26 results about "Differential spectrum" patented technology

The invention relates to the technical field of intelligent detection, in particular to a feed mycotoxin detection method and system based on spectral analysis. The method comprises the following steps: acquiring a near-infrared diffuse reflection spectrum of a feed sample to be detected, constructing an original spectrum vector based on absorbance data of the near-infrared diffuse reflection spectrum, and determining a projection component of the original spectrum vector on a pre-constructed qualified matrix space; subtracting the projection component from the original spectrum vector to obtain a residual vector, and calculating a first-order difference spectrum of the residual vector based on a preset sensitivity weight; and performing local feature weighting on the first-order difference spectrum to determine a weighted feature value, determining a residual entropy value of the residual vector and a range value of the residual vector, and determining a detection result of the to-be-detected feed sample by using the weighted feature value, the residual entropy value and the range value. According to the technical scheme, a more accurate detection result of the mycotoxin of the feed can be obtained.

The invention discloses a satellite hyperspectral coastal saline-alkali soil salinity inversion method based on a machine learning model, and the method comprises the steps: S1, collecting an original multispectral image and ground actual measurement data, and obtaining standardized initial feature space data through the radiometric calibration correction preprocessing operation; s2, performing fractional order differential spectrum enhancement on the initial feature space data to obtain an enhanced high-dimensional feature set, and performing importance adaptive optimization on the enhanced high-dimensional feature set by adopting a Boruta feature selection algorithm to obtain an optimal high-dimensional feature subset; s3, based on the optimal high-dimensional feature subset, an initial inversion model is obtained by integrating Optuna hyper-parameter automatic optimization, and based on the initial inversion model, a salinity inversion model is obtained through feature contribution degree analysis optimization and embedding of an SHAP interpreter; and S4, performing salinity prediction on the global remote sensing image based on the salinity inversion model to realize sea saline-alkali land salinity inversion.

The application relates to a visual identification method and system of a liquid crystal angle color-changing material. Specifically, optical images of a material at a preset angle sequence are collected, Lab chroma and rotation-invariant uniform LBP texture features of each image are extracted, and a multi-dimensional initial feature sequence is formed. A double-branch identification model is constructed. One branch performs differential spectrum analysis on the feature sequence, and then extracts features through a multilayer perception machine. The other branch encodes contexts through a bidirectional GRU, and enhances key information by using angle-channel correlation attention, and then pools the key information into a context vector. The outputs of the two branches are concatenated, and are mapped into an encoding vector through a classifier. During training, a polarization loss function is used to drive the encoding vectors of the genuine product and the fake product to converge to two preset opposite polarization directions. During identification, the distances between the encoding vector of a material to be tested and the two polarization vectors are calculated, and the one with a shorter distance is determined as the genuine or fake product.

The present application relates to the field of on-line analysis of material chemical composition, and discloses an on-line monitoring process for the composition of a lead frame electroplating solution, comprising: calculating the concentration of the electroplating solution by conventional monitoring of the spectrum; calculating a spectrum entropy index based on the degree of deviation of the real-time spectrum from a pre-established statistical baseline model, and comparing the spectrum entropy index with a pre-set failure trigger threshold; starting a dynamic calibration step only when the spectrum entropy index exceeds the pre-set failure trigger threshold, the dynamic calibration step generating a model error vector by comparing the physically real differential spectrum with the model virtual differential spectrum through injection of a standard pulse, and correcting the chemometrics model using the model error vector, so that the present application realizes on-demand calibration, ensures real-time accuracy of the model, and reduces the consumption of the standard.

The present application relates to the technical field of nondestructive testing and intelligent sorting of agricultural products, in particular to a kind of soybean based on hyperspectral image nondestructive testing method, comprising: collecting the original hyperspectral image data of soybean sample, generate pre-processing hyperspectral data cube;Calculate spatial gradient field, and apply gradient partition threshold to construct morphological distribution model;Spectrum-morphology coupling operator is constructed, and the output self-referenced differential spectrum feature set;For long-wave band spectral response, execute anomaly detection, generate internal state determination result output detection classification data;According to error distribution characteristics, update the gradient partition threshold of morphological distribution model and the weight parameter of spectrum-morphology coupling operator, generate final detection result for downstream sorting equipment to call;The present application solves the problem that fixed model produces performance drift with sample change, ensures the stability of sorting precision.

