32results about How to "Reliable extraction" patented technology

The invention discloses an LED lamp strip rhythm flickering control method and system, a storage medium and a program product, and the method comprises the steps: calculating a relative entropy value among a high-frequency band, a middle-frequency band and a low-frequency band; establishing an energy flow diagram, and determining a dominant frequency band according to an energy flow topological structure in the energy flow diagram; a phase tracking function is constructed, and hierarchical level division is carried out on the beat mark points; establishing a recursive segmentation window, calculating a spectrum centroid trajectory, and extracting an inflection point sequence; establishing a corresponding relation between the polar coordinate angle and the color parameter of the LED lamp strip and a corresponding relation between the polar radius and the brightness parameter; a spatial propagation mode of the LED lamp strip is designed based on the topological structure of the energy flow diagram, so that the propagation direction of the light effect along the LED lamp strip is kept consistent with the energy flow direction between the frequency bands, and the propagation speed is adjusted according to the energy flow intensity; and generating a driving control sequence of the LED lamp strip according to the gradient function. The emotional progression and tension change of music are captured, and the LED lamp strip is accurately controlled to flicker according to the emotional progression and the tension change.

The invention belongs to the field of ionosphere satellite magnetic field seismic anomaly extraction, and particularly relates to a satellite magnetic field seismic anomaly extraction method based on complex non-negative matrix factorization, which comprises the steps of reading Swarm A satellite magnetic field data, and removing a main magnetic field of a ground core and a lithosphere magnetic field of an earth crust to obtain a residual magnetic field; performing short-time Fourier transform on the preprocessed data to obtain a complex time-frequency matrix in a rectangular coordinate representation form of a complex domain; decomposing the complex time-frequency matrix by using complex non-negative matrix decomposition to obtain R feature components; performing inverse short-time Fourier transform on each feature component to obtain a time domain feature component; a mean square root of a time domain seismic component of each orbit is calculated, a threshold value is set by using an over-limit rate method, and abnormal points are extracted from the seismic components of the orbits. The amplitude and phase information contained in the time-frequency domain of the satellite observation data is fully utilized, so that the component related to the earthquake is identified, and the earthquake anomaly extraction is more accurate and reliable.

The invention discloses a pollution source tracing method based on a three-dimensional fluorescence spectrum, and relates to the technical field of water quality monitoring, a water sample of an upstream specific pollution discharge unit is collected as a pollution source, a mixed water sample formed by mixing different pollution source water samples according to different volumes or proportions is prepared, and a pollution source database adapted to parallel factor analysis is established; according to the pollution source tracing method, a good linear model between a pollution source and a three-dimensional fluorescence spectrum characteristic value is established, a shortest path algorithm is proposed to quantitatively solve the volume or proportion of the pollution source in a monitored water sample, and the pollution source is traced to a specific pollution discharge unit. According to the invention, the sewage three-dimensional fluorescence spectrum database established by mixing upstream pollution source drainage improves the application range of parallel factor analysis, so that the three-dimensional fluorescence spectrums of sewage of different industrial systems can be subjected to effective parallel factor analysis together; compared with a parallel factor analysis-Tucker congruence coefficient method, the method has the advantage that the pollution source can be traced to a more specific level.

The present invention provides a variational mode decomposition optimization method based on a time-frequency correlation coefficient joint factor, comprising: S1, initializing the K value, setting K=2; S2, performing VMD decomposition on the signal to obtain K IMF components; S3, calculating the time domain correlation coefficient ρ between each IMF component and the original signal K,i and frequency domain correlation coefficient #imgabs0#Calculate the time-frequency correlation coefficient joint factor δ K,i ; S4, let K = K + 1, perform VMD decomposition again, and obtain K + 1 IMF components; S5, calculate the time domain correlation coefficient ρ between each IMF component and the original signal K+1,j and frequency domain correlation coefficient #imgabs1#Calculate the time-frequency correlation coefficient joint factor δ K+1,j ; S6, calculate the time-frequency correlation coefficient joint factor difference β between each IMF component when K takes different values ​​and the same decomposition number K,r ; S7, when β K,r ≤θ(θ=0.05), it is considered as an effective factor, otherwise it is considered as an invalid factor; S8, when K=m, the number of continuous effective factors reaches the maximum value, that is, when K=m+1, the number of continuous effective factors is less than K=m, it is considered that the decomposition is optimal, and K=m is finally selected as the optimal value. The present invention is simple to operate, the correlation feature extraction is comprehensive and reliable, and is less affected by the factors of the signal itself.

