36results about How to "Lower application threshold" patented technology

The invention provides an urban green infrastructure comprehensive benefit applicability evaluation method based on subjective and objective weights. The urban green infrastructure comprehensive benefit applicability evaluation method comprises the steps of establishing a basic evaluation system, screening indexes, calculating index weights, calculating quantitative indexes, calculating qualitative indexes, performing grading evaluation, performing standardized score assignment and performing comprehensive evaluation. The basic evaluation system is of a three-layer structure and covers four criteria of runoff regulation and storage benefits, environmental benefits, economic benefits and social benefits and 19 indexes; a subjective weight is determined through an analytic hierarchy process, an objective weight is determined through an entropy weight method, and a comprehensive weight is obtained in combination with a D-S evidence theory fusion method; quantitative indexes are calculated through methods such as hydrological simulation, and qualitative indexes are rated by public; the indexes are unified into dimensionless scores through standardized score assignment, and finally an evaluation result is obtained through weighted summation. According to the method, the applicability quantitative evaluation of matching of the typical green infrastructure and the site condition construction target is realized, and a scientific and quantitative decision support tool is provided for sponge city planning.

The invention discloses a method, a system and equipment for troubleshooting a communication terminal of an all-moving simulator, and belongs to the technical field of aviation simulation equipment fault diagnosis. The method comprises the following steps: acquiring data of a communication terminal operation state, an input / output system state and a central control console environment state, synchronizing time, and constructing a time sequence feature vector through denoising normalization processing; inputting a Kalman filtering state space model to calculate an observation residual error, and distinguishing a fault source and interference characteristics; identifying a fault type by using an integrated decision tree classification model, positioning a target fault point in combination with a fault tree and a Bayesian network, and matching historical cases to generate a troubleshooting scheme; after troubleshooting is completed, the fault case library is updated, and the model is optimized. According to the scheme, intelligent and standardized troubleshooting of faults is achieved, the problems that a traditional method is low in efficiency and poor in accuracy, and fault sources are difficult to distinguish are solved, the troubleshooting efficiency and accuracy are greatly improved, the operation and maintenance cost is reduced, and stable operation of the full-motion simulator is guaranteed.

The invention discloses a total variation regularization depth expansion network method for hyperspectral image unmixing. The method comprises the following steps: inputting hyperspectral image data and an end member spectrum library; constructing a hyperspectral data unmixing model fusing a sparse constraint term and a total variation regularization term and an optimization objective function of the hyperspectral data unmixing model; converting the optimization objective function into an equivalent augmented Lagrangian form, and performing iterative solution on each variable; expanding an iteration process into a deep learnable network structure with a plurality of layers, and parameterizing hyper-parameters and a plurality of approximate operators in iteration into learnable network parameters; constructing a comprehensive loss function, and continuously optimizing network parameters by using back propagation and gradient descent methods; according to the method, total variation regularization prior is fused into a sparse unmixing model, the sparse unmixing model is expanded into a learnable end-to-end neural network through iterative solution, and the method has excellent unmixing precision and robustness for noise-containing data.

The invention provides a log analysis method and system based on a large language model, and belongs to the technical field of log analys.The method comprises the steps that anti-fact rewriting is conducted on a to-be-analyzed log, and multiple log variants are obtained; log parameters are extracted from the log variants through the large language model to serve as intermediate variables, and the confidence coefficient of each intermediate variable is determined; performing correctness scoring on each intermediate variable based on statistical information of a historical log corpus knowledge base; determining a comprehensive score of each intermediary variable according to the confidence coefficient of the intermediary variable and the correctness score of the intermediary variable; and screening the medium variable with the highest comprehensive score as a log parameter of the log to be analyzed, and generating a log template of the log to be analyzed according to the log parameter of the log to be analyzed. According to the log analysis method, the stable semantic structure of the log can be more emphasized in the log analysis process, and the accuracy and generalization ability of log analysis are improved.

This invention discloses a method, apparatus, equipment, and storage medium for predicting the minimum miscibility pressure of crude oil. The method includes acquiring the target crude oil for which miscibility pressure prediction is to be performed and the target gas for injecting into the target crude oil for oil displacement. The target crude oil and target gas are subjected to component analysis to determine the necessary component parameters for predicting the minimum miscibility pressure. Based on a pre-fitted target miscibility pressure prediction model, the target minimum miscibility pressure for the miscibility of the target crude oil and target gas is determined according to the necessary component parameters. The target miscibility pressure prediction model is divided into a first miscibility pressure prediction model and a second miscibility pressure prediction model based on the gas type of the target gas. This invention is applicable to injection gas scenarios involving pure CO2 or flue gas containing different impurities. By separately calculating the miscibility pressure contributed by the oil phase itself and the miscibility pressure contributed by the injection gas, the accuracy and applicability of minimum miscibility pressure prediction can be significantly improved.

