148 results about "Gaussian function" patented technology

The invention discloses an SPH fluid rendering method and system based on a Gaussian function, and belongs to the technical field of computer graphics, and the method comprises the steps: receiving SPH simulated fluid particle data, and constructing a particle connection graph based on a particle neighborhood relation; based on the particle connection graph, extracting local geometric features of each particle through a neural network and aggregating the local geometric features into global geometric features; analyzing the global geometric features into Gaussian function parameters to generate a plurality of Gaussian functions for each particle, and superposing all Gaussian functions in space to form an implicit fluid surface; performing screen space rendering on the implicit fluid surface to obtain a final fluid rendering image; and after foam particles generated by SPH simulation are also mapped into a Gaussian function, a fluid rendering image containing a foam visual effect is obtained through screen space rendering. The method can improve the smoothness and reality of the rendering result while guaranteeing the calculation efficiency, and achieves the consideration of the efficiency and quality of fluid rendering.

A set of images of a scene re received. Each image includes temporal data and spatial data relating to the scene. Based on the spatial data of each image, three-dimensional (3D) Gaussian splatting data is generated. The temporal data of each image and the 3D Gaussian splatting data are inputted to a neural network to generate spatial-temporal 3D Gaussian embeddings. Offset data based on the spatial-temporal 3D Gaussian embeddings is generated. The video of the scene is rendered based on the 3D Gaussian splatting data and the offset data, allowing for improved rendering of video of the scene.

The invention relates to the technical field of unmanned aerial vehicle navigation and control, in particular to an improved sparrow searching method for three-dimensional unmanned aerial vehicle path planning. According to the method, a three-dimensional mountain terrain model is constructed through a Gaussian function superposition method, obstacles and no-fly zones are defined by adopting a cylindrical geometric model, and obstacle avoidance constraints, take-off and landing safety constraints and flight height constraints are set; a comprehensive cost function with path length, threat penalty, flight safety height and path smoothness as sub-targets is constructed, an improved sparrow search algorithm is introduced, and a sine and cosine strategy is adopted in discoverer position updating to dynamically balance horizontal exploration and vertical obstacle avoidance. A Levy flight mechanism is introduced in follower position updating so as to enhance the capability of local development and jumping out of local optimum; the method comprises the steps of population initialization, role division, iterative optimization, collision detection and an elitist retention strategy. According to the method, a safe and smooth three-dimensional flight path with the optimal comprehensive cost is output, and the method is suitable for an actual unmanned aerial vehicle navigation task.

The invention discloses an electric power engineering construction quality evaluation method and device based on three-dimensional Gaussian modeling, equipment and a medium. The method comprises the steps of abstracting a plurality of pieces of equipment in electric power engineering into a physical parameter group, and performing multi-view image fusion based on images of the electric power engineering; modeling is carried out based on a 3DGS three-dimensional scene, parameterized Gaussian distribution is used for representing the geometric morphology of the power equipment, and mapping of physical parameters and the three-dimensional scene is achieved; projecting the 3DGS model into a two-dimensional rendered image based on differentiable rendering, and realizing gradient propagation from pixels of the two-dimensional rendered image to physical parameters; constructing a loss function, and optimizing Gaussian function parameters and physical parameters based on the image of the electric power engineering and the two-dimensional rendering image; and after the optimization is completed, carrying out physical parameter extraction on the 3DGS model corresponding to the electric power engineering so as to realize quality evaluation, graded early warning or visual display. The invention belongs to the field of engineering evaluation. According to the invention, accurate evaluation of the construction quality can be realized.

The invention provides a multi-frequency time-domain interference stimulation simulation optimization method and device. The method comprises the following steps: constructing a corresponding Gaussian function according to time-domain interference stimulation parameters for a target clinical user, and obtaining Fourier series of each channel of a transcranial electrical stimulator; generating a brain model based on the brain T1 weighted imaging data, and setting corresponding target spot position information to optimize the target spot electric field direction in any direction; determining an effective electrode arrangement area of each axial section of the craniocerebral model and arranging electrodes; and optimizing the electrodes arranged on each axial section so as to determine each grid of the craniocerebral model and the intensity amplitude of the time domain interference stimulation electric field, and optimizing the input current of each channel. According to the invention, multi-frequency stimulation simulation optimization for each channel of the transcranial electrical stimulator can be realized, the focusing performance of time domain interference stimulation can be improved, the calculation complexity of time domain interference stimulation parameter configuration can be reduced, and personalized time domain interference stimulation simulation can be effectively and quickly realized in combination with clinical data.

