15 results about "Nonlinear scaling" patented technology

The invention discloses a multi-modal communication signal intelligent identification method based on a neural network, and relates to the technical field of multi-modal signal identification. Comprising the following steps: acquiring original signals acquired by a sound sensor and a camera; inputting the processed original signal into a cross-modal auto-encoder to obtain a purification feature vector: processing the purification feature vector to obtain a quality scoring scalar; performing time sequence quality evaluation processing on the original signal to obtain a time sequence quality scoring sequence; fusing the time sequence quality scoring sequence and the quality scoring scalar to obtain a joint quality coefficient; based on the joint quality coefficient, adjusting a basic weight value of the original signal through weight nonlinear scaling processing, and generating a weight vector; processing the purified feature vector based on the weight vector to obtain a fused feature vector; and processing the fused feature vector to obtain a probability value representing the equipment failure rate. And through dynamic weight adaptive adjustment based on quality scoring, the high-quality signal occupies a larger proportion in the fusion process.

The invention discloses a mucky soil foundation intelligent monitoring method based on multi-source data fusion, and particularly relates to the technical field of foundation monitoring and early warning. The method comprises the following steps: constructing an original soil body state data set of a mucky soil foundation, and carrying out anomaly elimination and time calibration based on a rheological constitutive model to generate a foundation state characteristic matrix; performing dynamic causal association tracking analysis on the feature matrix in a sliding window mode to generate a water-soil coupling drift parameter set; extracting a stress historical dependency coefficient set by using a nonlinear scale analysis method; and according to the water-soil coupling drift parameter set and the stress historical dependency coefficient set, the stability risk level of the mucky soil foundation is judged, and corresponding early warning information is selected from a preset graded early warning instruction database and output, so that the accuracy and adaptability of anomaly recognition and risk early warning are improved, and the accuracy and the adaptability of abnormal recognition and risk early warning are improved. The method is suitable for long-period stability monitoring and intelligent risk prevention and control of high-soft foundations or silt stratums.

This invention provides a quantum-derived Newton-Raphson optimal fractional-order spectrogram generation method, comprising the following steps: Step 1: Acquire the original audio signal and construct a fractional-order spectrogram based on fractional Fourier transform (FRFT); Step 2: Perform nonlinear scaling compression on the fractional-order spectrogram using a Mel filter bank to generate a fractional-order Mel spectrogram; Step 3: Construct an adaptive optimization framework with information entropy minimization as the objective function to measure the information fidelity between the spectrogram and the original signal; Step 4: Use the quantum-derived Newton-Raphson optimization algorithm (QNRBO) to globally optimize the fractional-order order, frame length, and frame shift hyperparameters to generate the optimal fractional-order spectrogram; Step 5: Input the optimal fractional-order spectrogram into a downstream speech recognition model. This technical solution aims to systematically solve core problems such as insufficient traditional time-frequency representation capabilities, rigid hyperparameter configuration, limited optimization algorithm performance, and feature-task disconnect.

The invention relates to the technical field of computer vision, and discloses a method and system for adjusting a video picture display proportion of a splicing window, and the method comprises the steps: recognizing the original resolution and content features of a to-be-spliced video stream; establishing a feature matrix for receiving the to-be-spliced video stream, and establishing a layout analysis network of the to-be-spliced video stream, so as to determine an initial layout of the to-be-spliced video stream through the feature matrix; calculating an adaptive proportion interval of the to-be-spliced video stream; constructing a Grad-CAM thermodynamic diagram of the to-be-spliced video stream, marking a high response region of the Grad-CAM thermodynamic diagram, and performing nonlinear scaling on the to-be-spliced video stream according to the high response region and the adaptive proportion interval to obtain a scaled to-be-spliced video stream; and calculating an attention index for zooming the to-be-stitched video stream to generate a proportion adjustment instruction and a transition animation parameter for zooming the to-be-stitched video stream, and executing picture display proportion adjustment for zooming the to-be-stitched video stream based on the proportion adjustment instruction and the transition animation parameter. The display effect of the video picture can be improved.

The invention provides a bistatic synthetic aperture radar rapid imaging method, electronic equipment and a storage medium, and the method comprises the steps: receiving and obtaining an echo signal based on a transmitting signal transmitted by a transmitting station in a bunching mode; determining a dechirp parameter according to the echo signal, and performing dechirp on the echo signal according to the dechirp parameter to obtain a dechirp echo signal; performing interpolation transformation on the echo signal after dechirp to obtain an echo signal after interpolation; and azimuth nonlinear scaling is carried out on the echo signal after interpolation transformation, a nonlinear scaling parameter is obtained through calculation, and a final SAR image is obtained through azimuth Fourier transformation. Through dechirp and interpolation change, the imaging quality and the imaging effect are greatly improved.