The invention belongs to the technical field of lithium ion batteries, and provides a power battery multi-state collaborative estimation method and system based on a large language model, and the method comprises the steps: initializing the large language model; aging feature extraction is carried out under the random charging behavior, and window extreme voltage and accumulated charging amount are extracted as aging features; based on the initialized large language model, determining input and output of the model according to a multi-state collaborative estimation task, fusing the extracted aging characteristics with battery state parameters to jointly serve as input characteristics, and adjusting an input type; inputting and outputting a base model, improving a large language model, and adding an output layer module for the large language model to adapt to a multi-state collaborative estimation task; based on the improved large language model and the actual data set, adopting a low-rank adaptation method to finely adjust the model; and processing the target fragmented charging data based on the fine-tuned model, and carrying out multi-state collaborative estimation. According to the invention, multi-state collaborative accurate estimation can be realized.

The present invention discloses a method and device for feature extraction and matching of satellite video images, which can solve the problem of real-time or near-real-time application of satellite video. The satellite video image to be matched is determined, corner points are extracted from the satellite video image to be matched; the corner points are filtered; based on the filtered corner points, the satellite video image to be matched is characterized, and a feature vector representing the satellite video image to be matched is obtained; and the feature vectors of the satellite video image to be matched are matched.

This application provides a method, apparatus, equipment, and storage medium for identifying agricultural irrigation districts based on the irrigation-enhanced evaporation stress index (IESI), relating to the fields of agricultural water resources monitoring and management and remote sensing applications. The method includes: constructing a multi-source remote sensing and auxiliary dataset for the target area; calculating potential evapotranspiration and actual evapotranspiration and coupling them with a vegetation growth stability factor to obtain quarterly IESI data for the target area; determining irrigation identification thresholds applicable to plains and mountainous areas respectively; and discriminating the IESI of the target area to generate a binary spatial distribution map of irrigated and non-irrigated areas. This application solves the problems of existing irrigation area identification methods in large-scale applications, such as reliance on statistical data or large training samples, strong empirical thresholds, insufficient adaptability to terrain and regional differences, and complex calculation processes. It achieves irrigation behavior identification centered on evaporation stress response, ensuring computational efficiency while also considering spatial accuracy and regional adaptability.

A system and method for synchronously monitoring vital signs of multiple human bodies based on millimeter wave sensing, comprising: a millimeter wave radar front-end module, a microcontroller, a processor, and a display and recording module, wherein: the millimeter wave radar front-end module is used to transmit linear frequency modulated continuous wave microwave or millimeter wave signals, and receive reflected echo signals to generate multi-channel intermediate frequency baseband signals, the microcontroller sets and controls waveform parameters, transmission power and signal gain parameters and monitoring processes, sets and controls acquisition parameters of the signal acquisition module of the processor, the processor extracts vital sign parameter information of the human body from the multi-channel intermediate frequency baseband signal, and the display and recording module displays and saves information obtained by the processor, including: multiple human position information, chest wall micro-displacement information and vital sign parameters. The present invention provides a non-contact vital sign monitoring system and method with high measurement accuracy, strong anti-interference ability, and reliable and stable for synchronous monitoring of vital signs of multiple human bodies.