The invention relates to the technical field of communication, discloses an electromagnetic spectrum abnormal signal monitoring method and system based on double acquisition and comparison, and aims to solve the problems of asynchronous double-channel sampling and high abnormal signal omission ratio and false alarm rate caused by clock drift and physical path delay in the prior art. According to the method, a hardware-level self-adaptive clock synchronization compensation circuit is constructed, clock path delay is quantized and offset in real time before analog-to-digital conversion, and a first sampling clock and a second sampling clock which are accurately synchronized are generated; a dual-channel ADC is used for synchronously collecting the same radio frequency signal, and abnormal signals are judged and identified through spectrum coherence analysis and differential spectrum threshold. The system comprises a main clock unit, a self-adaptive synchronous compensation circuit, a dual-channel ADC module and a data processing comparison module. The picosecond-level synchronization precision is achieved, external time service is not needed, the missing detection and false alarm rate is remarkably reduced, and the method is suitable for high-reliability frequency spectrum monitoring in the complex electromagnetic environment.

The invention relates to the technical field of agricultural product nondestructive testing and intelligent sorting, in particular to a green soy bean nondestructive testing method based on a hyperspectral image, which comprises the following steps: collecting original hyperspectral image data of a green soy bean sample, and generating a preprocessed hyperspectral data cube; calculating a space gradient field, and constructing a morphological distribution model by applying a gradient partition threshold value; constructing a spectrum-form coupling operator, and outputting a self-reference differential spectrum feature set; performing anomaly detection on the long-band spectral response, generating an internal state judgment result, and outputting detection grading data; updating a gradient partition threshold value of the morphological distribution model and a weight parameter of a spectrum-morphological coupling operator according to the error distribution characteristics, and generating a final detection result to be called by downstream sorting equipment; the problem that a fixed model generates performance drift along with sample change is solved, and the stability of sorting precision is guaranteed.

A method for diagnosing the noise level of a transformer comprises: a step of acquiring a frequency spectrum of noise from a stacked iron core 21 contained in a transformer, the stacked iron core 21 being formed by stacking electrical steel laminations; a step in which, for the calculation of the excitation vibration of the stacked iron core 21, the transverse moduli of elasticity are determined in two planes, each including a stacking direction of the stacked iron core 21; a step in which the excitation vibration for the stacked iron core 21 is calculated using a constitutive equation that includes an elasticity matrix comprising the determined transverse moduli of elasticity as elements, and a frequency spectrum of an excitation vibration of the stacked iron core 21 is acquired;a calculation stage of the difference between the excitation noise spectrum of the stacked iron core 21 and the excitation vibration frequency spectrum of the stacked iron core 21 as a differential spectrum; and a transformer noise level diagnostic stage based on the differential spectrum.

A method for measuring anisotropic film based on zero reflection difference differential signal relates to the field of film thickness detection, and comprises the following steps: measuring wide-band reflection difference spectrum of a sample by using a reflection difference spectrometer and a photoelastic modulation reflection difference spectrometer; determining an incident wavelength corresponding to a zero reflection difference signal: traversing all wavelength channels in the wide-band reflection difference spectrum, and determining a wavelength corresponding to a zero-crossing of the reflection difference signal; determining the thickness of the anisotropic film: determining the thickness of the anisotropic film according to a mapping relationship between the wavelength corresponding to the zero reflection difference signal and the film thickness. The application proposes using the position of the zero-crossing point of the reflection difference signal ΔR / R=0 to determine the thickness of the anisotropic film, so that the error introduced by the difference between the polarization angle of incident light and the crystal axis angle of the anisotropic film can be removed, and further, the time for determining the crystal axis can be saved.