Presented herein are systems and methods for providing improved computer-assisted display and analysis of nuclear medicine images. In particular, the systems and methods described herein provide for interactive adjustment of improved intensity window settings for displaying nuclear medicine images. The interactive intensity window selection tool described herein utilizes a nonlinear scaling function that maps user adjustments to the position of a display indicator widget on a scale to an intensity window threshold. The scaling function of the described form gradually amplifies user adjustments at the upper end of the scale, but remains linear at the lower end. The intensity window setting tool presented herein allows the user to adjust the intensity threshold over the full range of intensities encountered in an image, up to a maximum value, while still maintaining fidelity within an important range that includes lower intensities.

This invention discloses a method and system for constructing an equivalent model of GaNHEMT based on a nonlinear scaling rule. This method addresses the insufficient scaling accuracy of traditional linear scaling techniques by introducing an exponential term to modify the GaNHEMT scaling rule, thus constructing a large-signal model of GaNHEMT based on the nonlinear scaling rule. The impact of the nonlinear scaling rule on the S-parameters of GaNHEMT devices is analyzed. The accuracy differences between the nonlinear and linear scaling rules are compared, ultimately finding that the nonlinear scaling rule significantly improves the fitting accuracy of the S-parameters compared to the linear scaling rule. Without increasing model complexity, this method improves the modeling efficiency and accuracy of GaNHEMT devices. This method is applicable to modeling equivalent models of GaNHEMTs with different gate widths within the same batch.

The invention provides a light beam quality time-varying sequence efficient calculation method and device, and the method comprises the steps: carrying out the processing of an optical load, design parameters and a working time mode of an optical-mechanical system according to a data set standard format, and creating a to-be-predicted optical-mechanical system data set; processing a to-be-predicted optical-mechanical system data set through the linear scaling mapping function and the nonlinear scaling mapping function to obtain a predicted data set; the prediction data set is input into a trained long short-term memory recurrent neural network based on the attention mechanism, a light beam quality time-varying sequence conforming to a working time mode is output, and the long short-term memory recurrent neural network based on the attention mechanism is an encoder and decoder architecture. The invention provides a method and a device for efficiently calculating a light beam quality time-varying sequence, which are used for solving the problem that the accurate transmission light beam quality time-varying sequence of a designed optical-mechanical system under a specified optical load is difficult to calculate within finite time in the prior art.

The invention discloses a low-repetition-frequency fiber laser active coherent phase locking method based on a laser waveform self-reference optimization algorithm, and the method comprises the steps: firstly, carrying out the time domain nonlinear zooming processing of a coherent combined photoelectric signal through extracting the light intensity fluctuation characteristics of a full-period single-path pulse laser; the pulse intrinsic intensity fluctuation is effectively eliminated, the phase noise modulation characteristics are completely reserved, and a pseudo-continuous signal representing the real phase deviation is generated; a phase mismatch control signal is obtained based on bilateral symmetry phase disturbance operation, two laser phases are corrected through closed-loop feedback, and coherent combination output efficiency optimization is achieved. The method solves the problems that a traditional active phase locking method cannot effectively distinguish phase noise and pulse fluctuation under the condition of low repetition frequency, and phase noise information is lost due to the fact that a filtering algorithm is directly applied, and the threshold preset requirement is avoided through a waveform self-reference mechanism. The adaptability to different waveforms and repetition frequency parameters in coherent combination is improved, and high-precision phase locking control is achieved under the low repetition frequency condition (smaller than or equal to 5 kHz).

The present invention relates to a computer-implemented method and system (500) for providing a match metric for quantifying spectral similarity between a target coating and at least one sample coating, the system comprising a computing device (510) and a computer program product comprising computer executable code stored on a computer-readable storage medium (560) functionally coupled to the computing device (510) and causing the computing device (510) to perform a computational process when executed, the computational process comprising the steps of: receiving reflectance values ​​for a target coating and a sample coating for a plurality of wavelength values; normalizing each of the reflectance values ​​of the target coating and the sample coating by using a nonlinear scaling function; generating a normalized reflectance curve for the target coating and the sample coating for each wavelength value; generating a derivative value of the normalized reflectance curve of the target coating with respect to wavelength and a derivative value of the normalized reflectance curve of the sample coating with respect to wavelength for the plurality of wavelength values; generating a difference between the derivative value of the target coating and the derivative value of the sample coating for each wavelength value of the plurality of wavelength values; and generating a match metric for similarity between the normalized reflectance curves of the target coating and the sample coating based on at least the differences across all of the plurality of wavelength values.