The invention discloses an output method based on 32-bit single-chip microcomputer firmware upgrade memory anti-overflow protection, and relates to the field of single-chip microcomputer memory anti-overflow protection, and the method specifically comprises the following operation steps: S1, environment variable configuration and management; s2, performing core execution of the Python automatic script; s3, performing anti-overflow protection on the memory; and S4, cleaning redundant files and environment variables. According to the output method based on the 32-bit single-chip microcomputer firmware upgrade memory anti-overflow protection, the file path is defined as the environment variable and stored in the independent configuration file in a centralized mode, complete separation of the path and script logic is achieved, the non-transplantable problem caused by a hard coding path is avoided, meanwhile, the adaptation cost in different environments is reduced, and the adaptation efficiency is improved. Operators do not need to modify script codes and only need to adjust configuration files to complete deployment.

This invention discloses a high-precision feedback measurement system and method for triaxial magnetic fields based on the SPGD algorithm, comprising: an atomic magnetometer, a triaxial Helmholtz coil group, a signal acquisition and processing module, an SPGD algorithm control module, a multi-channel precision current source, and a magnetic field calculation module, with each module forming a closed-loop feedback link. The signal acquisition and processing module extracts relevant physical quantities from the performance evaluation function in the output signal of the atomic magnetometer. The SPGD algorithm control module generates control commands through parallel perturbation, gradient estimation, and parameter updates. The multi-channel precision current source converts the commands into a stable current to drive the coils to generate a compensating magnetic field. The magnetic field calculation module calculates the environmental magnetic field components by combining calibration coefficients and the driving current. This invention can be directly adapted to commercial atomic magnetometers, automatically decouples the triaxial magnetic field coupling, and achieves rapid convergence and accurate measurement. It is suitable for high-precision vector magnetic field measurement in fields such as geomagnetic navigation, biomagnetic imaging, and space exploration.

This invention relates to the field of spectral analysis and identification of genuine and counterfeit Chinese medicinal materials, and discloses a method for identifying counterfeit Cordyceps sinensis based on multispectral fusion. This method first collects multispectral image data of samples under standard conditions; eliminates equipment and environmental noise through standardized preprocessing to establish a reliable data foundation; then, it fuses extracted spectral derivative features and spatial texture features to construct a lightweight neural network identification model. Its core innovation lies in introducing a physical interference stripping model, which can actively correct spectral distortion caused by moisture and mold, obtaining intrinsic spectra to improve model robustness; and through dynamic verification and incremental learning mechanisms, the model can continuously evolve, effectively identifying new counterfeiting methods. Finally, the entire model is integrated into a multispectral portable terminal, achieving rapid screening within seconds and outputting a complete report including authenticity determination, type inference, and confidence level. Ultimately, this solves the industry problems of low efficiency, reliance on experience, and inability to be applied on-site by traditional methods.

This invention discloses a non-inertial navigation mobile indoor volume measurement and 3D reconstruction method, equipment, and computer-readable medium, belonging to the interdisciplinary field of precision engineering measurement, computer vision, and digital twin technology. The method includes the following steps: S1, dynamic scanning data acquisition; S2, adaptive data augmentation and cleaning; S3, trajectory and depth calculation based on motion continuity constraints; S4, cross-sectional topology construction and closure constraints; S5, 3D solid modeling preserving volume characteristics; and S6, precise volume calculation based on the solid model. This invention employs the aforementioned non-inertial navigation mobile indoor volume measurement and 3D reconstruction method, equipment, and computer-readable medium, and through a single-line 2D LiDAR combined with innovative algorithms, achieves mobile high-precision volume measurement and high-fidelity 3D reconstruction of irregular indoor spaces under low-cost hardware conditions.

The application discloses an industrial control method and device based on model training, and belongs to the field of industrial control. The industrial control method based on model training comprises the following steps: using a target algorithm prediction model to perform operation processing on real-time industrial big data in a target industrial scene, so that a target control algorithm and a target independent variable for controlling a target controlled object in the target scene output by the target algorithm prediction model are obtained; and inputting the value of the target independent variable in the real-time industrial big data into a target controller in which the target control algorithm is deployed, so that the value of a target control variable is obtained, to control the target controlled object. The industrial control method based on model training has high precision and accuracy of the obtained target control algorithm, can significantly reduce the labor and time cost, improve the control efficiency and control effect, and has high universality.