The application provides a kind of optical sensing component differential spectral responsivity test system, belongs to the technical field of spectral responsivity test, comprising: modulation light source, for emitting modulation light;Bias light source, for emitting bias light, bias light and modulation light form superimposed light;Standard optical sensing module, obtains the standard value of superimposed light intensity;Measured optical sensing module, obtains the measured value of superimposed light intensity;Signal and control unit calculates the differential spectral responsivity of measured optical sensing module under superimposed light.The application also provides a kind of optical sensing component differential spectral responsivity test method, solves the problem that the test method under darkroom condition in prior art is difficult to test and calibrate the spectral responsivity of optical sensing component under actual application scene condition, realizes the spectral responsivity of measured optical sensing module in the case of background light, and evaluates the anti-interference ability of measured optical sensing module.

The invention discloses a vibration densification process reasoning and mapping system based on decision tree machine learning, and relates to the technical field of machine learning, and the system comprises a data collection unit which is used for carrying out the differential spectrum matching of a target object spectrum at a frequency band of 900nm to 1950nm; the double-domain residual machine learning unit is used for extracting abnormal characteristic parameters window by window by taking the zero pulse of the vibration exciter as a reference signal, judging item by item according to a decision tree model, outputting a continuous strip-shaped result through track constraint and morphological opening and closing operation, and generating a continuous quality map; and the output and feedback unit is used for outputting a continuous compact quality index and speed regulation, pass increasing, vibration switching and water replenishing prompts through a decision tree classifier according to the implicit under-compact probability grid and the water content suitability by taking the vertical acceleration clock of the roller as a unified reference, and writing the continuous compact quality index and the speed regulation, pass increasing, vibration switching and water replenishing prompts into a continuous quality map. According to the invention, the dual-suitable monitoring of the density and the water-containing state is realized, and the problems of unstable abnormal identification and difficult data alignment in the prior art are solved.

The invention relates to the technical field of electrochemical in-situ spectrum monitoring, in particular to a method for monitoring dynamic change of an electrocatalytic reaction intermediate product in real time by using SERS (Surface Enhanced Raman Scattering) spectrum.According to the method, an ultrathin catalytic layer is modified on a high-SERS activity nano substrate in situ, an integrated working electrode is constructed, and a monitoring interface, namely a real reaction interface, is ensured. Through a hardware synchronization system, when dynamic electrochemical excitation is applied, synchronous acquisition of SERS spectrums and electric signals of multiple time scales is implemented. The method comprises the following steps: analyzing fingerprint information, concentration dynamics and evolution sequence of different intermediate products from a spectrum sequence by using algorithms such as dynamic differential spectrum, multivariate curve resolution and two-dimensional correlation spectrum, and finally accurately associating with electrochemical response to construct a dynamic reaction mechanism diagram. According to the invention, real-time and high-sensitivity dynamic analysis on generation, consumption and path evolution of transient intermediate products at a real catalytic interface is realized, and a powerful tool at a molecular level is provided for research on an electro-catalysis mechanism.

Aspects of the technology provide a method / system of sensing a change in temperature and / or strain in an optical fibre comprising a substantially continuous or quasi-continuous fibre Bragg grating. A difference spectrum is determined between a reference reflected spectrum of the fibre Bragg grating and a monitored reflected spectrum. The difference spectrum is analysed to determine whether the optical fibre is experiencing a change of type a)—a temperature increase or an extension—or of type b)—a temperature decrease or a contraction. The technology may be applied to sensing a change in temperature and / or strain in a superconductor or detecting a risk of a quench in a superconductor.

