence, and the method comprises the steps: extracting the importance feature of each target region, calculating a region importance score, generating a target region monitoring plan, and obtaining continuous frame images; an improved YOLOv5 model is adopted, continuous frames of images are input, a bounding box of a suspected target wild animal is output and mapped to an actual geographic coordinate system, and actual position coordinates are obtained; key points of the t-th frame image are detected, coordinates of the key points are extracted, visibility scores and confidence coefficients are calculated, and feature descriptors are formed; performing target detection and key point detection on the (t + 1) th frame of image to obtain actual position coordinates and feature descriptors of candidate targets, performing track matching, updating the actual position coordinates and the feature descriptors of the suspected target wild animals if the actual position coordinates and the feature descriptors are matched, and defining new target animals if the actual position coordinates and the feature descriptors are not matched; counting according to the track matching result to obtain the total number of the target wild animals.

The invention relates to the technical field of air pressure intensity adjustment, and discloses a pressure control method and device for accurately adjusting air pressure intensity, and the method comprises the steps: carrying out the digital conversion and temperature and humidity compensation of an original pressure signal collected by a pressure sensing array, and obtaining a pressure data matrix and a pressure gradient matrix; calculating a pressure distribution mass center and a pressure fluctuation frequency spectrum to obtain a pressure space distribution characteristic matrix and a pressure time domain characteristic matrix; inputting a three-state Markov model for dynamic state identification to obtain a pressure state transition probability matrix and a pressure response prediction matrix; constructing an objective function containing a pressure tracking error term and an organization safety constraint term, and solving to obtain a PID control parameter matrix and a pressure safety limit matrix; and constructing a three-stage control system comprising a basic controller, an interference compensator and a safety protector, and outputting an air pressure regulation driving signal. According to the method, high-precision pressure sensing control is realized.

The invention discloses a wireless spectrum intelligent allocation and edge computing cooperation method, which relates to the technical field of mobile edge computing and wireless communication, and comprises the following steps: acquiring spectrum occupation data and service load data; constructing a spectrum resource allocation matrix according to the spectrum occupation data, and constructing an edge computing resource allocation matrix based on the service load data; and inputting the spectrum resource allocation matrix and the edge computing resource allocation matrix into a trained deep reinforcement learning model, outputting a joint optimization strategy, and performing collaborative scheduling of spectrum resources and computing tasks through a bidirectional matching algorithm. According to the invention, a multi-dimensional resource data acquisition mechanism is established, so that real-time monitoring and evaluation of a network resource state are realized; quantitative expression and association mapping of heterogeneous resource characteristics are realized by constructing a mathematical model of a spectrum resource allocation matrix and an edge computing resource allocation matrix; a deep reinforcement learning model is adopted for strategy optimization, and the limitation of a traditional fixed rule algorithm is overcome.

The invention discloses a water level prediction model construction method based on image recognition, and relates to the technical field of water level monitoring. Comprising the following steps: establishing a visual reference system of a region, constructing a multispectral image data fusion module, constructing a water surface edge detection module, designing an anomaly detection mechanism based on self-supervised learning, constructing a water level height estimation model based on image semantic segmentation, and constructing a water level distribution prediction model combined with geographic information. A data feedback model based on meteorological dynamics is constructed, and a water level trend estimation module based on short-term time sequence prediction is introduced. According to the method, a water level prediction model is constructed by introducing technologies of multispectral image fusion, semantic segmentation and dynamic height estimation, and accurate extraction and prediction of water level information are realized under various environmental conditions. The robustness of the model under complex weather conditions is remarkably improved, and meanwhile, the problem of deviation caused by different visual angles and optical distortion is solved through dynamic adjustment factors and a geometric correction mechanism.

According to the coastal saline-alkali land underground water level adjusting system based on real-time monitoring and intelligent regulation and control, the system integrates the soil salinity sensor, the underground water level sensor and the central processing unit, future trend prediction is carried out by monitoring key parameters such as soil salinity and underground water level in real time in combination with a deep learning model, and the real-time monitoring and intelligent regulation and control are achieved. And precise intelligent regulation and control are realized. The system can effectively reduce soil salinization, prevent seawater from flowing backwards, improve the utilization efficiency of water resources and reduce the labor management cost at the same time. The method is characterized in that the application of multi-parameter comprehensive monitoring, a deep learning prediction model and a self-adaptive intelligent regulation and control algorithm provides a new technical solution for coastal saline-alkali land treatment.

