101 results about "Spectral distortion" patented technology

The invention discloses a panchromatic sharpening method based on multi-resolution panchromatic feature guidance, which comprises the following steps: firstly, designing a double-branch feature decomposition module for respectively performing feature decomposition on a panchromatic image and a multi-spectral image to extract low-frequency basic features and high-frequency detail features, and enabling the two branches to share the same shallow feature extraction module; correlation constraint is applied to the extracted basic features; then, aiming at the problem of information redundancy caused by direct serial connection and stacking of the extracted features, performing interactive fusion on the features by using an interactive feature recombination module, and enhancing a long-distance dependency relationship and a global context relationship among different modal features; and finally, an INN-based image reconstruction module is provided, and efficient fusion of the basic features and the detail features is realized through multi-level transformation and mapping. According to the method, cross-modal features can be effectively extracted and recombined and fused, spectrum distortion is remarkably reduced, and the spectrum distortion phenomenon caused by improper feature fusion is effectively avoided.

The invention relates to the field of image processing, and provides a snapshot compression imaging reconstruction method and system based on double low-rank regularization, which is characterized in that a double low-rank regularization model is designed to carry out two-dimensional processing on three-dimensional hyperspectral image data, so that the quality of original data is ensured; the problem that the reconstruction effect is poor due to data loss or errors is avoided, a low-rank group sparse prior algorithm and a deep low-rank decomposition algorithm are combined in the dual low-rank regularization model, and a dual low-rank regularization framework is formed. Priori information of global spectral correlation and local spatial smoothness is respectively extracted through a low-rank group sparse prior algorithm, texture detail features of a reconstruction result are enhanced, space factors and spectral factors are respectively extracted through a depth low-rank decomposition algorithm, deep information is quickly and effectively extracted, and the reconstruction efficiency is improved. The problems of texture detail loss and spectrum distortion of a reconstruction result are further avoided, and the reconstruction accuracy and authenticity of snapshot compression imaging are improved.

The invention discloses a barn pest monitoring and early warning method, system and device based on big data and a medium, particularly relates to the technical field of pest monitoring, and is used for solving the problem that the distribution trend of newly appearing pests in a barn cannot be accurately monitored in the prior art. The method comprises the following steps: dividing a granary into a plurality of monitoring areas, arranging an independent spectrum sensor network, carrying out real-time monitoring based on a spectrum analysis technology, carrying out contrastive analysis with a granary pest library, rapidly identifying new pests, and after the new pests are monitored, generating a distribution trend chart in real time through a spectrum imaging technology, according to the method, a trend chart is obtained, whether early warning needs to be sent out is judged according to the trend chart, and the data quality can be mastered more comprehensively by evaluating the spectral distortion hidden danger and randomness of spectral data in a high-risk monitoring area, so that the accuracy evaluation of the distribution trend chart of newly-emerged pests is improved, and the reliability and accuracy of a granary pest monitoring and early warning system are remarkably improved.

Disclosed is an audio codec that represents an immersive signal by a two-channel stereo signal that is a stereo rendering of the immersive signal and directional parameters. The directional parameters may be based on a perceptual model describing the direction of virtual speaker pairs to recreate the perceived location of dominant sounds. Audio processing at the decoder may be performed on the stereo signal in the frequency domain for multiple channel pairs using time-frequency tiles. Spatial localization of the audio signals may use a panning approach by applying weightings to the time-frequency tiles of the stereo signal for each output channel pair. The weightings for the time-frequency tiles may be derived based on the directional parameters, an analysis of the stereo signal, and the output channel layout. The weightings may be used to adaptively process the time-frequency tiles using a de-correlator to reduce or minimize spectral distortions from spatial rendering.

The invention relates to the technical field of component detection, in particular to a dairy product component real-time detection system based on near infrared spectroscopy, which is provided with a feature acquisition module, a preprocessing module, a feature pre-analysis module, a feature recognition module, a feature marking module and a feature judgment module. Whether a near infrared spectrum confidence coefficient abnormal risk exists or not is judged through a feature pre-analysis module, local surface images and near infrared spectrums of all recognition light paths are obtained through a feature recognition module, an incidence relation is established, feature recognition light paths are marked through a feature marking module, light path synchronism analysis is carried out through a feature judgment module, and the accuracy of light path synchronism analysis is improved. And judging whether abnormal early warning is sent out or not. According to the method, the abnormal risk of the confidence coefficient of the near infrared spectrum is identified, spectrum distortion is identified and eliminated according to the actual condition of the dairy product to be detected, and the reliability and accuracy of real-time detection of dairy product components are improved.