Disclosed in the present invention are a spectral difference analysis method and device for gas-solid interface flashover. The method comprises: acquiring several pieces of gas-gas spectral data acquired by a gas-gas gap discharge experimental device; acquiring several pieces of gas-solid spectral data acquired by a gas-solid surface flashover experiment device; performing averaging operations on the several pieces of gas-gas spectral data and the several pieces of gas-solid spectral data to obtain background spectral data and gas-solid flashover spectral data, the gas-gas spectral data corresponding to the background spectral data, and the gas-solid spectral data corresponding to the gas-solid flashover spectral data; and, on the basis of the background spectral data and the gas-solid flashover spectral data, calculating a differential spectral signal, such that a user performs gas-solid interface flashover analysis on the basis of the differential spectral signal. The present invention analyzes the effect of a spectrum generated by a gas itself on a solid flashover spectrum during the experiment of the gas-solid surface flashover experimental device, thus rendering gas-solid interface flashover analysis more accurate.

The invention discloses a microwave power detection system based on a diamond NV color center, which belongs to the technical field of precision measurement and comprises a laser, the diamond NV color center, a photoelectric detector, a microwave generating device, a lock-in amplifier and a main control device. The laser generates laser with a set wavelength, so that the diamond NV color center generates fluorescence after being irradiated by the laser; the photoelectric detector performs photoelectric conversion on the fluorescence to generate a voltage signal; the microwave generating device outputs a microwave signal modulated by the modulation signal; the lock-in amplifier receives the voltage signal and the modulation signal, demodulates the voltage signal and the modulation signal, and outputs an ODMR differential spectrum signal; the main control device generates a differential spectrum line covering a formant area to be measured according to the ODMR differential spectrum signal, the differential spectrum line is fitted, a spectrum line broadening parameter is obtained, and the microwave power is obtained through calculation according to the spectrum line broadening parameter. According to the invention, the problem that the experiment construction and debugging difficulty is large due to the complex structure of the light path and the control circuit in the prior art can be solved.

The application provides a Raman spectrum cosmic ray peak identification method based on multi-feature fusion, which can accurately identify typical high and narrow cosmic ray peaks and narrow but low, high but wide and overlapping variant peaks. The identification steps comprise: 1) performing first-order differential processing on the original Raman spectrum to obtain a differential spectrum, and then performing threshold screening on the positive and negative differential vertexes in the differential spectrum to obtain a candidate peak set; 2) taking the positive and negative differential vertexes of the candidate peak as starting points, performing percentile threshold detection on the differential spectrum to preliminarily locate the peak boundary, and then performing percentile threshold detection on the original Raman spectrum to accurately locate the left and right boundary points of the peak; 3) calculating the left edge pixel width, right edge pixel width, boundary point pixel width, half-peak height pixel width and differential vertex pixel width of the candidate peak, then performing weighted fusion calculation on the pixel width to obtain a score value, and finally identifying the cosmic ray peak according to the score threshold.

A method is provided for multi-energy X-ray imaging of a field of view in order to display a contrast material, located in the field of view, having an atomic number of at least 65, (e.g., at least 73), and / or a k-edge of at least 60 keV. In the method, two image datasets having different X-ray spectra are recorded and are combined in order to ascertain display information of the contrast material, wherein the X-ray spectra are generated by using at least one spectrally acting filter, starting from a source spectrum of an X-ray source in such a way that the value of their differential spectrum has a maximum, which includes a range, beginning at the k-edge of the contrast material, of higher absorption by the contrast material.

The invention discloses a wide dynamic range sulfur dioxide concentration detection method and a wide dynamic range sulfur dioxide concentration detection device based on UV-DOAS combined with wave band self-competition spectrum reconstruction. The method comprises the following steps: acquiring SO2 data to be detected; determining an SO2 spectral line, and processing the spectral data of SO2 into a differential spectrum of SO2 to be detected based on the determined SO2 spectral line; an adjustable sliding window is constructed, and dynamic wave band selection is achieved through the sliding window; a mapping matrix is obtained through the mapping relation of the standard spectrum, and the differential spectrum of SO2 to be detected is converted from a wavelength domain to an intensity domain by using the mapping matrix; in the intensity domain, noise and interference signals in the spectrum are removed in a first-order linear fitting mode for fitting, and a reconstructed linear spectrum is obtained; the reconstructed linear spectrum is converted from the intensity domain to the wavelength domain through the inverse matrix of the mapping matrix, the differential spectrum of the to-be-detected SO2 after denoising is obtained, the concentration of the to-be-detected SO2 is calculated, dynamic waveband selection is achieved by means of the waveband self-competition technology, and therefore the detection dynamic range is effectively widened.