The invention belongs to the technical field of fluid mechanical engineering, and particularly relates to a centrifugal pump and turbine dual-mode shaft power calculation method based on mechanical energy and internal energy analysis. Performing three-dimensional flow field calculation based on the pump mode system mass flow; calculating the specific mechanical energy difference between the outlet boundary and the inlet boundary of the pump mode system; calculating the specific internal energy difference between the outlet boundary and the inlet boundary; calculating pump mode shaft power of the centrifugal pump; determining a turbine mode system of the centrifugal pump; three-dimensional flow field calculation is executed based on the turbine mode system mass flow; calculating the specific mechanical energy difference between the inlet boundary and the outlet boundary of the turbine mode system; calculating the specific internal energy difference between the outlet boundary and the inlet boundary; calculating the turbine mode shaft power of the centrifugal pump; through collaborative analysis of mechanical energy and internal energy, the structure and operation characteristics of the pump and turbine dual mode of the centrifugal pump are adapted, and the technical blank of quantitative shaft power calculation under the pump and turbine dual mode of the centrifugal pump is made up from the perspective of system energy dissipation and conservation.

The present application relates to the field of temperature measurement, in particular to a kind of metal organic framework crystal material and its preparation method and application.The present application provides a kind of metal organic framework crystal material, wherein the crystal material is connected body 2,6-di (2,4-dicarboxyphenyl) -4- (phenyl) pyridine and Zn (II) connected with the crystal material with periodic three-dimensional network structure.The metal organic framework crystal material provided by the present application still keeps the afterglow characteristic under the condition of smaller particle size, and the charging time is shorter;At the same time, the decay time of the metal organic framework crystal material of the present application is seconds, and the light source only needs ordinary ultraviolet semiconductor light source, and the camera can record the time / color of afterglow decay, as the basis for temperature judgment, which greatly reduces the temperature measurement cost and simplifies the measurement process;In addition, the metal organic framework crystal material of the present application can be synthesized below 120 DEG C, which reduces the production cost and danger.

The invention belongs to the field of early disease diagnosis, and particularly relates to a core-shell nanoparticle labeled portable dual-mode immunoassay platform based on dynamic driving, which comprises an antibody-gold nanoparticle biological coupling system, an amplification system, a color development system and a colorimetric-atomic emission spectrum detection system. The Au-coated Cu core-shell nano-particles are dynamically driven to realize signal amplification, colorimetric preliminary screening and atomic emission accurate quantification are combined, CEA (0.02 ng / mL), cTnI (0.35 fg / mL) and A beta (10.3 pg / mL) can be detected in an ultra-sensitive manner, and the Au-coated Cu core-shell nano-particles are suitable for bedside detection of early diseases.

This invention discloses a method and system for evaluating DNA storage sequencing depth based on channel simulation, addressing the problems of inaccurate predictions and limited guidance in existing uniform distribution models. The method includes: when sequencing data is available, fitting real data to obtain log-normal distribution parameters μ and σ; when sequencing data is unavailable, obtaining μ and σ through simulation modeling based on experimental parameters; and combining these parameters to calculate the decoding ratio of the coding strand in the noiseless channel and the sequencing depth boundary in the noisy channel. The system includes input, storage channel modeling, sequencing depth calculation, and output modules. The model of this invention closely reflects reality, provides accurate predictions, reduces DNA storage and retrieval costs, improves the success rate of single-sequencing decoding, and is easy to use, facilitating the practical application of the technology.

This invention discloses a rapid early warning and estimation method for the structural state of prestressed concrete continuous rigid frame bridges based on long-term deflection changes, belonging to the field of bridge structural health monitoring and assessment technology. This method involves setting up permanent observation points during the operation of the continuous rigid frame bridge to collect long-term deflection data and establish a deflection characteristic equation. Combined with the prestress-deflection and main beam stress-deflection regression relationships, it achieves dynamic calculation of the structure's prestress and stress state. Simultaneously, it sets early warning thresholds based on deflection development trends to provide rapid early warning of abnormal structural deformation and completes the serviceability limit state and ultimate limit state assessments according to current standards. This invention solves the problems of reliance on short-term data, delayed early warning, and inaccurate state estimation in traditional bridge assessments. It achieves long-term dynamic tracking and accurate prediction of structural state, providing a scientific basis for bridge maintenance decisions and is applicable to the health monitoring and performance assessment of continuous rigid frame bridges during the operation period.

