708 results about "Spectral bands" patented technology

A system and method are disclosed for generating hyperspectral images from multi-modal sensor data including RGB, LiDAR, thermal, and near-infrared inputs. Training data includes hyperspectral images and corresponding multi-modal measurements. Spectral band grouping is performed based on correlation coefficients. A multi-modal decomposition network with cross-modal attention mechanisms generate reconstructed hyperspectral images by fusing complementary sensor information. A fine-tuning network creates reconstructed RGB images. A comprehensive quality assurance system analyzes spectral consistency, cross-modal coherence, and fusion artifacts to generate quality metrics. Missing data compensation strategies handle corrupted sensor inputs using information from other modalities. The system includes temporal integration for video sequences and multi-resolution processing for different sensor resolutions. Quality metrics guide network weight adjustments to improve reconstruction accuracy while maintaining robustness to sensor failures and environmental variations.

The invention relates to the technical field of agricultural product quality detection, in particular to an agricultural product quality detection method and system. The method comprises the following steps: carrying out spectrum scanning on an apple through a near infrared spectrometer, and removing baseline drift to obtain an accurate reflection correction spectrum; spectral bands of sugar, acidity and cellulose are identified based on a reflection spectrum, peak shape variation structural strength analysis is carried out, furthermore, quantitative processing of background interference bands is carried out on peak shape variation structural strength data, and unnecessary interference signals are eliminated; and finally, according to the quantitative data of the background interference wave band, constructing a quality detection and identification model, and identifying the content distribution of sugar, acidity and cellulose of the apples through the model to obtain accurate component content distribution data. The process realizes efficient and accurate quality detection, and ensures the reliability and intelligence of agricultural product quality control. The agricultural product quality detection technology is optimized, so that the agricultural product quality detection technology is more perfect.

The invention relates to the technical field of underwater fish school monitoring, and discloses an underwater fish school monitoring statistical system based on image fusion. According to the system, underwater video streams and sonar reflection intensity data of different spectral bands are acquired through an underwater multi-source image acquisition module, and time-space synchronous multi-modal image data streams are generated through timestamp alignment; a fish school contour reconstruction module is used for segmenting a fish school contour boundary and fusing visible light texture and sonar geometric features to generate an underwater three-dimensional fish school distribution set; the dynamic track mapping module tracks the mass center displacement, calculates the movement rate and the direction deviation angle, and correlates the water area depth to generate a dynamic track topological graph; the behavior anomaly analysis module extracts environment data based on the track mutation node, and detects aggregation density change and direction dispersion to mark an anomaly feature cluster; and the population statistics output module integrates the data, performs classified statistics on population distribution, a quantity threshold value and a migration path overlap ratio, and finally generates a fish school quantity distribution statistics thermodynamic map.

The invention provides a water body monitoring method and equipment based on satellite remote sensing data and a storage medium, and relates to the technical field of water body monitoring, the method comprises the following steps: firstly, acquiring remote sensing image data and water quality measured data (including water quality parameters and geographic positions) of a target area at least covering two spectral bands; performing radiation correction on the remote sensing image data, and extracting target image data corresponding to the water body by using the image segmentation model; matching the water quality parameters with the target remote sensing image data according to the geographic position information, and constructing a training set of time-space alignment; using a multilayer model fusion strategy to train a preset water quality inversion model based on the training set to obtain a target model; and finally, inputting an image to be detected to the target model to obtain water quality parameters, and mapping the water quality parameters to a geographic space to generate a water quality two-dimensional distribution diagram. According to the invention, the water body monitoring precision and monitoring details based on satellite remote sensing data can be effectively improved.

The invention relates to the technical field of image processing, and discloses a printed matter detection method and device. The method comprises the following steps: acquiring multispectral image data of a printed matter to be detected in different spectral bands; performing spectral feature extraction and multi-level fusion on the multispectral image data to obtain a comprehensive spectral feature map; acquiring sample spectral feature maps of a plurality of sample printed matters to perform frequent feature mode extraction, generating a spectral fingerprint database, and performing region division and matching degree search to obtain a reference spectral fingerprint map; carrying out multi-scale decomposition on the comprehensive spectrum feature map and the reference spectrum fingerprint map to obtain a plurality of scale layers, and carrying out local spectrum similarity calculation and adaptive threshold segmentation to obtain a binary mask map of a defect region; performing spectral feature clustering analysis to obtain defect categories; and performing defect root cause analysis and process adjustment on the printing process parameters to generate a process adjustment scheme, thereby improving the printing quality detection accuracy of the printed matter.