The invention discloses an offshore wind power plant regular layout optimization method considering submarine topography and turbulence intensity, and the method comprises the steps: carrying out the grid discretization of a target region, and defining a layout rotation angle; a mixed integer linear fractional programming model with the minimum leveling cost per kilowatt-hour as the target is constructed, a wake flow model based on an axisymmetric cosine function and a turbulence model based on an axisymmetric double-Gaussian function are comprehensively integrated, and piecewise linearization is conducted on a turbulence nonlinear term; the method comprises the following steps: pre-defining a layout mode and regular layout constraints, and matching with engineering constraints such as fan number, power and safety distance; and a bounded Dinkelbach algorithm is adopted to carry out iterative solution on the sub-problems decomposed according to angles, and a global optimal layout is obtained. According to the method, wake flow loss, turbulence fatigue and terrain-related cost can be effectively and synergistically optimized, and an economical and efficient layout scheme is provided for offshore wind plant planning.

The invention discloses a virtual reality level simulation method based on a material point method and gridding surface reconstruction, and belongs to the crossing field of computer graphics, physical simulation and virtual reality technologies. The method comprises the following steps: S1, classifying and storing position coordinates of material points according to material numbers of the material points, and traversing the coordinates to obtain a boundary of a bounding box of the material points; s2, selecting a grid size and a boundary extension distance for reconstructing the surface in a self-defined manner, and generating a grid; s3, selecting a Gaussian function to calculate the contribution of each substance point to the overall substance point density field, and obtaining parameters; s4, calculating the contribution of each substance point to the overall substance point density field, and performing accumulation to obtain the overall substance point density field; and S5, obtaining a contour surface of the overall material point density field, and exporting the contour surface as a rendering engine general three-dimensional model to obtain a simulation visual effect. According to the method, the problems that in the prior art, tens of millions of substance points cannot be processed, reconstruction parameters are difficult to select, and connection with a high-end rendering engine is not smooth can be solved, and the virtual reality level simulation visual effect is achieved.

The present application relates to the technical field of aquaculture, and particularly relates to an improved PSO-based water quality parameter prediction method and system for aquaculture, which comprises collecting water quality parameters at different positions and depths in a breeding pond; training an improved radial basis function (RBF) neural network using training set data; and optimizing the parameters of the improved RBF neural network model using an improved particle swarm optimization (PSO) algorithm. The present application introduces a mixed Gaussian function and an abnormal S-shaped function into the radial basis function of the traditional RBF neural network, thereby solving the problem of weak capability of the model in nonlinear data modeling. Furthermore, the present application improves the inertia factor and the learning factor in the traditional PSO algorithm, thereby solving the problems of slow parameter convergence speed and poor global search capability in the RBF neural network.

The invention discloses a comprehensive assessment method for weak links of a power system involving multi-dimensional indexes in extreme weather. The method comprises the following steps: firstly, establishing a multi-dimensional index system containing safety and optimized operation, wherein the system covers the line outage probability, the load loss probability, the electric energy insufficiency expectation, the line heavy load rate and the network loss rate; secondly, constructing a line fault model and a cascading fault model under extreme storm weather, describing fault probability change in combination with a Gaussian function and Weibull distribution, and simulating line segment failure by adopting Poisson distribution; a system state after a fault is calculated through a direct-current optimal power flow model, and a large amount of simulation data is obtained in combination with a non-sequential Monte Carlo method. Furthermore, a fuzzy analytic hierarchy process and a CRITIC method are adopted to calculate subjective and objective weights respectively, subjective and objective comprehensive evaluation is realized through weight fusion, and weak links of the system are identified. The method gives consideration to evaluation scientificity and calculation efficiency, and is suitable for real-time risk evaluation and weak link identification of a large-scale power system in extreme weather.

This invention provides a method and system for detecting impurities in flunarizine hydrochloride raw material. The detection method includes: acquiring the original chromatographic signal, performing empirical mode decomposition for denoising and baseline optimization to obtain a baseline-corrected and denoised chromatogram; determining the signal-to-noise ratio based on the energy relationship between noise and signal in preprocessing, determining the optimal analytical scale through a preset function, determining the number of chromatographic peak components by screening ridges through continuous wavelet transform, and extracting initial model parameters; establishing a mathematical model for the convolution of a standard Gaussian function and a modified Logistic distribution function for each component peak, using a trust-region reflection algorithm for nonlinear iterative fitting, with the objective function containing a least-squares residual term and a weighted regularization term; after the objective function converges, obtaining the optimal model parameters, and calculating the integral area of ​​each impurity component peak to determine its content.