The present application relates to the ICP emission spectrum adaptive baseline calibration method and system of water body metal detection, aims at improving the adaptive calibration ability of baseline drift in the continuous scanning period of inductively coupled plasma emission spectrometer. Through point-by-point differentiation and noise suppression of the original spectrum intensity sequence, the standardized differential spectrum sequence is generated and the Shannon entropy is divided into sub-windows to calculate the differential entropy time sequence reflecting the local chaos degree, the mean, standard deviation and slope change rate are extracted to form a three-dimensional entropy flow state vector, and the spectrum signal anomaly detection and baseline fitting parameter freezing are realized by combining the incremental update isolated forest algorithm. Then, the limited step gradient descent fine tuning is carried out in the local window, the dynamic calibration parameter group is generated and the baseline compensation is completed. The method effectively suppresses the baseline drift and parameter oscillation caused by environmental and sample changes, improves the stability of the spectrum calibration data, and provides a reliable data basis for subsequent metal characteristic peak detection.

The invention relates to a soil organic carbon functional group differential spectrum quantitative analysis method based on mineral peak correction. The method comprises the following steps: dividing a to-be-detected soil sample meeting a preset standard into two parts, firing one of the two parts to form an original soil sample and an ash content sample, collecting a functional group characteristic signal from the sample through a Fourier transform infrared spectrometer, generating an original soil spectrum and an ash content spectrum, and calculating the content of the ash content in the original soil spectrum and the ash content spectrum according to the original soil spectrum and the ash content spectrum. The method comprises the following steps: carrying out spectrum pretreatment on an original soil spectrum and an ash spectrum, carrying out spectrum subtraction treatment after pretreatment to obtain a differential spectrum, carrying out peak area integration and proportion calculation on the differential spectrum, and establishing a reliable correlation between the peak area of the differential spectrum and the absolute content of organic carbon so as to directly output the absolute content of a target carbon functional group. By adopting the method, the defect that the traditional spectral analysis method can only be qualitative or semi-quantitative can be effectively overcome, the anti-interference capability and the standardization degree are improved, and the method has better repeatability.

The present application relates to the technical field of drug component detection, and in particular to a method for detecting concentrations of multiple drug components in blood based on Raman spectroscopy. The method obtains differential spectrum data corresponding to Raman spectrum data of serum samples of a patient before and after administration of a target drug. Based on the differences between differential signals corresponding to each endogenous substance in adjacent frame differential spectra in the differential spectrum data, the differential signals corresponding to the target drug in each frame differential spectrum in the differential spectrum data are corrected to obtain corrected differential signals corresponding to the target drug in each frame differential spectrum. Based on the fusion weight corresponding to each endogenous substance, the corrected differential signals corresponding to all target drugs are weighted and fused to obtain a fusion differential signal of the target drug. Finally, based on the fusion differential signal of the target drug, the drug concentration of the target drug in each Raman spectrum is determined. The present application effectively improves the detection accuracy of drug concentration.

The invention discloses a pipeline tiny leakage detection method based on a dynamic differential spectrum and metric learning. The method comprises the following steps: firstly, acquiring an acoustic signal of a pipeline and generating initial time-frequency representation; determining a main frequency in the initial time-frequency representation in a preset high-frequency band, constructing a dynamic gain function based on the main frequency to generate enhanced time-frequency representation, and calculating a difference value between the enhanced time-frequency representation and the initial time-frequency representation to obtain a dynamic difference spectrum; inputting the dynamic difference spectrum into a deep neural network model to extract a high-dimensional feature vector; and finally, based on a measurement distance between the feature vector and a pre-stored sample in a reference feature database, a k-nearest neighbor algorithm is adopted to determine whether leakage exists. According to the method, weak features are enhanced in a specific frequency band, the differential spectrum is utilized to highlight the weak features, and reliable judgment is performed in combination with metric learning, so that the detection accuracy and robustness of the tiny leakage of the pipeline are remarkably improved.