The invention provides a geometric correction method and device for a satellite remote sensing image and a storage medium, and relates to the technical field of image correction, and the method comprises the steps: obtaining remote sensing image data of at least two different spectral bands, synchronously recording satellite attitude parameters and orbit position parameters, and carrying out the verification processing; based on the satellite attitude parameters and the orbit position, geometric coordinate conversion is carried out on each remote sensing image data, and target remote sensing image data is generated; dividing each piece of target remote sensing image data into a plurality of continuous grid units, and constructing a local geometric transformation model at the vertex of each grid unit; and resampling the original pixels in each grid unit according to the local geometric transformation model, and outputting a plurality of pieces of target remote sensing image data subjected to geometric correction. According to the invention, the geometric correction precision of the high-resolution satellite remote sensing image can be improved.

Techniques for determining a presence of an object, especially an object such as animal or debris, in a path of a vehicle, are discussed herein. For example, sensors of various modalities, which may include multispectral sensors, may capture data representing an environment the vehicle is traversing. In examples, one or more trained machine learned (ML) models, operating on a vehicle computing system, may detect and / or classify objects in the environment, based on input data of one or more modalities or spectral bands. The ML models may be pre-trained using training data including real sensor data, synthetic data, and / or augmented data, along with auto-generated annotations. In some examples, hyperspectral data may be used to identify materials associated with detected objects. A confidence score associated with the detection of the object may also be computed. The vehicle may be controlled based on detection of the object and its classification.

A system can use a plurality of co-located telescopes to generate enhanced telescopic imagery of space objects. The system may be configured to receive telescopic imagery data of a plurality of space objects obtained from the plurality of co-located telescopes. The system can receive at least two imaging criteria. The available imaging criteria can include a target signal sensitivity comprising a minimum signal to noise ratio (SNR), a target number of spectral bands comprising a minimum number of spectral bands, a spectral range associated with one or more of the spectral bands, a location of the plurality of co-located telescopes, a target data cadence comprising a minimum number of frames per minute, a target number of space objects to be tracked, a target minimum spatial resolution, among others. The system can transmit instructions to the plurality of telescopes and generate enhanced telescopic imagery.

The invention relates to the technical field of remote sensing surveying and mapping, in particular to a terrain recognition and three-dimensional reconstruction method based on a remote sensing surveying and mapping image, which comprises the following steps of: recognizing and selecting a key spectral band based on a remote sensing image, adjusting the configuration of a spectral filter, extracting key terrain data according to configuration data, and analyzing terrain change by utilizing terrain feature data. And recording related wave bands and positions, forming a comparison table, and combining natural colors and surface texture analysis to generate a three-dimensional model of the terrain. According to the method, the key spectrum sections are screened by adopting the preferable waveband index, the specificity and efficiency of data acquisition are improved, the spectral response section is accurately adjusted, it is ensured that a spectral region with remarkable earth surface reflection characteristics is focused on, and the practical value of spectral data is improved; the spectral configuration data optimizes the response performance of the sensor so that the data collected from the terrain is more sensitive and accurate.

The invention relates to the technical field of solar panel maintenance, in particular to a photovoltaic power station solar cell panel intelligent flushing system, which comprises an image analysis module, an intelligent control module, an image processing module, a data processing module, a data processing module, a data processing module, a data processing module and a data processing module, and is characterized in that the image analysis module is used for monitoring smudginess data of the surface of a solar cell panel and transmitting the smudginess data to the intelligent control module in real time; the multispectral camera collects surface images of a current cell panel and a next cell panel, the surface images comprise cell panel surface images collected under different spectral bands or angles, and smudginess distribution and smudginess degree of the next cell panel are obtained; a cleaning strategy is generated according to the cleaning requirement through a preset intelligent algorithm, a control instruction is output and sent to the intelligent control module through the communication unit, and the intelligent control module adjusts the working state of the flushing device. According to the system, the precise recognition capability of the system on the dirty area is improved, it is ensured that water flow can effectively cover the part needing to be cleaned in the cleaning process, and waste of the non-dirty area is avoided.