The application discloses a kind of vacuum switch contact ablation intelligent evaluation method based on physical information fusion, belong to vacuum circuit breaker technical field.The method of the present application includes: according to the data set of vacuum switch breaking type obtained from vacuum arc experimental data, construct physical information enhanced support vector machine to predict breaking type of working condition;Vacuum arc simulation model is extracted from each working condition Contact surface energy flow density radial distribution data, and is fitted by Gaussian function;Based on the energy flow density radial distribution curve obtained by fitting, a vacuum arc simulation data set is constructed, and a physical information enhanced Gaussian process regression is constructed to predict the energy flow density of the working condition;Combining the heat transfer model to solve the radial distribution of contact surface temperature, further quantitatively evaluate the ablation degree of contact surface, and output the ablation degree level.The present application embeds the physical law of vacuum arc into machine learning model, realizes the breaking type prediction, energy flow density prediction and contact temperature rise calculation and ablation degree calculation and evaluation.

This invention discloses a large image processing method and system based on adaptive slicing and feathering fusion, comprising: acquiring the original image and parameters; adaptively determining the slicing parameters through iterative calculation with a minimum overlap width requirement as a constraint, and outputting slicing configuration information; expanding the boundaries of the original image according to the filling amount, and extracting multiple image slices with sizes conforming to the input size of the target processing unit from the expanded image; and performing feathering fusion and layered stitching on the image slices using a weight mask generated based on a Gaussian function based on the overlap size to reconstruct a final image with the same size as the original image. This invention calculates the minimum slice grid layout required to meet seamless stitching quality requirements. Through the synergy of reflection filling and Gaussian weighted feathering fusion techniques, it achieves imperceptible pixel value changes at the seams while minimizing the total number of calls to downstream processing units, thereby resolving the contradiction between image quality and computational cost.

The demodulation method based on the ultra-weak fiber grating OFDR includes: performing fast Fourier transform on the time domain reflection signal collected by the OFDR system to obtain a frequency domain spectrum; setting an amplitude threshold, identifying the UW-FBG array reflection peak in the frequency domain spectrum, calibrating the spatial position of the sensing unit, and intercepting the window segment spectrum signal corresponding to each UW-FBG; performing inverse Fourier transform on the intercepted window segment spectrum signal to restore the spectrum signal of a single UW-FBG; for the reconstructed spectrum signal, selecting a fitting region containing the Bragg reflection peak, iteratively optimizing the Gaussian function fitting parameters by the least square method, and demodulating the peak Bragg wavelength; traversing the window segment spectrum signals of all UW-FBGs and repeating the above steps to obtain the Bragg wavelength shift distribution of the whole sensing region. The demodulation method based on the ultra-weak fiber grating OFDR improves the high-precision detection characteristics of the OFDR, realizes dynamic demodulation, significantly improves the sensing demodulation rate, and simplifies the algorithm complexity.

The application discloses a path planning method and device, a vehicle and a storage medium. The method comprises the following steps: acquiring obstacle information, the obstacle information being used for indicating the position and shape of an obstacle in an environment in which a vehicle is located; acquiring vehicle speed information; determining a target region containing the obstacle based on the obstacle information and the vehicle speed information; and determining a planning path based on a two-dimensional Gaussian function and the target region, wherein the planning path does not coincide with the target region. The technical scheme provided in the embodiments of the application can provide accurate driving guidance when the vehicle has an obstacle avoidance requirement, improve the accuracy of path planning in an obstacle avoidance scene, and thus improve driving safety.

The present application relates to the technical field of image data processing, and more particularly to an elevator car passenger distribution monitoring method for elevator energy feedback optimization, comprising: acquiring an overhead image of an elevator car, identifying passenger shape centers and quantity, constructing a deformation action matrix, generating a deformed two-dimensional Gaussian function, superimposing to generate a quality density field, calculating the coordinates of the actual centroid of the elevator car, based on the coordinate deviation of the actual centroid and the geometric center of the elevator car, calculating the additional friction power loss, dynamically compensating the energy feedback control parameter, and completing the elevator car passenger distribution monitoring for elevator energy feedback optimization.