The invention relates to the technical field of data synchronization, and particularly discloses a multichannel instrument data synchronization method and system. The method comprises the following steps: acquiring time sequence data and timestamps of each channel of a multi-channel instrument, calculating a data sampling period and a clock drift rate according to the time sequence data and the timestamps so as to quantify time synchronization deviation, extracting a transient energy ratio and a spectrum distortion degree of a time sequence signal, measuring signal stability, performing normalization processing, and combining dynamic drift risk indexes of the time sequence signal, the spectrum distortion degree and the signal stability so as to obtain a time sequence signal. The low-risk data form a synchronization group, the high-risk data are isolated and corrected and then merged into the synchronization group to form synchronization data, and through dynamic risk assessment and grading processing, the influence of single-channel abnormity on overall synchronization is avoided, the clock and signal stability is considered, and the synchronization reliability is improved.

Disclosed herein are scientific instrument support systems, related methods, computing devices and computer-readable media. A method of mitigating distortion of an optical emission spectrum obtained from an optical emission spectrometer is provided. The method may comprise a step of obtaining a spectrum recorded with the spectrometer and a respective one or more condition parameters indicative of an operating condition at a time of recording the spectrum. The method may further comprise a step of providing a model configured to output, in response to the one or more condition parameters, one or more transform parameters of a transformation to be applied to the obtained spectrum. A transformation may be applied in accordance with the obtained one or more transform parameters to the obtained spectrum to mitigate distortion of the spectrum due to a discrepancy between the operating condition and a baseline operating condition.

The invention discloses a hyperspectral and multispectral image fusion method based on adaptive multi-scale features. According to the method, firstly, a low-spatial-resolution high-spectral-resolution hyperspectral image and a high-spatial-resolution low-spectral-resolution multispectral image serve as input of a network and are subjected to up-sampling through a sampling module on channel interaction to achieve efficient feature expansion, and then the hyperspectral image and the multispectral image are fused into a super-multispectral image with the same size as a reference image; and a multi-scale adaptive spatial reconstruction module captures diversified features and reconstructs spatial information, optimized feature maps enter a spectrum self-attention enhancement module to perform spectrum attention enhancement, correlation between the feature maps is calculated to generate an attention map, and spectrum information is reconstructed. In the image fusion process, the problems of spatial detail loss, insufficient cross-domain generalization ability, spectral distortion caused by low spectral resolution and the like are avoided.

Methods and apparatus (10) for Raman spectral analysis of a sample (12), such as a pharmaceutical dosage form, are disclosed. Delivery optics (16) are used to deliver probe light to a delivery region (13) on the sample, and collection optics (20) are used to collect, from a collection region (17) on the sample spaced from the delivery region, the probe light following scattering through the sample. Each of a plurality of target Raman spectral features are measured in the collected light, and a spectral distortion of the collected light arising during scattering through the sample is determined. A property of the sample is then quantified using the target Raman spectral features in combination with the determined spectral distortion, such that the quantified property is compensated for the spectral distortion.

The invention discloses a light source system capable of continuously adjusting an attenuation ratio, and belongs to the technical field of optical radiation measurement and photoelectric detection. In order to solve the problems of wavelength dependence, discontinuous adjustment, spectrum distortion and the like existing in an existing light intensity adjusting scheme, a combined structure of an integrating sphere and a diaphragm cylinder is adopted, and the device comprises a light source, the integrating sphere with the high-reflection inner wall, a front-end adjustable diaphragm, the diaphragm cylinder with the extinction inner wall and a rear-end adjustable diaphragm. Continuous attenuation of output light intensity in a predetermined dynamic range is realized by adjusting the apertures of the front and rear diaphragms, the lower limit of attenuation ratio meets the requirement, and the dynamic range can reach more than three orders of magnitude. The system uses the integrating sphere to form a stable Lambert light source, the attenuation process depends on geometric parameters, spectral distribution and spatial uniformity basically remain unchanged, and the non-uniformity of a light-emitting surface is not greater than 5%. The invention further provides a calibration method, high-precision attenuation prediction is achieved through detector measurement and view factor model fitting, and the calibration method is suitable for weak cursor calibration, detector linearity testing and other applications.