The invention provides a wheat early-stage stripe rust detection method and system, and the method comprises the steps: inputting time dimension embedded multi-modal data, pre-training a spectrum reconstruction and time sequence comparison learning model, and mining universal feature modes of wheat stripe rust in different development stages and different spectral bands; and transferring the pre-trained weight to a multi-modal encoder of a multi-modal multi-scale target detection model, extracting interactive fusion features of different scales, inputting the interactive fusion features into a multi-scale decoder, comprehensively utilizing global context information and local detail information, and meanwhile, fully utilizing the features of different scales through a cross attention mechanism so as to obtain a multi-scale multi-scale target detection model. Outputting a stripe rust detection result; finely adjusting parameters of the multi-scale decoder by using data with labels in the early-stage stripe rust data to obtain a trained multi-modal multi-scale target detection model; and performing early stripe rust detection on wheat to be detected by using the trained multi-modal multi-scale target detection model. The kit can be used for detecting the early wheat stripe rust.

The invention relates to an all-fiber sub-hundred femtosecond laser based on an SESAM. The all-fiber sub-hundred femtosecond laser comprises the SESAM, a wavelength division multiplexer, a pumping source, a pumping protector, a gain fiber, a chirp fiber bragg grating, an isolator, a reflector, a compression unit and an autocorrelator. According to the invention, the chirp fiber bragg grating with the dispersion amount of 0.115 ps / nm is adopted to compensate positive dispersion in the cavity, so that mode-locked pulse with the spectral width of 23nm is realized. A transmission-type diffraction grating pair is adopted, the pulse width is compressed to 88fs, and the maximum spectral width 27nm is obtained under the dispersion of 0.00087 ps2 in a near-zero cavity. The condition that the optimal spectral bandwidth of the optimal seed source pulse based on the optical fiber chirped pulse amplification laser should be larger than 15 nm is met, and the compression pulse distortion of the amplifier is reduced. By adopting the linear cavity, the SESAM is adhered to the surface of the single-mode optical fiber patch cord for starting mode locking, so that an all-optical-fiber structure of the system is realized, and the single-mode optical fiber patch cord is suitable for industrial integrated application.

The invention discloses a traditional Chinese medicine pesticide residue detection method and system based on artificial intelligence and a medium, particularly relates to the technical field of image processing and intelligent detection, and is used for solving the problem that the capability of distinguishing complex surface textures and pesticide residue areas of traditional Chinese medicines is insufficient. The method comprises the following steps: extracting local gradient direction distribution features of a high-resolution image, and combining gradient magnitude clustering processing to generate a gradient feature map of a texture edge contour; carrying out multi-scale fusion on the gradient features and the color channel data, and constructing a multi-channel feature map representing textures, colors and spatial distribution; analyzing quantitative feature relevance based on covariance of a high-order color moment and a directional entropy, and dividing normal textures and abnormal residual regions through a mahalanobis distance classifier; and performing spectral band matching degree evaluation on the abnormal region, fusing geometric morphological characteristics and spectral scores, inputting the fused geometric morphological characteristics and spectral scores into a pre-trained classification model, and generating a pesticide residue detection result by using a nonlinear correlation decision, thereby realizing accurate distinguishing between the natural texture on the surface of the traditional Chinese medicinal material and the pesticide residue.

The invention relates to the technical field of quality detection, in particular to an agricultural product quality detection method and system. The method comprises the following steps: carrying out multi-plane spectrum scanning on an apple through a near-infrared spectrometer to obtain a reflection spectrum and carrying out baseline correction; then, analyzing the mildewing degree of the apple pulp through cellulose spectral band identification and peak intensity attenuation index deduction; thirdly, identifying a preposed abnormal spectrum risk factor based on the mildew degree data, generating a mildew preposed abnormal spectrum risk factor, and constructing a quality detection model through sequential evolution traceability association of the abnormal factor and the mildew degree; finally, the detection model is sent to a terminal to execute agricultural product quality detection. According to the invention, the quality detection technology is optimized, so that the quality detection technology is more accurate.

The invention relates to the technical field of image processing, can be applied to business scenes of agricultural insurance, medical health and the like, and discloses a data processing method, device and equipment based on a spectrum association weight and a medium, and the method comprises the steps: obtaining original spectrum data, carrying out the preprocessing of the original spectrum data, and generating the preprocessed spectrum data; analyzing the relevance between spectral bands based on the preprocessed spectral data, and generating a spectral relevance weight; determining activation state distribution of a plurality of filters according to the spectrum correlation weight, selecting an activation filter based on the distribution to perform feature extraction on the data, and generating screening features; and inputting the screened features into an intelligent decision model to generate a target decision result. According to the method, through a dynamic filter activation strategy, the calculation complexity can be effectively reduced, and meanwhile high target detection precision is kept. Through deep analysis and optimization of the correlation between spectral bands, the robustness in small target detection and dynamic change environments is improved.