The utility model discloses an air exhaust cylinder with an air exhaust function and a vacuum chuck device, and the air exhaust cylinder comprises a shell, a detection box, a vacuum pump, a battery assembly, a circuit board, a control button and a sealing ring; the detection box is fixed at one end of the shell, the vacuum pump, the battery assembly and the circuit board are installed in the shell or in a space formed by the detection box and the shell, and the vacuum pump, the battery assembly and the control button are all electrically connected with the circuit board; a first hole is formed in the end part of the detection box, and the vacuum pump exhausts air through the first hole; and the sealing ring is arranged on the outer side wall of the detection box or the shell. The electric vacuum chuck is used for being installed on a vacuum chuck body, an elastic piston of a traditional manual vacuum chuck is replaced, and a common vacuum chuck can be changed into the electric vacuum chuck.

The invention discloses an image-based flame motion feature extraction method, and belongs to the technical field of flame feature extraction. Comprising the following steps: step 1, carrying out preprocessing operations of gray level conversion and filtering and denoising on a flame image collected by a camera in real time; 2, inputting two continuous frames of original images, carrying out downsampling to obtain an image pyramid, calculating a top velocity field, and then carrying out deformation operation on the images; 3, calculating a bottom layer image of the image pyramid by using an iterative calculation formula to obtain a flame optical flow velocity field; and step 4, calculating a velocity mean value and a velocity direction mean value of each frame of pixel point in the flame image, and completing flame motion feature extraction. The method can accurately and reliably extract the effective feature information of the combustor flame combustion process in space and time, qualitatively and quantitatively reflect the motion features of the combustion process, and improves the economy, safety and environmental protection of a power plant.

The present application relates to the technical field of energy storage, more particularly to a portable intelligent shared charging method, system, medium and equipment. The scheme comprises matching a user serial number according to iris and face information, collecting a remaining battery percentage, a current temperature of a battery storage device and a distance index; obtaining one of a basic power index or an advanced power index according to the remaining battery percentage; obtaining one of a basic safety index or an advanced safety index according to the current temperature of the battery storage device; calculating a comprehensive evaluation coefficient according to the classification level; and providing the battery storage device to the user according to the comprehensive evaluation coefficient. The scheme learns user habits and positions through iris and face assistance, and combines online analysis of energy storage system power and system state to intelligently recommend a power source, thereby adapting to user habits and facilitating use while ensuring system safety and device service life.

The invention discloses a satellite communication frame structure security enhancement method and system based on dynamic fragmentation embedding, and the method comprises the steps: carrying out the fragmentation processing of the control information of a satellite communication frame, and carrying out the LDPC error correction coding operation; a sending end collects a satellite state parameter P in real time, and a dynamic mapping rule R is generated through a hash function; the encoded information fragments are embedded into a QPSK symbol phase of a data field according to a dynamic mapping rule R to form an integrated frame and transmit the integrated frame, and meanwhile, a satellite state parameter P is sent to a receiving end in a quantum key encryption mode; a receiving end performs decryption to obtain a satellite state parameter P, reproduces a dynamic mapping rule R, extracts information fragments from the integrated frame, recombines the information fragments into a control information set C, and analyzes service data; and updating the dynamic mapping rule R according to the satellite orbit type and the channel change. The invention provides a technical scheme which does not have an independent frame header and is used for dynamically distributing and embedding control information, and dual security enhancement of frame header physical disappearance and control information non-localization is realized.

The application discloses a lightning arrester resistance current detection method, device and lightning arrester detection system. The method comprises the following steps: determining three-phase third harmonic current capacitive components and three-phase third harmonic current resistance components; determining three-phase third harmonic resistance currents of a three-phase lightning arrester according to the three-phase third harmonic current resistance components; determining three-phase fundamental wave resistance currents of the three-phase lightning arrester according to three-phase interphase capacitances, three-phase fundamental wave current resistance components and three-phase fundamental wave voltage effective values; and determining that the three-phase lightning arrester is in an abnormal state if any one of the three-phase third harmonic resistance currents is greater than or equal to a first preset threshold value or any one of the three-phase fundamental wave resistance currents is greater than or equal to a second preset threshold value. The method can effectively eliminate power grid harmonics and interphase capacitance interference in the case that the power grid contains harmonics, and can realize reliable and accurate extraction of lightning arrester resistance currents in normal and abnormal states of the lightning arrester.