The invention discloses a method for promoting growth of intercropping crops based on a soil micro food web and application of the method, and belongs to the technical field of crop planting. According to the invention, it is found for the first time that by adding a specific type of bacterial-feeding nematodes into soil, synergistic improvement of growth indexes such as plant height, stem diameter and biomass of two crops can be obviously and stably promoted; the application method is simple and easy to operate, the cost is controllable, and no additional farming equipment or steps are needed; the method is high in environmental compatibility, is particularly suitable for a farmland peanut-corn intercropping system in alkaline and calcareous soil, can effectively improve the crop growth performance, farmland productivity and agricultural product yield and quality, can optimize the soil ecological function, continuously increases the economic benefits of farmers, and has wide application prospects. A brand-new, low-cost and easy-to-popularize candidate scheme is provided for agricultural green sustainable development of the alkaline soil area in North China, and the application prospect is wide.

The invention is suitable for the technical field of automobile aerodynamics and thermal management, and provides an automatic opening and closing heat dissipation and resistance reduction hubcap structure and a control method thereof.The hubcap structure comprises a hubcap base and a hubcap cover, the hubcap base is provided with base ventilation openings, the hubcap cover is provided with cover cover ventilation openings, and the hubcap ventilation openings are communicated with the hubcap base. The blades are connected with the hubcap base through reset springs; the traction and locking unit comprises a sliding way, a traction sliding block is installed in the sliding way, the traction sliding block is connected with the blade through a traction rope, a locking magnet is arranged in the sliding way, a traction spring connected with the locking magnet is arranged in the sliding way, and a locking iron sheet used for attracting the locking magnet is arranged on the blade. The device completely depends on the physical principle to achieve self-adaptive opening and closing, does not need external energy, is simple and reliable in structure, low in cost and convenient to install, can balance the requirements for high-speed resistance reduction and low-speed heat dissipation, and effectively improves energy efficiency and driving safety.

The invention discloses a large language model-driven cross-multi-source heterogeneous data platform collaborative abnormal behavior detection method and a large language model-driven cross-multi-source heterogeneous data platform collaborative abnormal behavior detection system. The method comprises the following steps: aiming at a target entity, respectively extracting local dynamic sub-graphs of the target entity in a plurality of heterogeneous data platforms, fusing specific structural relationships, attribute features and text contents of the platforms, and constructing a multi-source heterogeneous time sequence behavior snapshot sequence; a frozen graph encoder and a semantic model are adopted to independently encode each source sub-graph, and multi-channel snapshot embedding in a unified representation space is generated through cross-domain alignment projection under privacy protection constraint; designing a comparative learning strategy of federal perception, and in combination with a cross-platform behavior consistency regular item and a generative classification target, embedding and injecting the snapshots into a large language model in a multi-channel soft prompt form to perform parameter fine tuning; and performing end-to-end collaborative reasoning on the cross-platform behavior sequence by using the fine-tuned model, and outputting an abnormal risk level.

The application discloses a weakly supervised image semantic segmentation training and reasoning method, and relates to the technical field of image processing, comprising the following steps: S1, a causal graph model containing image features, category labels, co-occurrence context and segmentation results is constructed, the causal relationship between variables is determined, and the false causal path generated by the co-occurrence context as a confounding variable on segmentation prediction is identified; the complete causal graph model containing image features, category labels, co-occurrence context and segmentation results is constructed, the false causal path generated by the co-occurrence context as a confounding variable is identified, the confounding variable is sampled and adjusted by a backdoor, and the false causal path is cut off from the root of causal reasoning, meanwhile, a causal consistency loss function is designed to constrain the consistency of segmentation results under different context environments, and the effect that the false correlation problem of the co-occurrence context can be effectively solved without additional labeled data or complex multi-model cooperation is realized.

This invention discloses a precision irrigation system and method for crops based on root water demand signals, belonging to the field of precision irrigation technology for agricultural roots. It includes a signal sensing module, an intelligent decision-making module, a precision execution module, a data storage module, and a feedback unit. The signal sensing module consists of a root image visual sensor, a stem micro-change sensor, a canopy infrared temperature sensor, and a root zone microclimate sensor array. This invention utilizes non-destructive multimodal sensing technology, comprehensively employing multi-sensor fusion sensing such as root window vision, stem flow or stem micro-changes, canopy infrared temperature, and root zone water potential to directly capture early water demand and stress signals released by root physiological activities. Furthermore, for mixed or vertical planting scenarios, it achieves specific identification of water demand signals for different crops and different vertical spaces through three-dimensional adaptive deployment and a multi-crop signal feature library.