The invention discloses a spectral polarization imaging three-dimensional reconstruction method and device based on a pixel array, and a medium, and relates to the technical field of image processing, and the method mainly comprises the steps: collecting original images, which are fixedly obtained by pixels, of corresponding spectral bands and polarization angles; extracting sub-image information under each spectral band and polarization angle according to pixel positions, and stacking to generate a four-dimensional data volume; constructing a polarization geometric model according to the four-dimensional data body, and constructing an actual expression of the linear polarization degree in combination with a Fresnel reflection theory; constructing a redundant dispersion constraint equation and a theoretical expression of the relation between the diffuse reflection spectrum polarization degree and the incident angle and the imaging target refractive index; substituting the redundant dispersion constraint equation into a combined residual function of the theoretical expression and the actual expression, and performing parameter inversion on the combined residual function; and carrying out three-dimensional shape recovery under imaging target surface normal reconstruction based on an inversion result. According to the method, time delay and space registration errors caused by time-sharing acquisition are eliminated.

The invention relates to the technical field of tailless thread sleeve detection, and provides a multispectral imaging-based tailless thread sleeve deformation detection system and method, and the method comprises the steps: S1, collecting reflection maps of tailless thread sleeve standard parts and deformation flaw parts of different models under different spectral bands through a multispectral imaging device, and constructing a data set according to model classification; s2, carrying out preprocessing and feature extraction on the obtained reflection spectrum; s3, constructing a recognition model based on a convolutional neural network (CNN); s4, verifying the accuracy of the constructed recognition model by using a verification set; s5, updating the trained recognition model to a calculation unit; s6, adding newly added detection data into the training set regularly, retraining the model and updating the model to the calculation unit; according to the invention, various deformation types of the tailless thread sleeve can be rapidly identified, the calculation amount is small, and the detection efficiency is improved.

The invention is an apparatus for amplifying cooling through interaction with electromagnetic radiation for optical of objects and / or object surfaces with essentially three layers, which are a bottom layer that is comprised of a single or multi layered material configured to emit IR radiation; a middle layer that is comprised of a single or multi layered material configured to respond in anti-Stokes fluorescence upon absorption of electromagnetic radiation; and a top layer that is comprised of a single or multi layered material configured to reflect selected spectral band(s) and / or amplify selected spectral band(s) of the electromagnetic radiation transmittable to the middle layer.

Techniques for determining a presence of an object, especially an object such as animal or debris, in a path of a vehicle, are discussed herein. For example, sensors of various modalities, which may include multispectral sensors, may capture data representing an environment the vehicle is traversing. In examples, one or more trained machine learned (ML) models, operating on a vehicle computing system, may detect and / or classify objects in the environment, based on input data of one or more modalities or spectral bands. The ML models may be pre-trained using training data including real sensor data, synthetic data, and / or augmented data, along with auto-generated annotations. In some examples, hyperspectral data may be used to identify materials associated with detected objects. A confidence score associated with the detection of the object may also be computed. The vehicle may be controlled based on detection of the object and its classification.

A method is provided for determining soil properties for an area of land, from soil spectrum data. In an embodiment, the method includes receiving soil spectrum data records from hyperspectral sensors that represent a mean soil spectrum of a specific geo-location of the area of land and removing interference signals from the spectrum data records to create soil spectral bands. The method also includes predicting a plurality of soil property datasets based on a partial least-square regression and the soil spectral bands and selecting specific soil property datasets from the plurality of soil property datasets to represent soil properties of the specific geo-location, wherein the specific soil property datasets include property data and spectral band data for spectral bands used to determine the property data. The specific soil property datasets may then be used to generate a crop prescription of recommended hybrid seeds or population densities for the specific geo-location.

The invention discloses a cultivated land state monitoring method based on cultivated land boundary identification, and relates to the technical field of image analysis, and the technical scheme comprises the steps: carrying out the preprocessing of a periodically obtained cultivated land real-time image, and carrying out the screening to obtain a gray-scale map corresponding to a set optimal spectral band; based on the grey-scale map and the spatial distribution characteristics and the time sequence difference characteristics of the seedlings, calculating the comprehensive probability that each local peak pixel point represents the existence of the seedlings in the grey-scale map of the current period; based on the comprehensive probability, judging whether each local peak pixel point represents a seedling or not; dividing a cultivated land boundary based on a judgment result, and distinguishing a cultivated area and a non-cultivated area; according to the method, the existence probability of the seedlings is judged by utilizing the spatial distribution characteristics and the time sequence difference characteristics of the seedlings, the accuracy of cultivated land boundary identification can be improved, and the cultivated land state is periodically updated and monitored in real time.