This invention discloses a submillimeter-level diffusion magnetic resonance imaging (DTI) method for extracting the radial index of cortical layered cognitive structures, relating to the field of medical image processing technology. This invention aims to address the problems of existing technologies failing to accurately quantify the directionality of cortical microstructures and neglecting the heterogeneity of cortical layering. The method includes: acquiring DTI data and performing high-fidelity preprocessing such as MP-PCA denoising, Gibbs artifact removal, and Gaussian process distortion correction; estimating the diffusion tensor and extracting the principal feature vector based on the preprocessed data; reconstructing the cortical white matter and pia mater surface using T1 images and extracting the normal vector; calculating the absolute value of the dot product of the diffusion principal feature vector and the surface normal vector to construct a radial index map; and segmenting the cortex into multiple intermediate layers based on the principle of equal volume and extracting layering parameters. Through refined high-resolution data processing and geometric modeling, this invention can accurately quantify the radial alignment integrity of cortical nerve fibers, providing highly sensitive imaging markers for cognitive disorders such as schizophrenia.

The invention discloses a public facility detection method based on a convolutional neural network GoogleNet, and relates to the technical field of public facility detection, and the method comprises the following steps: collecting a public facility image data set, and carrying out the preprocessing of the data set; the preprocessed image data set is divided into a training set and a test set; and establishing a GoogLeNet network model, and inputting the preprocessed test set into the GoogLeNet network model to complete identification and detection of the public facilities. According to the method, two continuous 3 * 3 convolution layers are used for replacing an original single 5 * 5 convolution layer, and grouping convolution is used for replacing a normal convolution kernel, so that the parameter quantity and the calculation complexity of the network are reduced, the public facility feature extraction capability is improved, and meanwhile, the public facility identification calculation efficiency is also improved. And a Leaky ReLU activation function is introduced, so that the nonlinear processing capability of the network is enhanced, and a positive effect is achieved on the expression in public facility identification. Through standardized operation, the features of the public facilities are extracted more reliably, and gradient disappearance is prevented.

The present invention discloses a method, system, device and storage medium for automatically outlining and detecting the offset of a brain midline. The method comprises: obtaining an original brain CT image and then calibrating it to obtain a calibrated brain CT image; processing the calibrated brain CT image using a brain midline extraction model to obtain a feature image containing only a predicted brain midline; superimposing the feature image and the calibrated brain CT image to obtain a target brain CT image that automatically outlines the predicted brain midline; in the feature image, automatically generating an ideal brain midline according to the two endpoints of the predicted brain midline, and determining the offset of the predicted brain midline relative to the ideal brain midline. The present invention can accurately and reliably extract the slender and complex tubular structure feature of the brain midline contained in the brain CT image by introducing two different tubular convolutional layers in the encoder structure and the decoder structure contained in the brain midline extraction model.

The present invention belongs to the field of data processing technology, and in particular, relates to a method for extracting wind farm signal features and denoising and optimizing data. It includes: step 1. obtaining wind farm measurement signal parameters; step 2. obtaining EEMD parameters in the measurement signal; step 3. establishing a Hilbert-Huang transform model of the white noise clustering empirical method according to the measured signal parameters; step 4. analyzing the obtained signal to determine the filter window length; step 5. sampling the analytical signal for the number of points m; step 6. setting the number of iterations K; step 7. calculating the derivative value and average value of each sampling point of the analytical signal; step 8. calculating the weight of each point in the analytical signal filter window to denoise each signal point; step 9. iterating K times; step 10. obtaining accurate wind farm analytical signal characteristic values. The present invention can effectively and accurately extract signal features and denoise optimization of wind farms, is easy to implement, and has broad market application prospects.

