27 results about "Nonlinear interpolation" patented technology

The invention discloses a three-order precision nonlinear interpolation smooth factor design method based on a five-point template, and belongs to the field of computational fluid mechanics, a high-precision format realizes unification of high precision, no oscillation and multi-scale adaptability in complex flow simulation through nonlinear weighting, multi-order combination and smooth factor design, and the method has the advantages of high precision and high precision. The device is an indispensable numerical tool for aerospace. Wherein the smooth factor is a core mechanism for realizing shock wave capture and high resolution in a high-precision format, and weight distribution is dynamically adjusted through smoothness of a quantitative solution, so that balance among shock wave capture, high precision and low dissipation is achieved. Therefore, the invention provides the Z-type third-order precision nonlinear interpolation smooth factor which is based on the five-point template and meets the condition that the optimal precision of the nonlinear weighting framework of the multi-order candidate template is sufficient. A new consistent high-order smooth factor meeting the optimal interpolation precision is designed by using a consistent high-order smooth factor method.

The invention relates to the technical field of vertical shaft guide detection, in particular to an industrial intelligent visual detection and diagnosis system for a vertical shaft guide. The coal dust shadow quantization unit extracts a coal dust region by adopting a double-peak self-adaptive threshold segmentation combined region growing algorithm, and introduces a space weight factor to calculate an average gray value and a coverage area proportion, and the dynamic coupling calibration unit depends on a three-dimensional lookup table calibrated by multiple working conditions and nonlinear interpolation; the optimal deviation compensation coefficient of the jitter and coal dust combination is matched, a dynamic gray segmentation threshold value is generated, and a defect diagnosis decision unit accurately calculates the abrasion depth through partition self-adaptive segmentation, multi-scale morphological filtering and denoising and gradient amplitude and curvature double-constrained sub-pixel edge detection. A hierarchical control instruction is output according to the real-time depth and the historical trend, coupling interference is eliminated, the cage guide abrasion detection accuracy is improved, and a reliable basis is provided for safe operation and maintenance.

The invention discloses a method for determining a parameter lookup table of a neural network activation function, a method and system for calculating the neural network activation function based on the parameter lookup table and a storage medium, and belongs to the technical field of neural network accelerators, and the method comprises the steps: dividing the neural network activation function into a linear region and a nonlinear interpolation region; dividing the nonlinear interpolation region into a plurality of sub interpolation regions; performing linear fitting on the multiple segments of sub-interpolation regions to obtain slopes and intercept of multiple linear functions in one-to-one correspondence with the multiple segments of sub-interpolation regions; and determining a parameter lookup table according to the input variable value range of the neural network activation function corresponding to the multiple segments of sub-interpolation regions and the corresponding relationship between the slopes and intercept of the multiple linear functions, and storing the parameter lookup table in the memory. According to the method, the activation function is divided into the linear region and the nonlinear interpolation region, the activation function of each partition is approximately calculated through the linear function, the parameter lookup table is determined, and the activation function calculation is realized.

The invention belongs to the technical field of CT images, and particularly relates to a double-domain CT metal artifact removal method based on deep learning. Comprising the following steps: 1) acquiring CT images of different scanning parts, and processing the CT images to obtain simulated CT images containing metal artifacts; 2) correcting and simulating the CT image containing the metal artifact through a traditional nonlinear interpolation algorithm to obtain an image after nonlinear interpolation correction, and preprocessing the image; 3) constructing a metal artifact removal model based on a deep convolutional neural network, and training the metal artifact removal model; 4) reasoning the preprocessed image by using the trained model to obtain a metal artifact-removed image, and counting the training precision; and 5) carrying out post-processing on the metal artifact-removed image, and carrying out reverse normalization to an original CT magnitude. The method not only performs metal artifact correction in the projection domain, but also performs metal artifact correction in the image domain, so that the metal artifact removal precision is ensured, and the model convergence can be accelerated.