This application discloses a downhole nuclear magnetic resonance (NMR) inversion method, apparatus, equipment, and storage medium based on pruned Bayesian methods, relating to the field of NMR data processing technology. The method includes: constructing a forward dataset based on fluid relaxation time and acquisition parameters, and adding Gaussian white noise to obtain the inversion input signal; initializing prior parameters, posterior mean, and variance; constructing a kernel matrix based on relaxation time points; constructing a kernel-dictionary matrix using Gaussian functions; updating the fitted echo data and fitting residuals using hierarchical Bayesian methods, thereby updating the prior parameters; updating the posterior mean through evidence maximization and calculating the proportion of each fluid component; pruning the kernel-dictionary matrix according to a preset pruning threshold and proportion, iterating until a termination condition is met, and finally mapping to obtain a two-dimensional NMR relaxation spectrum. This method solves the problem of waiting time and loss of diffusion relaxation dimension information caused by improved logging efficiency, which makes accurate multidimensional NMR logging inversion difficult.

The present application belongs to the field of image classification, and particularly relates to a hyperspectral image classification method. The hyperspectral image classification method of the present application obtains Retinex features of different scales by selecting different standard deviation values of Gaussian functions, and fuses the Retinex features to form fused space-spectrum features for hyperspectral image classification, so as to take into account the global features and the detail features of the hyperspectral image, and thus the fused features used for classification contain more image information, thereby improving the classification accuracy. Moreover, the standard deviation values of the Gaussian functions capable of making the corresponding single-scale Retinex features reach the maximum classification accuracy and the adjacent standard deviation values thereof are selected to obtain Retinex features and fusion, so as to obtain feature values with higher classification accuracy. Finally, through comparison with the classification results of the spectral features, the single-scale Retinex features and other classification methods, it is verified that the classification method of the present application has higher classification accuracy compared with other methods.

The application discloses a kind of scintillator decay time test methods based on waveform sampling, the steps of test method include: scintillator is coupled to the optical window of photomultiplier tube placed in shielding box, using radioactive source excitation sample, open oscilloscope automatic acquisition not less than 100 pulse signals and transmission to control computer;Pulse waveform data acquired by oscilloscope is handled calculation, obtains the decay time constant corresponding to single pulse data;The above steps are repeated for processing all collected pulse data, obtain decay time constant set;With decay time constant set histogram, data should present normal distribution;Using Gaussian function to the normal distribution is fitted, the decay time value corresponding to the peak obtained is the decay time of the sample.The application can accurately test the decay time of different scintillating materials, especially suitable for high-throughput decay time measurement.

The application relates to a long-term deformation MT-InSAR adaptive fusion measurement method and device for a coal mining subsidence area surface, wherein the method comprises the following steps: obtaining surface deformation monitoring data by using multiple modes respectively, and performing at least one of time-space alignment, noise removal and error correction on the surface deformation monitoring data to obtain surface deformation data; the coherence index of the surface deformation data is calculated; a bimodal Gaussian function model constructed based on the deformation law of the coal mining subsidence area is used to generate a spatial adjustment coefficient; an adaptive weighting method is constructed based on the coherence index of the surface deformation data and the spatial adjustment coefficient to calculate the weight of multi-source InSAR data; and the surface deformation monitoring data is weighted and fused by using the weight of the multi-source InSAR data to obtain the final surface deformation result. Therefore, the problem that a single InSAR technology cannot comprehensively reflect the time-space evolution characteristics of long-term deformation of a wide-area coal mining subsidence area surface, resulting in insufficient monitoring accuracy, incomplete coverage, poor adaptability and the like is solved.

This invention relates to an edge localization method and system for optical communication chips. The method includes: performing multi-directional gradient convolution detection on an image of the optical communication chip to obtain an edge gradient map; based on the gradient magnitude of the edge points, selecting a fitting strategy of a quadratic curve or Gaussian function to measure the sub-pixel offset of the edge points and performing directional compensation based on the offset to obtain a sub-pixel edge point set; traversing the sub-pixel edge point set, and performing directional classification and decoupling of the sub-pixel edge point set by comparing the offset angles of candidate corner points and filtering by direction; performing region-aware separation of outliers in the decoupled sub-pixel edge point set based on the mode of coordinates to locate the main peak region, and iteratively fitting straight lines through neighbor point sampling and a multi-dimensional scoring strategy to obtain edge straight lines in each direction; performing intersection point calculation and geometric constraint optimization on the edge straight lines in each direction to generate the edge localization result. This invention can achieve high-precision and robust sub-pixel-level edge localization of optical communication chips.