The invention provides a multifunctional passive cooling optical metamaterial fiber, yarn and fabric and a preparation method thereof. The passive cooling optical metamaterial fiber is composed of high-refractive-index inorganic micro-nano particles and / or a micro-nano functional material and a polymer substrate material. Wherein wavelength / sub-wavelength scale micro-nano particles are uniformly distributed in a substrate material as a random scattering medium to form a random optical metamaterial system, and reflection enhancement of the optical metamaterial fabric in a solar spectrum band is realized by utilizing a multiple Mie resonance collective effect. According to the passive cooling optical metamaterial fiber, a circular structure, a skin-core structure and the like are prepared through melt spinning, high-refractive-index inorganic micro-nano particles are added into a skin layer and used for sunlight band wide spectrum regulation and control, and a micro-nano functional material is added into a core layer, so that multiple functions are effectively achieved.

The invention discloses a tortoise-shell glue transparency detection method and system based on multispectrum, and the method comprises the steps: S1, obtaining a multispectral image sequence of a tortoise-shell glue sample, carrying out the wave band correction, constructing a multispectral image data cube, and carrying out the spatial registration; s2, generating a spectral angle mapping graph according to the registered multispectral image data cube; s3, extracting a tortoise-shell glue region through spectral angle mapping and a region growing algorithm, and constructing a spectral feature matrix; s4, based on the spectral feature matrix and a cross validation method, obtaining an optimal tortoise-shell glue transparency qualified product identification model, and performing qualification judgment on the transparency of the tortoise-shell glue sample to be detected; and S5, constructing a transparency quality evaluation model, and carrying out quality grade division on the transparency of the tortoise-shell glue. And carrying out quality grading on the transparency of the tortoise-shell glue. According to the method, objective and comprehensive evaluation of the transparency quality of the tortoise-shell glue can be realized by acquiring the image information of the tortoise-shell glue sample in a plurality of spectral bands and combining an intelligent analysis algorithm.

Implementations are described herein for obtaining a sequence of high-elevation images capturing a particular geographic area during a particular time period in a plurality of spectral bands; applying the sequence of high-elevation images as input to upstream machine learning model(s) to generate remote sensing embeddings indicating terrain feature(s) of the particular geographic area; applying agricultural data obtained from a local agricultural knowledge graph as input to additional upstream machine learning model(s) to generate agricultural knowledge embedding(s); inferring a natural-language description of a status of the particular geographic area based on generating: an aggregate representative embedding that semantically represents a plurality of agricultural conditions of the particular geographic area, and a natural-language description of one or more of the plurality of agricultural conditions of the particular area using a large language model; and causing a user device associated with an agricultural entity to present the natural-language description of the status.

The present invention provides a method for enhancing mobile phone imaging images driven by near-infrared, including: Step 1, an additional near-infrared spectral sensor is mounted on the mobile terminal to select the near-infrared spectral band; an RGB image is captured by the visible light camera of the mobile terminal, and a near-infrared image is obtained through the near-infrared spectral sensor; Step 2, a visible light and near-infrared joint decoupling module is established to map the near-infrared image and the RGB image to the same feature space, and then the light and dark components are output through a decoder; Step 3, a light and dark component embedding framework is established to embed the light and dark components into the image enhancement network in a plug-and-play manner to obtain the final enhanced result. The present invention can make full use of the additional near-infrared band information to provide a prior with clear physical information guidance for mobile phone imaging, and to a certain extent, can alleviate problems such as the dependence on data and poor interpretability of the end-to-end deep learning method.