The invention relates to the technical field of textile fiber component analysis, and discloses a textile fiber component qualitative online analysis system and method.The system comprises a spectrum collection module, a data processing terminal and a component recognition center, and the component recognition center comprises a spectrum preprocessing unit, a feature matching unit and a component judgment unit. The spectrum acquisition module acquires multi-band spectrum data of a textile sample in real time, and the data processing terminal receives an instruction to perform spectrum data format conversion. The spectrum preprocessing unit determines an interference section by recording spectrum distortion and a baseline correction interval, and controls the processing terminal to generate a spectrum characteristic matrix to calibrate component distribution on the surface of a sample; a feature matching unit generates a component report according to matched fiber types; when the component determination unit detects an abnormal spectrum, re-measurement is controlled in an interference section. The system realizes online real-time qualitative analysis of the textile fiber components, enhances the anti-interference capability, and can be widely applied to the field of textile fiber component analysis.

The invention discloses a method for designing and evaluating a spectral range of a surveying and mapping load wave band, which comprises the following steps of: firstly, transmitting configuration parameters based on radiation under an imaging condition; the entrance pupil radiation characteristic result, the load imaging performance, the image texture characteristic and the spectrum distortion characteristic of the detection wave band designed based on different spectral range parameters at the surveying and mapping load entrance pupil after atmospheric radiation transmission are calculated; and secondly, according to the obtained entrance pupil radiation characteristic result, the imaging performance parameter calculation result, the image texture characteristic parameter calculation result and the spectral distortion characteristic calculation result, evaluating the comprehensive influence of different spectral range parameters on load imaging. According to the invention, the imaging quality and the image information amount of the surveying and mapping load can be evaluated by different spectrum broadening, and the optimal configuration of the spectrum range of the load detection wave band is realized.

The invention discloses online detection equipment for a medicine material, relates to the technical field of medicine material detection, and solves the problem of detection distortion caused by slight inclination of a medicine sample. The equipment comprises an equipment shell, a console and a detection bin, wherein a state detection assembly and an infrared detection assembly are arranged in the detection bin. An integrated circuit board in the console is integrated with a sensing acquisition module, a data processing module and a control module, the data processing module comprises a tiny inclination identification unit, a spectrum offset correction unit and a component data correction unit, and the real component content is calculated after the inclination angle is identified through pressure data and spectrum distortion is corrected. And the control module judges the qualified state of the sample in combination with a preset standard, controls the equipment to operate and sends a regulation instruction to an external production system. According to the equipment, the detection accuracy and the operation convenience are improved, the linkage of detection and production is realized, the online quality control requirement is met, and reliable support is provided for medicine production quality control.

The present invention relates to a method for processing and calibrating data of a circular variable filter spectroradiometer, including: receiving detector data sent in packets by the instrument end, splicing them to obtain complete detector data; correcting the collected original data to obtain detector data after drift correction; performing wavelength registration and then passing registration verification to obtain precisely registered spectral data; comparing the precisely registered spectral data with spectra at different temperature points recorded within the interval, and obtaining the responsivity function of the target to be measured through linear interpolation. The present invention proposes a drift correction method based on covariance cross-correlation and cyclic offset correction, and on the basis of CVF spectral calibration, uses the gas absorption method to re-register the wavelength after drift correction, solving the problem of spectral distortion caused by drift during the rotation of the CVF; the present invention can solve the non-linear response problem brought about by a large measurement temperature range and a wide working wavelength band, ensuring the accuracy of the measurement results of the instrument.

This invention discloses a multi-objective optimization-based sparse unmixing method for hyperspectral images. The method acquires the hyperspectral image to be unmixed and an initial endmember spectral library; constructs a structured sparse control operator based on dynamic temperature decay and adaptive gating; performs adaptive dynamic grouping of endmember topology based on gradient feedback and cosine similarity; constructs a multi-objective optimization model combining reconstruction error and group sparsity error for iterative optimization; and uses the maximum curvature method to obtain the final unmixing result from the Pareto front. Compared with existing technologies, by constructing an adaptive dynamic grouping mechanism for endmember topology, this invention fully utilizes gradient feedback and cosine similarity to capture spectral distortion differences, ensuring that the identification of similar endmembers is directional. Simultaneously, this invention combines a structured sparse control operator and uses a pass-through estimator to effectively filter false endmember features. Through the synergistic effect of these two methods, multi-objective sparse unmixing of hyperspectral images is ultimately achieved, minimizing reconstruction error and optimizing group sparsity error.