The invention belongs to the technical field of spectral analysis and chemical quantitative detection, and particularly relates to a reverse solvent spectral stripping quantitative analysis method, which comprises the following steps of: acquiring a Raman spectrum by selecting reference solutions with different concentrations, performing spectral subtraction by taking a high-concentration reference spectrum as a reference spectrum and a low-concentration spectrum as a to-be-processed spectrum, and eliminating a solvent background, so as to obtain a Raman spectrum; the method comprises the following steps: obtaining a difference spectrum only containing solute characteristic peaks, carrying out curve fitting on the difference spectrum, extracting a solute Raman characteristic peak intensity value, establishing a linear relation model of concentration and characteristic peak intensity, taking a spectrum of a solution to be detected as a spectrum to be processed, and subtracting the spectrum from a high-concentration reference spectrum to obtain a reverse difference spectrum; and calculating the solute concentration in the solution to be detected according to the established model. The problems of poor detection precision, high detection limit and the like of the low-concentration solute in the prior art are solved, thorough deduction and signal enhancement of a solvent background can be realized under the condition of not depending on pure solvent reference, and high-precision quantitative analysis of the low-concentration solute is realized.

This invention belongs to the field of degraded oil regeneration technology, and specifically discloses a photoluminescence decolorization and regeneration device for transformer dismantling oil. The device includes a main body, a transparent processing container inside, a catalytic adsorption composite material at the bottom of the processing container, a sunshade outside the processing container, a lid on the top of the main body, and an ultraviolet lamp at the bottom of the lid. The ultraviolet lamp is inserted into the processing container and connected to a power source via wires and a plug. This invention has the advantages of simple structure and high-quality regeneration of dismantled oil. Furthermore, this invention discloses a GC-MS differential spectral detection method for coloring substances in transformer dismantling oil.

This invention belongs to the field of optical thin film measurement technology, specifically a film thickness measurement method and system based on differential feature-assisted deep learning. It involves constructing a polarization angle-resolved hyperspectral database based on an optical theoretical model; determining the initial film thickness value by coarsely matching the sample's polarization angle-resolved hyperspectral data with the database, and defining upper and lower bounds for the film thickness based on the initial value and thickness tolerance; constructing differential feature spectra by performing differential processing on the sample's polarization angle-resolved hyperspectral data with the corresponding polarization angle-resolved hyperspectral data at the upper and lower bounds, separating redundant information to highlight sensitive feature spectra caused by subtle film thickness variations; inputting the differential spectra into a pre-trained neural network, outputting the film thickness deviation from the upper and lower bounds, and calculating the final film thickness. This invention combines the global positioning capability of a physical model with the local fine-tuning capability of deep learning, achieving high-precision and rapid online measurement, and solving the problem of slow inversion speed and reliance on initial values ​​in ellipsometric measurement in complex multilayer film systems.

The invention relates to the technical field of maturity detection, and discloses a ramie fiber maturity detection method and system, and the method comprises the steps: obtaining the spectral data of a ramie fiber sample to be detected, and calculating the spectral torsion moment; based on the spectral torsion moment, through a nonlinear mapping function, determining a fractional differential order; processing the spectral data to obtain a differential spectral curve; fourier transform is carried out to obtain a frequency domain signal; determining an accumulated energy percentile threshold value, and dividing the frequency domain signal into a low-frequency band and a high-frequency band by taking a frequency when an accumulated integral value of the energy of the frequency domain signal from low frequency to high frequency reaches the accumulated energy percentile threshold value as a division frequency; respectively calculating energy integrals of a low-frequency band and a high-frequency band, and calculating a high-frequency and low-frequency energy ratio; and determining the maturity of the ramie fiber sample through a support vector regression model. Noise can be suppressed while effective signals are enhanced, and the detection precision of the ramie fiber maturity is improved.