The invention belongs to the technical field of image recognition and processing, and aims to provide an intelligent homework reviewing method and system, electronic equipment and a program product. According to the method, page structure adaptive recognition is realized through template matching, high-precision image correction is completed by using anchor geometric constraints, character recognition is performed through a single character recognition model, and objective question answering information is automatically scored. According to the invention, the problem of low recognition accuracy of subjective questions and complex answering scenes caused by recognition of handwritten texts of students depending on OCR in the prior art is solved, and the problem of low recognition accuracy of OCR caused by poor handwriting of students in the prior art is effectively solved; in addition, full-process high-precision intelligent reviewing from image acquisition, area positioning, character recognition to automatic objective question scoring and subjective question scoring summarization is realized, and homework reviewing efficiency and accuracy are greatly improved.

The embodiments of the present application relate to the field of map processing technology, and disclose a method for determining a target area, a self-mobile device, and a computer-readable storage medium. The method includes: obtaining a grid map, deleting an area between two unknown grids that meets the first type of inaccessible condition, deleting an area between an unknown grid and an obstacle grid that meets the second type of inaccessible condition, and searching for a connected domain containing the current machine position in the optimized map to obtain a target area. When a self-mobile device detects and constructs a grid map, it is easily affected by the position of the gap and obtains a slit area. The embodiments of the present application regard an area between two unknown grids that meets the first type of inaccessible condition, and an area between an unknown grid and an obstacle grid that meets the second type of inaccessible condition as a gap area, and delete both of them, thereby ensuring that a reliable target area is extracted.

The invention discloses an AWTA-ELCNN-based fault diagnosis method and diagnosis system. The method comprises the steps of obtaining a strong time-varying fault vibration signal of a rotating machine; inputting the strong time-varying fault vibration signal into the trained AWTA-ELCNN fault diagnosis model to obtain a fault diagnosis result of the rotating machine; wherein the AWTA-ELCNN fault diagnosis model adopts a single-layer architecture, and the AWTA-ELCNN fault diagnosis model comprises a scale adaptive wavelet convolution layer, an adaptive threshold time sequence attention enhancement layer, a global average pooling layer, a first full connection layer and a second full connection layer. Dynamic matching of strong time-varying fault features is realized through the scale adaptive wavelet convolution layer, feature enhancement and noise suppression are cooperatively performed through the adaptive threshold time sequence attention enhancement layer, and a single-layer lightweight architecture is adopted, so that the calculation efficiency is balanced while the time-frequency feature precision is guaranteed, and the diagnosis performance under a complex working condition is remarkably improved.

This invention provides a method for segmenting rock mass structural surfaces in a three-dimensional mesh model of the tunnel face, relating to the fields of three-dimensional laser point cloud processing and tunnel geotechnical engineering. The method involves collecting multimodal sensing data of the tunnel face during excavation, generating a colored three-dimensional mesh model based on this data, and performing tensor quantization on the colored three-dimensional mesh model data. Each vertex of the three-dimensional mesh model is encoded as a normal vector tensor, and the saliency of the vertices is determined based on the tensors to extract the vertex data of the structural surfaces. The structural surface vertices are then clustered, and structural surfaces are extracted from the obtained clustered vertices. Based on the orientation of the structural surfaces, they are grouped, and rock mass parameters are calculated. This invention solves the technical problems in existing technologies where structural surface rock mass segmentation easily leads to over-segmentation or under-segmentation, and where manual geological sketching is inefficient and inaccurate.

The invention provides a method for optimizing a shiny-leaved yellowhorn tea extraction process. Comprising the following steps: adopting a single factor test in combination with a response surface analysis method, selecting methanol concentration, liquid-material ratio and extraction time as investigation factors, taking total flavone and tea polyphenol content and extract yield as comprehensive indexes, and carrying out a verification test to finally obtain optimal process parameters; the optimal process parameters obtained by the method are as follows: the methanol concentration is 60%, the liquid-material ratio is 40: 1mL / g, and the extraction time is 35min. A verification test is repeated for three times according to the improved optimal process condition, the actual comprehensive score of the obtained shiny-leaved yellowhorn tea extraction process is 0.7033 and is close to a theoretical value 0.7130, the relative error is only 1.36%, and it is indicated that the method for extracting the process parameters through the model with the optimized response surface is quite accurate and reliable.