Embodiments of the present disclosure relate to a nonlinear interpolation unit. The interpolation logic described herein provides a good approximation of bicubic interpolation without performing all of the computations normally required for bicubic interpolation, which is generally smoother than bilinear interpolation. This allows the approximation of the smooth bicubic interpolation to be performed on devices with limited processing resources, such as mobile devices. At each position in a set of predetermined interpolation positions within an array of data points, a set of predetermined weights represents a bicubic interpolation that can be applied to the data points. For a plurality of predetermined interpolation positions surrounding a sample position, the corresponding set of predetermined weights and the data points are used to determine a plurality of surrounding interpolated values that represent the results of performing bicubic interpolation at the surrounding predetermined interpolation positions. Then, a linear interpolation is performed on the surrounding interpolated values to determine an interpolated value at the sample position.

This invention relates to a high-precision digital predistortion phase interpolator with a constant input load, comprising: a multi-phase clock generation circuit, an input buffer circuit, two pre-filter circuits, a phase interpolation core circuit, a CML2CMOS circuit, and a control code decoder circuit. The two pre-filter circuits convert two full-swing square wave signals into two sinusoidal clocks. The phase interpolation core circuit performs nonlinear interpolation on the two sinusoidal clocks with a constant input load based on two weighted control codes, outputting a small-swing signal. This phase interpolator significantly improves interpolation accuracy. Digital predistortion of the interpolation weights can be achieved by adjusting the control code decoder, further enhancing the accuracy of the phase interpolator and enabling high-speed, high-precision output clocks. It is suitable for clock data recovery circuits and phase-locked loops in high-speed SerDes systems.

The application discloses a channel state information compensation method, a communication device and a computer program product. It relates to the technical field of network communication, and the method comprises the following steps: acquiring a target moving speed of relative motion between a first device and a second device, and a plurality of channel estimation results of a channel between the first device and the second device; determining a target nonlinear interpolation coefficient corresponding to the target moving speed based on a corresponding relationship between the moving speed and the nonlinear interpolation coefficient; and compensating channel state information of the channel based on the plurality of channel estimation results and the target nonlinear interpolation coefficient to obtain a channel state information compensation result. The application solves the technical problem that in the related art, the channel considered when performing channel compensation is mostly a channel in a static or slow moving environment, and in the case that the terminal moving speed is relatively high, the channel has strong time variation, the channel reciprocity is poor, and the channel compensation accuracy is low.

The invention discloses a 5G-based underground coal mine unmanned inspection control method, and relates to the technical field of underground coal mine unmanned inspection, and the method comprises the following steps: constructing a multi-dimensional electromagnetic propagation sensing frame on the moving track of an underground coal mine inspection robot, collecting the instantaneous signal intensity, phase deviation and polarization direction, and generating a time sequence electromagnetic feature sequence; a space energy density mapping mechanism is operated based on a time sequence electromagnetic characteristic sequence, a signal distribution state is converted into a space-time continuous energy field, and sequence abnormal points are corrected through nonlinear interpolation to form an energy distribution pattern. According to the method, dynamic description of underground signal propagation and refined display of the risk situation are achieved through multi-dimensional electromagnetic sensing and energy mapping, an inverse phase wave train is generated by introducing a time domain inversion mirror algorithm to actively compensate extremely uneven signals, a propagation channel is uniform and stable, and operation continuity and safety of an inspection robot are ensured.

The present application relates to the vertical shaft cage detection technical field, specifically, it relates to a kind of industrial intelligent vision detection and diagnosis system for vertical shaft cage, acquisition unit synchronously obtains cage flange surface image and cage car pitch angle, coal dust shadow quantization unit uses bimodal adaptive threshold segmentation combined with region growing algorithm to extract coal dust area, introduce space weight factor to calculate average gray value and coverage area proportion, dynamic coupling calibration unit relies on three-dimensional look-up table and nonlinear interpolation of multi-working condition calibration, match the optimal deviation compensation coefficient of the combination of jitter and coal dust, generate dynamic gray scale segmentation threshold, defect diagnosis decision unit is through zoned adaptive segmentation, multiscale morphological filtering denoising, combined with gradient amplitude and curvature double-constraint subpixel edge detection accurately calculate wear depth, according to real-time depth and historical trend output grading control instruction, eliminate coupling interference, improve cage wear detection accuracy, provide reliable basis for safe operation.