The invention discloses a high-precision calibration method of a hyperspectral laser radar for photon counting, which belongs to the technical field of laser radars and comprises a spectrum calibration stage and a radiation calibration stage. Spectral calibration is used for extracting spectral response parameters of each channel based on wavelength-by-wavelength scanning of a monochromator and synchronous measurement of a beam splitter; a detection link is strictly decoupled through radiometric calibration, the effective attenuation coefficient of a water body is extracted in situ through a slope method, full-link radiation correction is carried out in combination with an air whiteboard reference and window transmittance, and the target intrinsic spectral reflectivity without interference is inverted. The method can be used for data correction, target accurate classification and identification and high-fidelity quantitative detection of the hyperspectral laser radar for photon counting, and solves the problem of serious spectral distortion of an original signal caused by non-uniform light source energy of the hyperspectral laser radar and complex scattering absorption of a water body.

The present invention discloses a hyperspectral and multispectral image fusion method based on adaptive multi-scale features. First, a hyperspectral image with low spatial resolution and high spectral resolution and a multispectral image with high spatial resolution and low spectral resolution are used as the input of the network, and efficient feature expansion is achieved through upsampling by a channel interaction upsampling module. Then, they are fused into a super-multispectral image with the same size as the reference image. Next, diverse features are captured and spatial information is reconstructed via a multi-scale adaptive spatial reconstruction module. The optimized feature map enters a spectral self-attention enhancement module for spectral attention enhancement, calculates the correlation between feature maps to generate an attention map, and reconstructs spectral information. This avoids problems such as loss of spatial details, insufficient cross-domain generalization ability, and spectral distortion caused by low spectral resolution during the image fusion process.

The invention discloses a multi-level two-dimensional interactive remote sensing panchromatic sharpening method based on Transform, which comprises the following five steps: S1, inputting a low-resolution multispectral LRMS image and a panchromatic PAN image, and respectively projecting the two images to a feature space by using two different 1 * 1 convolution layers; s2, constructing two different branches by using a spectrum encoder-decoder MED and a space encoder-decoder PED to extract multi-level space details; s3, a space interaction module AIM and a spectrum interaction module EIM capture local and global potential interaction of each level respectively; s4, sending the multi-layer interactive representation to a reconstruction block to generate an HRMS image; and S5, designing a loss function to reduce the difference between an HRMS image generated by the model and Ground Truth and GT labels. According to the method, the complementary information of the PAN image and the MS image in the spatial dimension and the channel dimension is fully utilized, and the problems of spectral distortion, spatial blur and the like occurring in the fused image are relieved.

The invention relates to the technical field of vision and spectral measurement, and discloses an automobile ornament chromatic aberration detection method based on image recognition, which comprises the following steps: setting a known spectral reference spot on site and collecting multiband imaging response and spectral color measurement data, and constructing closed linear mapping to estimate a pixel-level discrete reflectance spectrum; performing energy normalization on the estimated spectrum, and calculating a spectrum gravity center, a spectrum width, first-order differential energy and second-order differential energy to form a spectrum signature; constructing a self-adaptive threshold value based on a local reference region statistical median and a robust scale, and classifying abnormal pixels by adopting a determination logic of a spectrum angle and a signature standardization distance; the method is suitable for online and offline appearance detection scenes of a production line, and can distinguish the overall color cast and spectral distortion, thereby improving the interpretability and robustness of the whole area array detection.

The present invention relates to the field of image processing, and proposes a snapshot compression imaging reconstruction method and system based on dual low-rank regularization. By designing a dual low-rank regularization model, three-dimensional hyperspectral image data is two-dimensionalized, thereby ensuring the quality of original data and avoiding the problem of poor reconstruction effect caused by data loss or error. The dual low-rank regularization model combines a low-rank group sparse prior algorithm and a deep low-rank decomposition algorithm to form a dual low-rank regularization framework. The low-rank group sparse prior algorithm is used to extract prior information of global spectral correlation and local spatial smoothness, respectively, and the texture detail features of the reconstruction result are enhanced. The deep low-rank decomposition algorithm is used to extract spatial factors and spectral factors, respectively, and deep information is extracted quickly and effectively, further avoiding the problem of texture detail loss and spectral distortion in the reconstruction result. The present invention improves the reconstruction accuracy and authenticity of snapshot compression imaging.