The invention discloses a legal map construction method based on multi-subject behavior recognition and role extraction, and belongs to the technical field of natural language processing and judicial intelligent analysis. The technical problems that multi-subject behavior semantics in criminal case texts are difficult to completely recognize, master-slave roles are inaccurate to judge, and behavior cross relations are difficult to extract are solved. According to the technical scheme, the method comprises the following steps of 1, data acquisition and preprocessing; step 2, semantic role labeling SRL; step 3, multi-level semantic feature extraction; 4, performing combined extraction on entities and relationships; step 5, behavior cross relation identification; and step 6, knowledge storage and application. According to the method, the behavior logic and role relationship of multiple subjects can be accurately described in a single case context, and efficient data support is provided for judicial semantic understanding, evidence chain construction and sentencing assistance.

The utility model discloses a stock solution extraction module which comprises a mounting table, a stock solution bottle cap breaking mechanism, an extraction mechanism, a transfer bottle cap taking and placing mechanism, a transfer bottle capping mechanism and a transfer bottle shifting mechanism, wherein the stock solution bottle cap breaking mechanism, the extraction mechanism, the transfer bottle cap taking and placing mechanism and the transfer bottle capping mechanism are sequentially arranged along a processing production line and are arranged on the mounting table; the industrial personal computer is electrically connected with the stock solution bottle cap breaking mechanism, the extracting mechanism, the transfer bottle cap taking and placing mechanism, the transfer bottle capping mechanism and the transfer bottle shifting mechanism; the transfer bottle shifting mechanism is used for carrying the stock solution bottles and the transfer bottles and conveying the transfer bottles; the stock solution bottle cap breaking mechanism is used for breaking caps of stock solution bottles, and the extracting mechanism is used for extracting stock solution; the transfer bottle cap taking and placing mechanism places bottle caps and bottle cap plugs into the transfer bottles, and the transfer bottle capping mechanism is used for capping. By means of the scheme, the device has the advantages of being simple in structure, reliable in extraction and the like, and has high practical value and popularization value in the technical field of stock solution extraction.

The utility model discloses a kind of power supply circuit of photovoltaic sensor, including DC-DC power supply chip, voltage sampling circuit is connected with the input end of DC-DC power supply chip, the output end of DC-DC power supply chip is connected with freewheeling circuit, one end of positive feedback circuit is connected in the input end of DC-DC power supply chip, and the other end is connected in the output end of freewheeling circuit.The utility model is sampled in real time to photovoltaic panel input voltage by voltage sampling circuit, and hysteresis threshold is formed using positive feedback mechanism, ensure that power supply is stable and reliable operation near starting voltage critical point, effectively prevent the system frequent start-stop problem caused by photovoltaic panel output voltage fluctuation due to illumination fluctuation, to ensure that power supply circuit can efficiently, stably extract photovoltaic panel energy in very wide input voltage range, provide continuous reliable electric energy for rear-end sensor.

The utility model provides a kind of ion exchanger unloading system and ion exchange system, the ion exchanger unloading system includes ion exchanger, infusion assembly and suction assembly, ion exchanger includes reaction cavity and filter screen, reaction cavity is divided into first chamber and second chamber, first chamber and second chamber are oppositely located the two sides of filter screen;Infusion assembly includes liquid storage bin, infusion pump and infusion pipeline, one end of infusion pipeline is connected liquid storage bin, the other end of infusion pipeline is connected second chamber, infusion pump is located on infusion pipeline, and infusion pump is used to make the liquid in liquid storage bin flow to second chamber;Suction assembly includes suction pipeline, resin storage bin and suction pump, one end of suction pipeline is connected first chamber, the other end of suction pipeline is connected resin storage bin, and suction pump is located on suction pipeline, and suction pump is used to make the resin in first chamber flow to resin storage bin.The utility model can improve the efficiency of resin replacement, improve the continuity and stability of production.