This invention discloses an AI-based virtual restoration and display system and method for grotto statues, belonging to the field of grotto statue restoration technology. It includes a data acquisition module, an intelligent restoration processing module, and a display output module connected in sequence. This invention acquires complete 3D and texture data of the statues through a non-contact data acquisition module, completing the entire virtual restoration process in digital space without any physical contact or intervention with the grotto statues. This effectively avoids the risk of secondary damage caused by traditional physical restoration, conforming to the principle of minimal intervention in cultural relic protection. Simultaneously, this invention can generate multiple differentiated restoration plans through a CSN artificial intelligence model, supporting experts to switch, adjust, and revoke restoration operations at any time, making the restoration process completely reversible. Furthermore, a secure encryption unit completely records the operation trajectory and model parameters of the entire restoration process, achieving full traceability and auditability of the restoration process, meeting the standardized management requirements of cultural relic restoration.

The invention relates to a defect detection system based on a large projection model, and relates to the technical field of machine vision defect detection. The technical problems that an existing defect detection technology needs to depend on specific scene customization design, the cost is high, and the detection effect is limited are solved. The system comprises a hardware module, a data acquisition module, a model training module and a detection application module, the hardware module forms an imaging facility with a fixed pose by a DLP (Digital Light Processing) and an area-array camera, the data acquisition module obtains a 70-channel initial imaging effect picture and a structured cue word of multi-period fringe projection, and the model training module realizes cooperative training of a projection large model and a detection large model through reinforcement learning. And the detection application module completes full-process automatic defect detection. Manual customization design is not needed, the technical threshold is reduced, time and labor cost are saved, the detection precision and universality are remarkably improved, hardware equipment is small, exquisite and portable, and the method is suitable for detection scenes of various materials, surface types and defect types.

The invention relates to the technical field of intelligent agriculture, in particular to a fig water and fertilizer digital management method based on a low-code platform. The core of the method comprises the following steps: constructing a water and fertilizer decision model integrated with time sequence prediction and an attention mechanism through a low-code platform, processing an environmental data sequence by adopting an improved Transform architecture, and capturing periodic characteristics in combination with learnable time perception codes; establishing a segmented attention weight distribution strategy based on a growth stage, and dynamically adjusting the attention degrees of different environment characteristics; and configuring a model updating system comprising an online learning mechanism, and realizing continuous optimization of model parameters through an elastic weight consolidation algorithm. Accurate prediction and self-adaptive regulation and control of fig water and fertilizer requirements are realized, and the water and fertilizer management precision and the system availability are effectively improved.

The invention discloses a self-adaptive weight UWB-ultrasonic fusion positioning method. According to the technical scheme, the method is characterized by comprising the following steps that S1, parameters are collected, wherein UWB positioning data, ultrasonic positioning data, environment state parameters, equipment working state parameters and positioning task requirement parameters are synchronously collected; s2, acquiring effective positioning data: performing time synchronization, space calibration, anomaly detection and smoothing processing on the acquired positioning data to acquire the effective positioning data; s3, outputting an evaluation result: based on the positioning data quality, the environment suitability and the task matching degree, carrying out comprehensive evaluation on the UWB and the ultrasonic positioning module, and outputting the evaluation result; s4, obtaining an adaptive weight; s5, outputting a real-time positioning coordinate; and S6, performing closed-loop control. According to the method, the positioning precision is greatly improved, the method is adaptive to a complex environment, and fusion positioning can be accurately adaptive to different environment changes through environment state sensing and self-adaptive weight dynamic adjustment.

The application discloses a kind of precision calculation method and system of carbon fiber cloth surface acoustic impedance, it is related to the cross field of acoustics engineering and composite material technology, including: obtaining the fiber volume fraction of carbon fiber cloth, warp and weft weaving density, fiber diameter, single layer thickness, layer number, adjacent layer stacking angle, back cavity depth and acoustic angular frequency;According to warp and weft weaving density and fiber volume fraction, anisotropic correction coefficient and equivalent weaving density are calculated;Utilize the anisotropic correction coefficient, equivalent weaving density and other parameters to calculate acoustic resistance;Utilize acoustic angular frequency, fiber volume fraction, single layer thickness, fiber diameter, back cavity depth, layer number and stacking angle to calculate acoustic reactance;Acoustic resistance and acoustic reactance are integrated into the surface acoustic impedance of complex number form.The application can accurately adapt single layer, multi-layer stacking configuration carbon fiber cloth, with small calculation error and high calculation efficiency, and can be widely applied to aerospace acoustic lining, high-end equipment noise reduction and other scenes.