The invention relates to the technical field of photoelectric display, and provides an area array chromaticity data multi-point correction method based on a two-dimensional Gaussian basis RBF interpolation method, and the method comprises the steps: obtaining a first chromaticity data array of a measured surface and second chromaticity data of M correction points in the measured surface; constructing a two-dimensional Gaussian radial basis function interpolation model sharing a group of shape parameters based on the positions of the correction points; solving a weight parameter of each Gaussian basis function according to the first chrominance data array and the second chrominance data at the correction point; and performing point-by-point correction on the first chrominance data array by using the solved model. According to the method, fitting precision and smoothness of chromaticity space nonlinear change are effectively balanced through adaptive shape parameters, the chromaticity relation consistency is kept by adopting a multi-channel coupling correction mode, and correction precision and efficiency are effectively improved.

The invention discloses a method for generating a hollow light array with rectangular distribution. The method comprises the following steps: constructing a kernel function based on a spiral phase and a Gaussian function; superposing hyperbolic cosine functions to construct a weighting function so as to obtain a multi-cosine correlation vortex light field; introducing a partially coherent light construction formula, and constructing a cross spectral density function of the multi-cosine correlation vortex light field at the initial plane in combination with the kernel function and the weighting function; a free space transmission formula is introduced, a light intensity expression of the multi-cosine correlation vortex light field at any transmission distance is constructed in combination with a cross spectral density function, parameters of the transmission distance and the multi-cosine correlation vortex light field are adjusted, and light intensity distribution regulation and control of the multi-cosine correlation vortex light field are obtained. According to parameter regulation and control, the array form that the sub-beams are hollow beams is obtained, and a rectangular distribution hollow light array is obtained; according to the invention, the controllable transmission of the multi-cosine Gaussian correlation vortex beam can be realized, and the effective regulation and control of the array distribution with the sub-beams being hollow beams can be realized.

This invention discloses a method for enhancing the terahertz absorption spectrum of materials. The method uses a terahertz time-domain spectroscopy system to measure the material spectrum, calculates the material's transfer function, amplitude, and phase using the acquired reference signal and sample signal, fits the material's absorption spectrum by constructing a multi-Gaussian function model, and enhances the absorption spectrum through a coefficient enhancement factor. This strengthens the ability to identify minute differences in the terahertz spectrum and improves the ability of terahertz spectroscopy to distinguish substances with similar properties, similar compositions, and trace amounts.

The application discloses a spraying robot nozzle flow adjustment method based on a Gaussian convolution kernel and reinforcement learning, and comprises the following steps: S1, measuring two-dimensional plane data of the distance between a spraying robot nozzle and an unsprayed wall surface; S2, defining a nozzle flow two-dimensional distribution function; S3, defining a two-dimensional Gaussian function as a convolution kernel; S4, weighting the convolution kernel by using the nozzle flow function, and performing convolution operation on the distance data and the two-dimensional Gaussian function; S5, calculating the roughness of the wall surface after spraying according to the convolution operation result; S6, optimizing the nozzle flow function by using reinforcement learning according to the roughness, and stopping the optimization when the roughness reaches the requirement; and S7, outputting the nozzle flow function result and spraying the wall surface according to the result. The method of convolution calculation and reinforcement learning is used to adjust the nozzle flow of the spraying robot, the problem of 'fixation and lagging adjustment' of the nozzle flow control can be solved, and thus good spraying effect is realized.

A polyolefin-polystyrene-based multiblock copolymer and method of making the same are disclosed herein. In some embodiments, a polyolefin-polystyrene-based multiblock copolymer satisfies conditions (a) a weight average molecular weight is 100,000 to 300,000 g / mol; (b) molecular weight distribution is 1.5 to 3.0; (c) measured results of gel permeation chromatography, a graph having an x-axis of log Mw and a y-axis of dw / dlog Mw, are fit to a Gaussian function, where all constants satisfy −0.01<A<0.03, 4.8<B<5.2, 0.8<C<1.2, and 0.6<D<1.2; and (d) a polyolefin block comprises one or more branch points, where a carbon atom at the branch point is represented by a peak of 36 to 40 ppm, and a terminal carbon atom of a branched chain from the branch point is represented by a peak of 13 to 15 ppm.