The application provides a magnetic suspension rotor fundamental frequency vibration signal real-time monitoring method and system, and belongs to the technical field of magnetic suspension bearings; the method comprises the following steps: collecting an original displacement signal; using an adaptive wave trap to pre-process the displacement signal; selecting an initial voltage signal according to the real-time rotating speed frequency of the magnetic suspension rotor V c , and inputting the pre-processed displacement signal into a phase-locked loop unit; the phase-locked loop unit is used for tracking the fundamental frequency component in the displacement signal which is synchronous with the rotating speed, and outputs an instantaneous frequency value and an instantaneous phase value which are synchronous with the fundamental frequency component, and two reference signals; a synchronous demodulator is configured, and based on the instantaneous frequency value, the instantaneous phase value, the two reference signals and the original displacement signal output by the phase-locked loop unit, the real-time amplitude of the fundamental frequency vibration component is extracted; and a rotor unbalanced vibration state vector including the real-time amplitude of the fundamental frequency vibration component, the real-time instantaneous phase value and the real-time instantaneous frequency value is output.

The application belongs to the technical field of machine learning, and specifically discloses a plateau wetland landscape component extraction method, system and program product. Through the design of a spectrum-space double-branch interaction model, the spectral features of the image elements are extracted through the image element dimension increasing module and the ViT architecture Transformer encoder in the spectrum branch, the neighborhood spatial features of the image elements are extracted through the multi-scale feature extraction module and the feature enhancement in the space branch, then the features of the two branches are integrated through the cross attention fusion module, and the image element classification result is output, so as to determine the plateau wetland landscape component regions in the remote sensing image according to the image element classification results, and the efficient and reliable extraction of the plateau wetland landscape components can be realized. The application can effectively alleviate the salt and pepper noise problem existing in the traditional pixel-by-pixel classification method, improve the classification accuracy of the plateau wetland landscape components, ensure the lightweight of the network, and provide more accurate technical support for the long-time sequence plateau wetland landscape component extraction.

The invention discloses a carrier speed estimation method and device based on a millimeter wave radar, and a storage medium, and relates to the technical field of carrier speed estimation. The method comprises the following steps: firstly, carrying out interference suppression processing on a distance-velocity spectrum matrix to obtain an interference suppression spectrum matrix; the first path performs velocity dimension energy aggregation on the interference suppression spectrum matrix to generate a static background energy distribution spectrum and determine a first velocity estimation value according to the static background energy distribution spectrum; a second path screens and verifies an effective ground reflection target from the interference suppression spectrum matrix, and a second speed estimation value is obtained by combining the pitch angle of the effective ground reflection target; and finally, respectively carrying out time sequence filtering and weighted fusion on the outputs of the two paths, and outputting the final carrier speed. According to the method, the problem that a traditional point cloud method fails in an open scene is fundamentally solved, and the robustness, accuracy and stability of a spectrum analysis method in a complex interference environment are remarkably improved.

The application discloses a T-shaped corner weld identification method and system based on laser vision, and belongs to the technical field of automatic welding and machine vision, and comprises the following steps: collecting images, performing enhancement processing on the collected images, and extracting an effective laser stripe area; center lines are extracted from the effective laser stripe area, and the center line coordinates are converted into three-dimensional point cloud data; the three-dimensional point cloud data is subjected to geometric feature optimization processing, and a RANSAC algorithm with spatial geometric constraints is used to fit and identify the geometric features of a T-shaped corner weld structure; based on the geometric features of the T-shaped corner weld structure, the three-dimensional coordinates of the key nodes of the weld are calculated through linear end point distance constraints, three-dimensional intersection calculation and Z-axis height screening. The application can realize automatic and high-precision acquisition of the position of the weld, realize automatic conversion of the weld from a two-dimensional image to a three-dimensional robot coordinate system, and significantly improve the degree of welding automation.