The invention discloses a wine grape gray mold grade discrimination method and device based on unmanned aerial vehicle remote sensing and a medium, and relates to the technical field of unmanned aerial vehicle remote sensing. The method comprises the following steps: preprocessing unmanned aerial vehicle image data to obtain a multi-period hyperspectral image; extracting region-of-interest spectral data from the multi-period hyperspectral image based on the geographic position data; carrying out vegetation index feature extraction and physiological and biochemical feature extraction on the spectroscopic data of the region of interest; the vegetation index features and the physiological and biochemical features are combined with the original spectral band to be connected according to a time sequence to form a multi-temporal feature vector; correlation analysis is carried out on the multi-temporal feature vectors, and target feature vectors most relevant to gray mold are screened out; according to the method, the grape gray mold grade is taken as a target feature, a disease monitoring model based on a time sequence is constructed, training prediction is carried out by utilizing a target feature vector, and a grape gray mold monitoring model is obtained, so that efficient, accurate and high-adaptability wine grape gray mold grade judgment is realized.

The invention discloses a lithium battery laser welding detection method, and belongs to the technical field of nondestructive testing. The method comprises the following steps: acquiring a three-dimensional path point sequence of a to-be-detected welding seam; controlling the transient thermal excitation unit to apply transient thermal excitation along the path; a high-speed multispectral imaging unit is used for synchronously collecting the thermal radiation attenuation process of the weld joint under at least two different spectral bands, and a space-time spectrum data cube is formed; and finally, inputting the data cube into an artificial intelligence model for defect identification. The system correspondingly comprises a three-dimensional visual sensor, a transient thermal excitation unit, a high-speed multispectral imaging unit and a synchronous control and data processing unit. According to the method, different types of welding defects such as air holes and incomplete fusion can be effectively distinguished by analyzing the heat diffusion characteristic difference under different spectral bands, the interference of uneven surface emissivity is inhibited, and the online detection accuracy and reliability of the three-dimensional complex welding seam are improved.

The invention relates to the technical field of automatic control, in particular to a control system for a mosquito killing lamp and a control method thereof, and the control system comprises the following modules: a mosquito behavior analysis module, a spectral efficiency analysis module, a trap lamp parameter configuration module, a voltage parameter adjustment module and an automatic counting module. By analyzing the relationship between the mosquito activity number and various environmental conditions, identifying the mosquito activity mode, combining with real-time monitoring of the environmental conditions, optimizing the start-stop parameters and spectral band selection of the mosquito killer lamp, improving the adaptability and trapping effect of the trap lamp, reducing the invalid working time, and improving the efficiency of the trap lamp. Real-time parameter adjustment of a high-voltage power grid, optimization of the working efficiency of the power grid and reduction of energy consumption are achieved by analyzing voltage parameters needed for killing various mosquitoes, the number of trapped mosquitoes is counted in real time by analyzing current changes, the trapping efficiency is evaluated, the working parameters of a lamp and the power grid are dynamically optimized according to a result, and the trapping effect is improved. And the working efficiency and the resource utilization efficiency are improved.

The invention relates to the technical field of remote sensing image recognition, in particular to a method for determining a moso bamboo forest snow disaster affected area based on a GEE platform. The method mainly solves the problems of low efficiency, high cost and difficulty in realizing large-range rapid evaluation caused by dependence on manual field investigation in the prior art. According to the technical scheme, the method comprises the following steps: firstly, obtaining a multi-temporal Sentinel-2 image before and after a snow disaster, and carrying out the preprocessing of the multi-temporal Sentinel-2 image; extracting spectral bands, vegetation indexes and texture features to form an initial feature set; then, a key feature variable combination is determined through statistical significance filtering and machine learning optimization; on this basis, establishing a logistic regression discrimination model and determining an optimal classification threshold; and finally, carrying out snow disaster state classification on the moso bamboo forest region by utilizing the trained model, and generating a disaster region spatial distribution diagram.

The invention relates to the technical field of image enhancement, in particular to a traditional Chinese medicinal material infrared microscopic image enhancement method, which comprises the following steps of: performing two-dimensional Fourier transform on each spectral band image based on input original traditional Chinese medicinal material infrared spectral microscopic image data to obtain frequency domain representation, identifying a target frequency peak value position generated by interference in the frequency domain representation, and obtaining a target frequency peak value; and generating a stripe-removed hyperspectral data cube. According to the method, frequency domain analysis is carried out on an original infrared spectrum microscopic image through two-dimensional Fourier transform, stripe noise generated by light path interference is recognized and eliminated, a stripe-removed hyperspectral data cube is generated, and the input data quality of subsequent component separation is improved. And performing nonlinear decomposition on the spectrum vector after stripe removal, constructing a full-pixel end member contribution weight set, organizing weight values into an original abundance image set in a two-dimensional matrix form by utilizing an end member spectrum index, and realizing spatial decoupling and independent expression of multi-chemical component distribution.