The invention provides a ground surface settlement monitoring method and system fusing Beidou and InSAR data. The method comprises the following steps: receiving positioning data from a Beidou satellite system and earth surface deformation data acquired by a synthetic aperture radar interference measurement system; based on the positioning data and the earth surface deformation data, correcting the earth surface deformation data according to the environmental factors through a space-time fusion model to obtain corrected earth surface deformation data; by means of the corrected surface deformation data, a multi-stage analysis technology is adopted, the development trend of surface subsidence is determined, a surface subsidence distribution diagram is generated, and the surface subsidence distribution diagram is used for displaying spatial distribution characteristics of surface subsidence; and according to the ground surface settlement distribution diagram, the future ground surface settlement condition is monitored by using the space-time fusion model, and a ground surface settlement monitoring report can provide a scientific basis for early warning and prevention of geological disasters. The ground surface settlement monitoring efficiency and monitoring effect are improved.

The invention relates to the technical field of digital interaction, in particular to a digital interaction enhancement system based on multi-modal voice, which comprises a multi-modal sensing module comprising a voice sensing unit, a visual sensing unit and an environment sensing unit and used for capturing multi-modal data input by a user; the multi-modal fusion module comprises a data synchronization unit, a feature extraction unit and a multi-modal fusion unit and is used for performing time synchronization, feature extraction and fusion processing on the multi-modal data; and the interactive understanding module comprises a natural language processing unit, a context understanding unit and an emotion analysis unit and is used for understanding the intention and the emotion state of the user. The deep learning technology is adopted to fuse voice, vision and environment data, and efficient and natural user interaction experience is achieved through the steps of feature extraction, feature fusion, intention recognition and the like. The system has excellent performance and a wide application prospect in a complex environment.

The invention relates to the field of environment monitoring, and discloses a mobile VOCs rapid detection and positioning system, which comprises a data acquisition module, a data fusion module, a concentration gradient analysis module, a Bayesian filtering module and a particle filtering optimization module, the data acquisition module is used for acquiring VOCs concentration data and multi-source positioning data; the data fusion module outputs high-precision position information for the positioning data; the concentration gradient analysis module is used for calculating a concentration gradient based on the VOCs concentration data and meteorological information and simulating concentration field distribution; the Bayesian filtering module dynamically updates the probability distribution of the pollution source position in combination with the concentration field distribution and the fusion position information; and the particle filtering optimization module optimizes the Bayesian filtering result. The system is suitable for various complex environments, can realize rapid detection of VOCs and high-precision positioning of pollution sources, and has good real-time performance, precision and environmental adaptability.

The invention discloses a gas leakage source positioning system based on multispectral imaging, which comprises a gas leakage positioning supervision platform, and the gas leakage positioning supervision platform is in communication connection with a holder data acquisition unit, a gas leakage centering analysis unit, a gas data comprehensive processing unit and an early warning information linkage unit. According to the invention, the accurate position of the leakage source is positioned and calculated by realizing omnibearing scanning of the target area and comparing and converting the parallax among the spectral imaging display and the value of the overlapping area, and the positioning result is fed back on the user interface for display, so that the accuracy and environmental adaptability of gas leakage detection are improved, and the detection efficiency is improved. Rapid detection and accurate positioning of dangerous chemical gas leakage can be realized, the reliability of positioning a leakage source is improved, and the emergency response capability of a petrochemical park is effectively improved.