The application discloses a diesel engine speed self-disturbance control method based on nonlinear scaling, which comprises the following steps: step 1, constructing a diesel engine speed dynamic model; the diesel engine speed dynamic model comprises total disturbance f(x, t); step 2, constructing a nonlinear scaling function g(z) to scale and reconstruct a tracking error e x ; step 3, based on the reconstructed variable x m of step 2, constructing an extended state observer of the lifting order, and feeding back an estimated value of the total disturbance f(x, t) to the diesel engine speed dynamic model of step 1 as the total disturbance f(x, t); step 4, constructing a low-gain control law, outputting u(t), and feeding back to the diesel engine speed dynamic model of step 1 to form a closed-loop system. The application can effectively inhibit high-frequency noise disturbance and reduce control input fluctuation.

The application discloses a kind of different scale model impact environment equivalent effect evaluation method and system, comprising: the similarity law of underwater explosion is derived by dimensional analysis, the relationship of shock wave pressure, energy impulse, energy density and charge mass and blast center distance is established.Based on Hopkinson similarity law, a scaled experimental model is designed, and nonlinear scaling optimization is carried out.The application has the advantages that: by dimensional analysis and similarity law, the relationship between the model and the actual environment is accurately established.Using Hopkinson similarity law and nonlinear scaling, the reliability of the model response is ensured.The technical obstacles of data integration and sharing of different scale models are effectively solved.Provide scientific and reliable technical support for engineering design and environmental assessment.

This application discloses a drought-flood transition prediction method, system, and storage medium that integrates artificial intelligence and physical mechanisms. The method includes: collecting multi-source data within a selected drought-flood transition study area; optimizing key factors affecting terrestrial water storage at each grid point within the study area; constructing and calibrating a temporal convolutional network deep learning model; using a long series of key factor sets to drive the deep learning model and reconstruct a long series of terrestrial water storage inversion datasets; extracting atmospheric circulation characteristics during drought-flood transition events and drought-flood transition events to construct a series of atmospheric circulation-drought-flood transition intensity pairings; deriving a nonlinear scaling relationship between drought-flood transition intensity and atmospheric circulation characteristics; and predicting the intensity and risk of drought-flood transitions at each grid point in the future. This invention provides a new perspective for drought-flood transition prediction, has important scientific significance for disaster prevention and mitigation, and provides a theoretical and technical foundation for assessing the environmental and hazard effects of Earth system evolution.

The application discloses a low-field susceptibility weighted imaging method, system, medium and product, relates to the field of susceptibility weighted imaging, and comprises the following steps: acquiring K-space original data obtained by scanning a target object, and performing image reconstruction on the K-space original data; obtaining multi-echo amplitude data and multi-echo phase data according to the obtained reconstructed image; unwrapping the multi-echo phase data, and merging the obtained multi-echo unwrapped phase data into three-dimensional phase data; merging the multi-echo amplitude data into three-dimensional amplitude data, and performing nonlinear scaling on the three-dimensional amplitude data; performing a high-pass filtering operation on the three-dimensional phase data based on complex domain phase separation, performing nonlinear phase mask processing on the obtained high-frequency phase data, multiplying the obtained phase mask and the scaled three-dimensional amplitude data, and performing minimum density projection on the obtained original susceptibility weighted image. The application can effectively improve the contrast of the SWI image under the low field.

The invention discloses a low-field magnetic sensitive weighted imaging method and system, a medium and a product, and relates to the field of magnetic sensitive weighted imaging, and the method comprises the steps: obtaining K-space original data obtained by scanning a target object, carrying out the image reconstruction of the K-space original data, and obtaining multi-echo amplitude data and multi-echo phase data according to an obtained reconstructed image; unwrapping the multi-echo phase data, and merging the obtained multi-echo unwrapped phase data into three-dimensional phase data; combining the multi-echo amplitude data into three-dimensional amplitude data, and carrying out nonlinear scaling on the three-dimensional amplitude data; and performing high-pass filtering operation based on complex field phase separation on the three-dimensional phase data, performing nonlinear phase mask processing on the obtained high-frequency phase data, multiplying the obtained phase mask by the zoomed three-dimensional amplitude data, and performing minimum density projection on the obtained original magnetic sensitive weighted image. According to the invention, the SWI image contrast in a low field can be effectively improved.