The present invention proposes a panchromatic and multispectral remote sensing image fusion method based on GAN and Transformer to solve the problems of loss of detail information such as texture and spectral distortion. This method adopts a specific generative adversarial network structure and designs parallel base layer and detail layer branches to effectively extract the texture and spectral features of the image. The source image is decomposed by guided filters. By using convolutional neural networks and Transformer networks, this method can effectively extract the local texture and global spectral features of the image respectively. The discriminator part contains two specially designed sub-networks to discriminate the base layer and detail layer respectively. This method meets the high-quality requirements of remote sensing image fusion to ensure that in the fusion process, the texture information of the panchromatic image with high spatial resolution is retained, and the spectral information of the multispectral image with high spectral resolution is transmitted as much as possible, providing a powerful tool for remote sensing image analysis.

This invention discloses a remote sensing image fusion method combining ratio transformation and distribution transformation, comprising: performing mean filtering on a panchromatic image and an upsampled multispectral image to obtain high-frequency components of the panchromatic image and the upsampled multispectral image; based on this, obtaining the missing high-frequency details in the multispectral image, denoted as the first high-frequency details; performing standard normalization on the first high-frequency details to obtain the second high-frequency details; calculating the mean and standard deviation of each pixel in each channel of the upsampled multispectral image; concatenating the upsampled multispectral image and the first high-frequency details and inputting them into a convolutional network to generate two affine transformation parameters; injecting the obtained mean and standard deviation into the second high-frequency details to generate high-frequency details with the same distribution as the upsampled multispectral image; and combining this with the upsampled multispectral image to obtain the final fused image. This method solves the problems of spectral distortion and detail distortion in existing remote sensing image fusion algorithms.

This invention discloses a method for verifying and confirming numerical prediction thrust spectrum results of propellers, belonging to the field of computational fluid dynamics (CFD) result verification and confirmation. While maintaining consistency between the physical model and boundary conditions, this invention encrypts the CFD calculation settings with joint discrete parameters and quantitatively characterizes simulation errors and uncertainties based on Richardson extrapolation. This invention employs the Lomb-Scargle periodogram method to directly process non-uniformly sampled data, avoiding spectral distortion caused by resampling interpolation and improving the accuracy of spectrum estimation. This invention automatically selects the optimal AR model order through information criteria, models colored noise backgrounds in actual engineering, and achieves adaptive noise modeling. This invention simultaneously evaluates the statistical significance of the global maximum peak value and each expected harmonic frequency in a single Monte Carlo simulation cycle. This invention uses a block bootstrap method to calculate the frequency band power confidence interval, maintaining the local correlation of the time series and achieving reliable confidence interval quantification.

The invention relates to the technical field of remote sensing image processing, in particular to a hyperspectral and multispectral remote sensing image fusion method, electronic equipment and a storage medium, and the hyperspectral and multispectral remote sensing image fusion method comprises the following steps: grouping a hyperspectral image and a multispectral image to be fused according to a spectral range; wherein the spectral range of the hyperspectral image in each group is matched with the spectral range of the multispectral image; for each group, fusing the hyperspectral image and the multispectral image in the group to obtain a fused image; and overlapping the fusion images corresponding to the groups to obtain a final fusion result. According to the scheme, the hyperspectral image and the multispectral image are grouped according to the spectral range, so that the spectral range of the hyperspectral band and the spectral range of the multispectral band in each group are matched, and the problem of spectral distortion caused by inconsistent spectral ranges is avoided.

The invention provides a hyperspectral image reconstruction method and system based on flow matching prior assistance, and the method comprises the steps: generating a low-dimensional condition vector based on compressed two-dimensional test data, and enabling the low-dimensional condition vector to be used for constraining a flow matching model process, and obtaining a hidden space vector; setting an initial iteration frequency and an initial signal vector, and determining a sensing matrix and measurement data; and based on the initial signal vector, the sensing matrix and the measurement data, performing loop iteration of N stages until a final reconstructed hyperspectral image is obtained, wherein the k-stage iteration specifically comprises two processes of generalized alternating projection and a de-noising device. And after N stages of iteration are completed, taking the hyperspectral image reconstruction estimation value of the Nth stage as a final reconstructed hyperspectral image. According to the method, the continuous probability modeling capability of stream matching is utilized, the spatial-spectral information recovery capability of the hyperspectral image is enhanced, and the problems of spectral distortion and spatial blur in a complex degraded scene in a traditional method are avoided.