The invention discloses a color image vision enhancement method. The method comprises the following steps: acquiring an original color image and separating three color components; a Laplace operator is used to sharpen each component image, a Sobel operator is used to extract horizontal and vertical gradients respectively, the gradients are transformed based on a Gaussian function and normalization processing is completed, natural logarithm operation is carried out on an original component image and a normalized gradient image respectively, two obtained logarithm domain images are added and then exponent recovery is carried out, and a final component image is obtained. Restoring the component image to a gray domain, and performing Sigmoid function transformation on the restored component image; and calculating the global maximum and minimum values of the Sigmoid enhanced images of the three component images, performing linear stretching, and finally combining the stretched color component images to obtain a visual enhanced color image. According to the invention, the permeability of the color image can be improved, the brightness of the color image under the condition of low illumination can be improved, and the color effect matched with the visual habit of human eyes can be kept.

The invention provides an intelligent monitoring and early warning method and system for real-time ship dragging, and the method comprises the steps: firstly obtaining trajectory data in a normal anchoring state based on a ship control simulator, extracting a bias width, a to-ground navigational speed and a heading change rate as dragging monitoring indexes, and constructing a standard cloud model of each index through employing a Gaussian function; inputting the indexes collected in real time into the corresponding cloud model to obtain a membership degree, and calculating a real-time dragging probability in combination with an entropy weight method; net resultant force is calculated based on wind, flow and anchor chain parameters, and the drift velocity vector range after anchor dragging is estimated; aIS data are used for fitting an actual anchor position, a drift end point set is generated through Monte Carlo simulation, and four drift danger level intervals of danger, partial danger, partial safety and safety are divided; generating a ship space density distribution diagram covering the interval in combination with kernel density estimation; finally, the dragging probability and density information are fused to calculate a comprehensive risk value, acousto-optic grading early warning is triggered according to a preset threshold value, and high-precision and early-warning intelligent prevention and control of the dragging risk are achieved.

The application relates to the technical field of supply chain scheduling, and discloses a supply chain scheduling method and system based on AI, which comprises the following steps: acquiring order, resource, logistics topology data and historical performance data of a supply chain; generating a delivery uncertainty embedding vector by using a graph neural network; determining the center time and slope of an asymmetric trapezoidal function and the mean and variance of a Gaussian function according to the delivery uncertainty embedding vector, and constructing a fuzzy time window for a scheduling task of each destination node; acquiring a reinforcement learning model, and outputting a weight combination of a cost term, a time term and a robustness term in a scheduling target function; determining the scheduling target function according to the weight combination, and solving the scheduling target function; and performing cross operation of a genetic algorithm, wherein the cross operation is key subgraph exchange, and the key path subgraphs of two parent scheduling schemes are identified and exchanged to realize the cross operation. The application can not only accurately depict complex performance risk distribution, but also adapt to real-time changes, and the calculation amount is reduced.

This invention relates to a method, device, and electronic equipment for underwater vehicle formation control. The method includes: constructing a formation configuration maintenance framework based on rigid graph theory, and determining the relative positional relationships between vehicles through distance constraints; establishing an obstacle avoidance control model based on the artificial potential field method, calculating the potential field force based on the distance between the vehicle and the obstacle and the distance between vehicles, thereby achieving obstacle avoidance and collision prevention between vehicles; performing online identification and compensation of unmodeled dynamics and external disturbances in the vehicle dynamics model based on a radial basis function neural network, using a Gaussian function as the activation function; and vector-fusing the formation maintenance control force, potential field force, and compensation output of the radial basis function neural network to generate control commands for each vehicle, thereby achieving formation control. This invention can achieve high-precision formation maintenance, effective obstacle avoidance, and adaptive disturbance compensation capabilities in complex underwater environments.

The invention relates to the technical field of image correction, in particular to a cloth image correction method under the condition of uneven cloth feeding speed of a cloth inspecting machine. According to the method, a to-be-corrected cloth arch area is screened out through extreme value analysis in an image, and a Gaussian function is used for simulating a cloth arch shape; determining a distance to be subjected to interpolation correction through the shape of the Gaussian Bush arch simulation curve corresponding to segments; and based on the correction distance and the total number of to-be-interpolated values, determining the number of to-be-interpolated values under the segments for interpolation, and correcting the pixel values of the whole pixel points on the to-be-corrected cloth arch area according to the deviation of the pixel values after interpolation to obtain corrected image data. According to the method, the to-be-corrected area is determined through the light and shadow influence, and the Gaussian curve is used for simulating the shape of the cloth arch for correction, so that the completeness of the cloth cover under the condition that the cloth arch wrinkles are generated is improved, subsequent more accurate flaw detection is facilitated, and the accuracy of cloth cover quality evaluation is improved.