The invention provides a gastrointestinal three-dimensional dynamic imaging analysis method based on conductivity distribution modeling, and the method comprises the steps: applying at least three composite electric signals with different frequencies to a gastrointestinal region, collecting voltage response signals, and constructing a dynamic conductivity parameter matrix based on frequency domain feature separation. And then, spatial discretization processing is carried out by using a nonlinear interpolation algorithm, a three-dimensional conductivity distribution map is generated, dynamic modulation and demodulation processing is carried out on the map, and a frequency band weight coefficient is adjusted. And mapping the modulated data to a dynamic imaging coordinate system through an inverse projection algorithm to generate a three-dimensional dynamic imaging result. According to the space coordinates and the change rate of the conductivity gradient mutation area, the position and the activity state of the gastrointestinal abnormal area are judged, and a self-adaptive feedback mechanism is adopted to update boundary condition parameters of the initial conductivity distribution model. The accuracy and efficiency of gastrointestinal disease diagnosis can be improved.

The application provides a wide-range high-altitude task environment aircraft high-dimensional temperature prediction method, comprising the following steps: step one: temperature-time first prediction based on regression analysis; step two: temperature-time secondary prediction based on nonlinear fitting and error random representation; step three: temperature-height prediction based on empirical formula and dynamic interpolation; step four: temperature-longitude and latitude prediction based on dynamic interpolation; and step five: high-dimensional temperature fusion prediction. The application comprehensively utilizes mathematical statistics methods such as regression analysis, nonlinear fitting, linear (nonlinear) interpolation and empirical formula to construct a temperature factor prediction model related to four-dimensional variables including time, height, longitude and latitude, so that a reasonable temperature long-term prediction value can be output when any four-dimensional coordinate in the given aircraft task process, thereby providing guidance for the task process.

The application discloses a curve determination method for point-by-point integration of a centrifugal pump and belongs to the technical field of centrifugal pumps. u The curve determination method for point-by-point integration of the centrifugal pump is a parameterized modeling method based on Euler equation and momentum conservation, the inlet and outlet boundaries of the v r curve are calculated according to design requirements, a smooth transition curve is constructed by combining flow channel geometry and angle of attack constraints, the curve is constructed by dynamic correlation of streamline arc length through nonlinear interpolation, and the relationship between the streamline arc length and the curve is established; the centrifugal pump blade profile can be designed according to the given design conditions and the streamline arc length s through the parameterized method.

The application provides a high-precision hole plate alignment method for a multilayer circuit board, and belongs to the technical field of printed circuit board manufacturing. The method collects measured coordinates of each key layer pad array in the laminated multilayer board through an X-ray perspective system, takes each layer pad array as a distributed deformation sensor, and pairs with the design nominal coordinates to obtain a deviation vector; affine transformation separation and radial basis function nonlinear interpolation are respectively performed on each key layer to reconstruct an anisotropic deformation field covering the whole board area; the range of layers passed through is determined according to the drilling type, and weighted fusion is performed on the deformation field of each layer to calculate the optimal compensation vector of each drilling hole, and a compensated drilling program file is generated. The method upgrades the traditional target alignment and two-step global scaling method to a three-stage compensation process of distributed sampling, nonlinear deformation field reconstruction and hole-by-hole coordinate correction, and controls the hole plate offset to within 25 microns without changing the existing drilling equipment hardware.