A foldable pushing chain type lifting platform belongs to the technical field of pushing chains and lifting platforms and comprises a meshed pushing chain system, an X-axis limiting mechanism and a Y-axis limiting mechanism, one side of the bottom of the meshed pushing chain system is fixedly connected with a base, the meshed pushing chain system outputs pushing chains upwards in the Z-axis direction, and the X-axis limiting mechanism and the Y-axis limiting mechanism are fixedly connected with the base. The top end of the pushing chain is fixedly connected with a top table used for bearing a lifted object, an X-axis limiting mechanism is further connected between the top table and the base, and the side, away from the meshing type pushing chain system, of the X-axis limiting mechanism is connected with a Y-axis limiting mechanism. The X-axis limiting mechanism is of a folding arm structure which only rotates around the X axis and limits the pushing chain and the top table to move in the X-axis direction. According to the invention, the pushing chain is used as a driving part, and the device has the advantages of high speed, large bearing capacity, high precision, long maintenance period, multi-point control, good stability, low initial height, large stroke, small occupied space and strong environmental adaptability.

The invention relates to an intelligent photovoltaic power generation control device which integrates a photovoltaic power generation panel, a series connection base, a supporting assembly, a water collection groove, a rainwater collection assembly, a cleaning assembly and a control assembly. A series base and a series interface are arranged below the photovoltaic panel, so that modular installation is facilitated; the supporting assembly comprises two supporting rods, the light following function is achieved, and the photovoltaic panel is automatically adjusted to face the sun. A water collecting groove above the photovoltaic panel and a rainwater collecting assembly at the bottom of the photovoltaic panel can collect and store rainwater; the cleaning assembly comprises a spraying unit and a cleaning brush and is used for automatically cleaning the surface of the photovoltaic panel; the control assembly monitors operation of the system and adjusts the posture of the photovoltaic panel and work of the cleaning assembly. Redundant electricity generated by the photovoltaic power generation panel in sunny days is stored for the system to use, and the light following function is not started in cloudy days, so that electric energy is saved; through the design, the efficiency and sustainability of photovoltaic power generation are improved; the energy conversion efficiency can be improved, the reliability and automation level of the system are enhanced, and the utilization and environmental friendliness of rainwater are improved.

The invention discloses a broadband wide frequency spectrum analysis method and system based on deep learning. The method comprises the steps of spectrum signal acquisition, spectrum preprocessing, abnormal interference detection, spectrum trend prediction and spectrum scheduling management. The invention relates to the technical field of wide-frequency spectrum analysis, in particular to a broadband wide-frequency spectrum analysis method and system based on deep learning, which adopts a three-step method of abnormal interference detection, spectrum trend prediction and spectrum scheduling management, and adopts a layer-by-layer progressive deep learning and reinforcement learning technology. The real-time detection of spectrum abnormity, the accurate prediction of future trend and the optimal allocation of dynamic resources can be realized; carrying out abnormal interference detection by adopting a joint improvement multi-classification support vector machine method combined with feature enhancement; performing spectrum trend prediction by adopting a time sequence integrated network combined with one-dimensional convolution feature extraction; and performing spectrum scheduling management by adopting a reinforcement learning method enhanced by a near-end strategy optimization algorithm.

The invention provides a photovoltaic power generation power prediction method, system and device and a storage medium, and relates to the technical field of photovoltaic power generation power prediction, and the method comprises the steps: collecting historical data and real-time data of a photovoltaic array region; performing synchronous preprocessing to generate a historical data set and a real-time data set; establishing an initial generation power model, and drawing a curve graph of generation power changing along with time; analyzing climate data and cloud cluster image data in the real-time data, judging the illumination condition of the photovoltaic array area, and calculating the moving direction and the moving speed of the photovoltaic array area under the illumination condition; calculating the actual illuminated area of the photovoltaic array area under each time node, introducing the information into the initial power generation power model for optimization, and generating new prediction data; and carrying out comparison verification on the optimized prediction data and subsequent actual generation power. According to the method, the prediction accuracy is remarkably improved, and the influence of environmental factors on the prediction result is effectively reduced.

The invention discloses a large indoor direct shear tester for geotechnical engineering, and relates to the technical field of geotechnical shear tests.The large indoor direct shear tester for geotechnical engineering comprises a sample container system, a shear loading system, a data acquisition control system and a base protection system.The sample container system is formed by splicing a plurality of modular sample containers; the modularized sample container is connected through a detachable connecting fastener, a replaceable lining plate is arranged in the container and is used for adjusting the size and interface conditions of the sample container, and an embedded porous pressure plate is arranged at the bottom of the container and is used for measuring the pore pressure change in the sample seepage process. By installing the base, the protection system and the normal loading system, the function of uniformly applying normal force in the shearing process is achieved, the problem that the test result is affected by eccentric errors in the shearing process is solved, and the precision and stability of the shearing test are improved.