Systems and methods for waveform repair. The methods comprise: receiving a signal of interest having a spectral distortion, the signal of interest having a sequence of symbols; generating an estimate for each symbol of a sequence of symbols in the signal of interest; and applying reconstruction filtering to the estimates to correct the spectral distortion.

The invention relates to the technical field of image processing, in particular to a liquid crystal filtering multispectral image generation method based on quantum inspiration. According to the invention, a quantum inspiration optimization structure is combined with liquid crystal filtering hardware, the filtering parameter optimization speed and precision are improved by referring to quantum superposition and parallelism characteristics, and the spectral band dynamic switching is realized by using the electrically-controlled adjustable characteristic of liquid crystal filtering; the problems that a traditional mechanical filtering structure is poor in adjustability and low in integration level and a traditional algorithm is low in high-dimensional data processing efficiency are solved, and high-quality original signals and efficient calculation support are provided for multispectral image generation. According to the method, an image quality analysis and feedback mechanism is constructed, the image quality is evaluated based on parameters such as spectral angular distance and residual noise energy, bad images are fed back to an optimization link to be reprocessed, an optimization frequency threshold value is set to avoid invalid circulation, the quality stability of output images is effectively guaranteed, and the method is suitable for large-scale popularization and application. And the problems of spectrum distortion, noise and difficulty in balancing details are solved.

The invention provides a hyperspectral image super-resolution reconstruction method and device, equipment and a medium, and relates to the technical field of remote sensing image processing. According to the method, the input low-resolution hyperspectral image is acquired and preprocessed, and the full-image semantic object mask is generated by using the pre-trained semantic segmentation model, so that the semantic index map is obtained; then, constructing a hierarchical global spectrum prototype memory bank, and aggregating cross-sample category-level spectrum priori through momentum update; designing a deep reconstruction network formed by a plurality of cascaded semantic prototype interaction units, and performing feature enhancement on the input initial feature map by combining local prototype perception, object pixel cross attention, window self-attention and a feedforward network; and finally, a high-resolution image is output through sub-pixel convolution reconstruction. According to the method, auxiliary images are not needed, spectral distortion and boundary blur are effectively relieved, and reconstruction fidelity and detail representation are improved.

The invention discloses a hyperspectral image compression network and compression method based on multi-scale spectrum and spatial feature enhancement, and mainly solves the problems of insufficient hyperspectral image spectrum modeling, insufficient spatial feature extraction and high network complexity in the prior art. The network comprises a main encoder, a main decoder, a super-prior encoder, a super-prior decoder and an entropy model. The main encoder comprises a spectral attention gating data unit, a convolution unit and a multi-scale spatial adaptive feature attention enhancement unit, and is used for converting an input image into potential representation and removing spatial and spectral redundancy; the main decoder and the main encoder are symmetrical in structure; the super-prior encoder comprises a convolution layer and an activation layer and is used for extracting auxiliary information from the output of the main encoder; the super-prior decoder and the super-prior encoder are symmetrical in structure; the entropy model is Gaussian distribution based on output parameters of a super-prior decoder. After the network is trained, lossy compression of a hyperspectral image can be realized. The method reduces the network complexity and spectral distortion, improves the reconstruction quality of complex ground feature details, and is suitable for earth observation, meteorological monitoring and the like.

The application provides an image fusion method and device and a storage medium, and belongs to the technical field of image fusion.The method comprises the following steps: importing a high-resolution hyperspectral image to be fused, a high-resolution multispectral image to be fused, a plurality of original high-resolution hyperspectral images and an original high-resolution multispectral image; performing down-sampling processing on the original high-resolution hyperspectral images respectively to obtain original low-resolution hyperspectral images; and constructing a training model, wherein the training model comprises a feature analysis network, a target vector field construction network and a predicted vector field construction network.The application effectively avoids the spectral distortion problem caused by the difference in spectral response of a sensor in a traditional fusion method, and overcomes the defects of lack of constraints in the intermediate process and insufficient stability of results in a single-step regression method.