The invention relates to a GA-BPNN-based spectral calibration method for a prism dispersion type spectral imager, which is mainly used for solving the problems that in spectral calibration, Gaussian fitting is carried out by using a least square method, so that the requirement on data distribution is high, and a local optimal solution is easily caused; and the technical problems of Runge phenomenon, sensitivity to data noise, local inflexibility, boundary error accumulation and the like when polynomial fitting is used for nonlinear interpolation are solved. According to the GA-BPNN-based spectral calibration method for the prism dispersion type spectral imager, Gaussian fitting is carried out through a genetic algorithm GA to obtain a spectral response function of a selected channel, and nonlinear interpolation is carried out through construction of a BP neural network to determine a spectral response function of a whole channel. The method can adaptively find a global optimal solution of Gaussian fitting and nonlinear interpolation, has strong nonlinear mapping capability and fault-tolerant capability, effectively solves the problem of dispersion nonlinearity of the prism, and improves the quantitative measurement capability of the prism dispersion type spectral imager.

The invention relates to the related technical field of pointing correction, in particular to a method for correcting pointing of a two-dimensional phased array by using a difference-sum ratio. The method comprises the steps that pointing information can be rapidly extracted by adopting an electric scanning mode in combination with a phased array antenna and difference phase information, and pointing precision correction can be rapidly completed by adopting a polynomial interpolation method for accurately extracting the pointing information. The method has the technical effect that the corresponding relation between the beam pointing angle and the beam control angle is accurately positioned by using the difference-sum ratio. The data acquisition efficiency is improved by utilizing an electric scanning mode; and carrying out nonlinear interpolation encryption and correction on the grid by utilizing MATLAB (matrix laboratory).

The application discloses a design method of a three-order precision nonlinear interpolation smoothing factor based on a five-point template, belongs to the field of computational fluid dynamics, and realizes the unification of high precision, non-oscillation and multi-scale adaptability in complex flow simulation through nonlinear weighting, multi-order combination and smoothing factor design, and becomes an indispensable numerical tool for aerospace. The smoothing factor is a core mechanism for realizing shock capturing and high resolution of the high-precision format, and the weight distribution is dynamically adjusted through quantitative solution smoothing, so that a balance is achieved between shock capturing, high precision and low dissipation. Therefore, the application proposes a Z-type three-order precision nonlinear interpolation smoothing factor based on a five-point template, which meets the optimal precision sufficient condition of a multi-order candidate template nonlinear weighting framework. The method uses the method of expanding the template, and uses a consistent high-order smoothing factor method to design a new consistent high-order smoothing factor which meets the optimal interpolation precision.

The invention relates to the technical field of geotechnical mechanics and geophysical exploration, and discloses a three-dimensional electrical resistance tomography method for geotechnical materials and a data acquisition device. The method aims to solve the technical problems that in the prior art, a general linear interpolation method is prone to generating a block mosaic effect and is extremely unfavorable for image analysis, and a nonlinear interpolation method is poor in accuracy, possibly causes calculation difficulty and cannot achieve real-time rapid imaging. The method comprises the following steps: (1) collecting electrode data; (2) screening and summarizing data; (3) carrying out hierarchical plane inversion imaging; and (4) depth three-dimensional interpolation. The spatial electrode arrangement mode is optimized, a numerical inversion imaging system capable of being used for engineering scale measurement is formed, real-time, rapid and accurate monitoring of geotechnical engineering projects is achieved, the detection speed is high, the imaging efficiency is high, the calculated amount is relatively small, the requirement for non-intrusive engineering monitoring is met, and engineering condition damage cannot be caused.

The application discloses a transient power quality disturbance detection method based on digital frequency conversion and nonlinear interpolation, which comprises the following steps: collecting three-phase voltage and three-phase current of a power grid under different disturbances; sequentially performing waveform interception, spectrum segmentation, spectrum shift, sinusoidal interpolation, resampling, empirical mode decomposition and correlation coefficient calculation on single-phase voltage or current; extracting required components and feature values according to correlation coefficients of the components; training an OS-ELM classifier; and detecting transient power quality disturbance and disturbance categories through the OS-ELM classifier.