The invention discloses a scale-effect-free mine dynamic disaster fine monitoring and dynamic early warning method, which comprises the following steps of: arranging a multi-point micro-seismic monitoring network on a coal mine working face, and continuously acquiring and preprocessing micro-seismic signals; an innovative analysis positioning algorithm is adopted to accurately analyze the accurate position and time information of the micro-seismic source; in combination with signal features and positioning data, a four-dimensional joint kernel density estimation and confidence interval theory is innovatively applied, a rock burst risk assessment model is constructed, and a danger level classification standard is determined; and the system calculates risk indexes in real time and dynamically issues early warning. According to the invention, the precision, timeliness and reliability of monitoring and early warning of the dynamic disasters in the mining area are obviously improved, and key technical support is provided for safe production of a coal mine.

The invention relates to the technical field of automatic control, in particular to an unmanned aerial vehicle tourism photography automatic control system which comprises a light and climate monitoring module, a photography configuration adjusting module, a scenic spot coverage planning module and a real-time shooting execution module. According to the invention, environmental data is collected and analyzed in real time, the photographing parameters, including aperture, exposure time and ISO, of the unmanned aerial vehicle can be rapidly adjusted to adapt to the rapidly changing external environment, the flight path is dynamically calculated and optimized by means of the GIS technology, the energy consumption efficiency and coverage of the photographing task of the unmanned aerial vehicle are maximized, and the photographing efficiency of the unmanned aerial vehicle is improved. Through the automatic adjusting function of the real-time shooting execution module, the unmanned aerial vehicle can adjust the flight height and the camera angle according to actual climate conditions, the automation degree of operation and the adaptability of the system are remarkably improved, the response speed and the environment adaptability are remarkably improved, errors caused by manual operation delay are reduced, and the shooting quality and efficiency are improved.

The invention relates to a high-frequency and high-performance iron-nickel alloy magnetic powder core and a preparation method and application thereof, belongs to the field of metal magnetic powder cores, and solves the problems that in the prior art, when a metal magnetic powder core is applied at high frequency, high magnetic conductivity, high magnetic conductivity stability and high quality factor are difficult to consider, and the saturation flux density is low. And the preparation process is tedious. The invention discloses a preparation method of a high-frequency high-performance iron-nickel alloy magnetic powder core. The preparation method comprises the following steps: S1, preparing raw materials; s2, performing pretreatment; s3, the pretreated iron-nickel alloy powder and nano SiO2 powder are blended, and inorganic insulation coating powder is obtained; s4, adding the inorganic insulating coating powder into a first coating agent to obtain inorganic and first organic insulating coating powder; s5, adding the inorganic and first organic insulation coating powder into a second coating agent to obtain inorganic and double-layer organic insulation coating powder; and S6, compression molding. The comprehensive magnetic performance of the iron-nickel magnetic powder core is improved, and the preparation process is simplified.

The invention provides magnesium phosphate cement-based concrete. The magnesium phosphate cement-based concrete is prepared from the following raw materials in parts by weight: 10-20 parts of phosphate, 30-50 parts of active magnesium oxide, 0.5-3 parts of a retarder, 5-15 parts of a mineral admixture, 0.5-4 parts of a functional additive, 20-30 parts of fine aggregate, 2-5 parts of reinforced fibers and 10-28 parts of water. The invention also provides a preparation method of the magnesium phosphate cement-based concrete. The magnesium phosphate cement-based concrete provided by the invention has a multi-layer anti-corrosion barrier, a compact microstructure, an intelligent ion barrier and good environmental adaptability, and can meet the use requirements in complex and severe environments.

The invention discloses a high-efficiency artificial cultivation method for gordon euryale seeds in northeast marsh, and belongs to the technical field of production and cultivation of economic plants in marsh, the method comprises the following steps: 1, selecting and germinating wild seeds; 2, seedling raising and temporary planting; 3, field transplanting and field planting; 4, growth management and harvesting. According to the method, the maturing rate of the gordon euryale seeds can be remarkably increased only by adopting a scientific method during transplanting, keeping the plant spacing and reasonably regulating and controlling the water level after transplanting, operation is easy and convenient, and the seed yield per unit area is increased by 5%. The method can effectively solve the problems of low survival rate, uneven growth vigor and unstable yield of wild gordon euryale seeds in northeast marsh.