The present application relates to the technical field of remote sensing image processing, and particularly relates to a remote sensing identification method for gordon euryale seed traditional Chinese medicinal materials based on deep learning, which comprises the steps of data acquisition, feature extraction, construction of an optimized feature set, construction of a convolutional neural network model and model training. The method can distinguish crops with high spectral similarity with high precision, and is fully automated. No manual intervention is required in the processes of phenological feature extraction, feature optimization and intelligent interpretation, subjective bias caused by manual experience is avoided, the overall processing efficiency is improved compared with the manual auxiliary method, the method can be quickly popularized to automatic interpretation in a large area; meanwhile, the method is compatible with optical and radar multi-source data, and can still maintain high classification accuracy in the case that optical images are disturbed by clouds and rain or crop plots are severely fragmented, thereby significantly improving the applicability in complex scenes.

This invention relates to the field of simulation technology, specifically to an intelligent control system for heat treatment of drill pipes based on digital twins. The system collects environmental parameters, medium parameters, and workpiece state parameters from the heat treatment production line and outputs a time-series input vector. It reconstructs heat exchange boundary conditions using the medium parameters contained in the time-series input vector. Combining these boundary conditions, it drives a strongly coupled calculation based on a phase transformation-induced strain mechanism, involving a temperature field submodule, a microstructure evolution submodule, and a stress field submodule. This outputs a virtual state matrix containing the temperature distribution across the entire cross-section, grain size distribution, phase transformation volume fraction distribution, and stress distribution. Based on this virtual state matrix, a multi-objective optimization algorithm is executed in a multi-dimensional process parameter space to find the optimal balance point of strength and toughness indicators while satisfying microstructure and stress constraints. This generates an optimal heating power curve and dynamic quenching parameters. Finally, it optimizes production through a process instruction set.

This invention discloses a method and system for dynamically correcting suspended sediment transport during the flood season in run-of-river hydropower stations, relating to the field of hydropower engineering technology. The method first establishes instantaneous sediment transport rate relationship models, cumulative sediment transport relationship models, and flood peak lag relationship models based on historical time-series data. Then, it determines the current lag time based on the real-time rainfall intensity and previous soil moisture content of the current flood event. Based on this lag time, it reconstructs the phase of the sediment concentration process line on the time axis to obtain the reconstructed sediment concentration process line. Next, based on the cumulative sediment transport relationship model and inflow forecast data, it calculates the phase-corrected suspended sediment transport at each future moment. Finally, it generates prediction results and dispatch decision support information based on the corrected sediment transport. This method can quantitatively describe and correct the phenomenon of asynchronous water and sediment sources during the flood season, reducing the calculation errors caused by the assumption of synchronous water and sediment in traditional water-sediment relationship methods.

This invention discloses a method and apparatus for optimizing the formulation of a waste-decomposing complex enzyme. The method includes: obtaining a set of formulation information for the complex enzyme; the set of formulation information includes a component vector for each formulation; the component vector includes the component ratio values ​​of lipase, protease, cellulase, and amylase; conducting waste decomposition experiments on the complex enzyme based on the set of formulation information to obtain a set of decomposition effect scores; and optimizing the set of formulation information based on the set of decomposition effect scores to obtain optimized formulation information.

This invention belongs to the field of defect recognition technology, and particularly relates to an intelligent defect recognition method for power lines based on image processing. The method first performs multi-scale, multi-directional phase consistency calculations on visible light images of power lines, generating a maximum phase consistency response map, a feature direction map, and an energy map. Then, it fuses gradient magnitude maps and extracts key component regions such as conductors and insulators through watershed transformation and a hierarchical merging strategy based on minimum spanning trees, combined with geometric constraints. Next, it extracts multi-dimensional local features based on these key component regions, constructs a normal feature subspace using principal component analysis, and generates an anomaly score map by reconstructing residuals to achieve unsupervised defect detection. Finally, it extracts defect contour features and outputs a defect severity score using a random forest model. This invention exhibits strong resistance to illumination changes and background interference, requires no large amount of defect annotation data, has high recognition accuracy, and can quantify defect severity, making it suitable for power line inspection scenarios.