The application discloses an infrared focal plane array POCC parameter solidification loading method and particularly relates to the field of infrared image sensors; a basic pixel output matrix is collected at a calibration temperature, and initial POCC parameters are calculated; pixel response data is collected at multiple temperature points, and a temperature drift response model is constructed; dynamic prediction parameters are generated based on the initial parameters and the temperature drift model, and a temperature-adaptive POCC solidification matrix Pfused is formed through weighted nonlinear interpolation fusion; automatic loading is performed in the system power-on initialization stage; in the running process, if it is detected that the temperature change exceeds a set threshold, a corresponding subset in the Pfused is called to perform offset correction; the application has the characteristics of temperature drift sensing, autonomous updating and solidification availability, improves the offset correction accuracy and running stability of an infrared imaging system in a complex thermal environment, and is suitable for application scenarios of multiple types of infrared focal plane arrays such as satellite-borne, airborne and vehicle-borne infrared focal plane arrays.

The embodiment of the invention provides an image reconstruction model training method and device, an image reconstruction method and device and electronic equipment, and relates to the technical field of image processing. The image reconstruction model training method comprises the following steps: acquiring a first sample RAW image, a second sample RAW image and a real HDR image acquired from the same scene at the same visual angle; performing image splicing on the first sample RAW image and the second sample RAW image to obtain a predicted spliced image; using an encoder to extract prediction down-sampling features of the prediction spliced image; performing up-sampling processing on the predicted down-sampling feature by using a decoder to obtain a predicted RAW image; performing nonlinear interpolation on the predicted RAW image to obtain a predicted RGB image; and based on the difference between the real HDR image and the predicted RGB image, adjusting model parameters of the image reconstruction model until the image reconstruction model converges, and obtaining a trained image reconstruction model. Based on the RAW image with more complete and more accurate image information, the HDR image with higher quality can be obtained.

The invention provides a multi-terminal sharing high concurrent adaptive synchronization method and system and a storage medium, and the synchronization method comprises the steps: carrying out the multi-dimensional priority calculation, and calculating the synchronization priority coefficient of a first user and each other user in real time; performing hierarchical LOD transmission; performing client prediction and smoothing, wherein a receiving end adopts client dead reckoning prediction and nonlinear interpolation smooth transition; a P2P relay technology based on geofences is introduced into P2P relay based on geofences and comprises geofence division, temporary edge node selection, Mesh network construction and data relay, so that the problem of state synchronization in high concurrency and weak network environments is solved, smooth and stable XR interaction experience can still be obtained in a complex outdoor network environment, and the method is suitable for being applied to the field of communication. The transmission quantity of useless data is reduced, and the communication problem of a signal blind area is also solved. In order to realize the above effects, the invention provides a multi-device and multi-user shared high concurrent adaptive synchronization method and system oriented to a super-large space augmented reality environment, and a storage medium.

This disclosure provides a video generation method and apparatus. The method includes: extracting first image features from a first image; obtaining multiple intermediate image features through nonlinear interpolation based on the first image features and second image features, wherein the second image features are image features of the second image; and reconstructing an image using an image generation model based on the first image features, the second image features, and the multiple intermediate image features to generate a target video, wherein the target video is used to show the process of transitioning from the first image to the second image. Thus, by using nonlinear interpolation, the quality of the intermediate image features is improved, thereby improving the image quality of the intermediate frames of the target video, i.e., improving the video quality of the target video.

The embodiment of the invention provides a method for constructing an earthquake vulnerability curve. The method comprises the following steps: acquiring earthquake intensity data and actual earthquake loss data of a target area; determining a fragility curve corresponding to the target building according to the seismic oscillation intensity data of the target building in the target area; determining an initial vulnerability curve corresponding to the target building according to the fragility curve and a preset nonlinear interpolation parameter; the preset nonlinear interpolation parameters comprise a gradient control parameter, a reference offset parameter and a demarcation point parameter; taking the obtained actual seismic loss data as an optimization target, and adopting a preset optimization algorithm with boundary constraint to iteratively adjust parameters of the initial vulnerability curve to obtain a target vulnerability curve corresponding to the target building; the parameters of the initial vulnerability curve comprise regression parameters of the initial vulnerability curve and preset nonlinear interpolation parameters.