The invention relates to the field of sponge materials, and provides an antibacterial degradable sponge material and a preparation method thereof. The material is prepared from protocatechuic acid modified quaternized chitosan, quaternary ammonium salt grafted gelatin, sodium alginate, polyhydroxybutyrate, polycaprolactone, a polylactic acid-glycolic acid copolymer, hexamethylene diisocyanate, a foaming agent, nano cellulose, vitamin E, polyethylene glycol, triethylamine and dimethylformamide. Protocatechuic acid and quaternized chitosan are subjected to amidation modification, and chitosan and gelatin are subjected to quaternization treatment by adopting 2, 3-epoxypropyl trimethyl ammonium chloride, so that the antibacterial property and biodegradability of the material are improved. In the material preparation process, through vacuum degassing, foaming vacuum degree and humidity control and a multi-step drying process, it is ensured that the average pore size of the material is uniform (50-200 microns) and the porosity is high (50%-90%), and the obtained material has excellent antibacterial property and degradability and has wide application value.

The invention relates to an aerosol moisture absorption and activation process observation system and an observation method thereof. The aerosol moisture absorption and activation process observation method comprises the following steps: collecting sample gases under different humidity conditions, and carrying out particle size screening on the sample gases to respectively obtain aerosols with different particle sizes; measuring scattering coefficients of the aerosols with different particle sizes under different humidity conditions; calculating scattering moisture absorption growth factors of the aerosols with different particle sizes according to the scattering coefficients, and performing inversion calculation on the scattering moisture absorption growth factors so as to obtain moisture absorption, water content, volume concentration and surface area concentration of the aerosols; according to the scattering coefficients corresponding to the aerosols with different particle sizes in the environment state, the dry state critical particle sizes of the aerosols with the dry state increased to 1 micrometer and 2.5 micrometers through water vapor condensation and the scattering moisture absorption growth factor of the aerosols with the environment state ranging from 1 micrometer to 2.5 micrometers are calculated in an inversion mode, and the effective supersaturation ratio of the cloud mist is obtained through inversion. And the aerosol moisture absorption and activation process observation system is arranged according to the aerosol moisture absorption and activation process observation method.

The invention discloses an NURBS curve robot trajectory offset point focusing fitting method based on adaptive optimization, which remarkably improves trajectory precision and calculation efficiency through adaptive optimization focused on an error offset point. The method comprises the steps of initial path planning, NURBS curve fitting, error evaluation and offset point extraction, adaptive optimization and real-time feedback adjustment. In error evaluation, double indexes of a root-mean-square error and a maximum absolute error are adopted, an offset point is accurately identified, a control point, a weight and a node vector are focused and optimized by using a genetic algorithm, and local optimum is effectively avoided. A regularization item is introduced into a fitness function, curve complexity and precision are balanced, and smoothness and feasibility of a track are achieved. And finally, real-time adjustment of the trajectory is realized through a multi-feedback mechanism, and the robustness of the robot in a dynamic environment is improved. Compared with the prior art, the precision, the real-time performance and the environmental adaptability of track generation are realized, and particularly, an efficient and accurate track fitting effect is achieved in a complex dynamic environment.

The invention discloses a construction site inspection method, inspection system and equipment and a nonvolatile storage medium, and relates to the technical field of intelligent construction, and the method specifically comprises the steps: carrying out the inspection planning according to the inspection requirements of a construction site; based on the inspection plan, executing an inspection task, and obtaining inspection data of a target area; and analyzing the inspection data by constructing a deep learning algorithm model, and carrying out risk event classification based on an analysis result. According to the invention, a closed-loop processing mechanism of data in a platform-nest-unmanned aerial vehicle-platform is driven through control of the unmanned aerial vehicle cruise by the intelligent construction transportation and management platform for construction engineering, and a mobile supervision mode for engineering quality and safety is innovated.