The application relates to the technical field of industrial process control, and specifically discloses a regulating valve flow opening degree optimization method and system based on deep learning, which comprises the following steps: collecting multi-source operation data of a split-range control system, analyzing control signals, valve actions and flow responses, judging whether a switching point gain mutation exists, if the mutation exists, predicting future flow demand and estimating a current measurement noise level by using a pre-trained time sequence convolution network and an attention mechanism hybrid model, dynamically calculating an optimal switching point and a dynamic overlap area width according to the flow prediction sequence and the noise level, and adaptively adjusting the dynamic overlap area width according to a flow change rate; and comparing flow characteristics of large and small valves, adopting a linear distribution method if the characteristics are the same, or designing a nonlinear interpolation method to distribute valve opening degrees in the dynamic overlap area, so that the gain mutation problem near the switching point in the split-range control is solved, the adaptive optimization of the flow opening degree is realized, and the system stability and control precision are improved.

The invention discloses a battery performance detection method and system introducing implicit features, and relates to the technical field of electrochemical detection.The method comprises the steps that timestamp remapping is conducted on a voltage sequence, temperature field distribution and an acoustic emission signal through a space-time convolution attention module, and a unified time axis is established; a nonlinear interpolation model is adopted; constructing a dual-channel feature extraction network, inputting the dynamic alignment data into an improved 3D-ResNet18 network, and outputting dominant feature vectors; extracting a potential degradation mode from the target domain battery through domain adversarial transfer learning, and generating an implicit feature vector; generating a final feature by adopting a feature enhancement fusion layer; the chemical characteristics comprise the negative pulse amplitude of the differential capacitor at the set voltage, and the acoustic characteristics comprise the energy accumulation value of the set frequency band; and calculating a lithium precipitation risk index. The safety boundary is redefined through a double-channel framework, and remarkable economic benefits are created through service life prolonging and accident avoidance.

The application relates to a fundus photograph-based method, system and device for calculating the diameters of retinal blood vessels. The method comprises the following steps: acquiring a fundus photograph to be processed; segmenting the fundus photograph to be processed to obtain a fundus photograph with a tree structure and single-branch blood vessels; calculating the normal vector of each branch blood vessel segment pixel by pixel; and through nonlinear interpolation and smoothing processing, the average, maximum and minimum diameters of each branch blood vessel are obtained. The method aims to extract a target region based on an automatic segmentation algorithm, then traverse each branch blood vessel according to the tree structure to calculate the normal vector, and then calculate the parameters of each branch blood vessel through interpolation operation, so that the important clinical significance and potential application value of the method in the prediction and analysis of cardiovascular diseases and systemic diseases are discovered.

The invention relates to an indicating value automatic identification method and system device of a pointer type multimeter, the device realizes automatic fixing and gear switching of the multimeter through an automatic platform comprising an electric clamping jaw, a guide rail and a rotating mechanism, acquires images of the multimeter through an industrial camera, and quickly positions a dial plate area by using a YOLOv8 target detection network; and carrying out pixel-level segmentation on the dial plate and the gear plate by adopting a U2-Net-based deep learning model to generate a pointer, a scale and a gear mask. And geometric calculation is carried out on the binary mask pixels to determine the circle center of the dial plate. Through polar coordinate expansion, conversion from fan-shaped distribution to linear one-dimensional data is realized. And finally, outputting an indicating value in combination with scale peak quadratic fitting optimization and a piecewise nonlinear interpolation technology. And meanwhile, gear information acquisition and indicating value integration are completed by combining the circle center of the gear disc and the mass center of the pointer which are determined by K-Means clustering. The method effectively adapts to different types of multimeters and complex backgrounds, and efficiency and accuracy are remarkably improved.