108 results about "Spectral similarity" patented technology

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 discloses a large-area crop rapid mapping method integrating multi-source remote sensing data, and the method comprises the steps: obtaining a multi-source remote sensing satellite image, carrying out the preprocessing, amplifying a sample based on the spectrum similarity of a field collection sample point and an amplification sample point, and screening the amplification sample point based on a dynamic time warping distance; drawing a crop distribution reference map of the research subarea in cooperation with multi-source data; and based on the crop distribution reference diagram of the research sub-region, constructing a research region sub-pixel drawing sample set and predicting a sub-pixel distribution diagram of each crop type. According to the method, existing field collection sample points are fully utilized, the drawing potential of multi-source remote sensing data and different classifiers is explored, and crop sub-pixel rapid drawing of a large-scale research area is achieved.

The invention provides a fruit tree canopy three-dimensional model reconstruction method and a fruit tree canopy three-dimensional model reconstruction system based on multi-modal data, and relates to the technical field of fruit tree canopy three-dimensional model reconstruction based on multi-modal data. Sub-pixel-level space registration is achieved by means of ground control points, voxelization division is carried out based on the range of a point cloud bounding box, and the number of point clouds is screened to construct an initial seed set; and adopting a structure-spectrum collaborative growth strategy, taking the adjacent voxels which are not accessed as candidate voxels, calculating a spectrum similarity mean value of the candidate voxels and directly adjacent voxels in the seed set, stopping growth when no new voxels are added in continuous three times of iteration, and outputting a final seed set. Meanwhile, the set is converted into a three-dimensional canopy model fusing space coordinates and full-wave-band spectral reflectivity, voxel-level pathological labeling is achieved based on the chlorophyll sensitivity index and the blue-red slope, and the whole process of reconstruction and analysis is completed.

The invention relates to the technical field of light emitting of LED lamp beads, and discloses a uniform control method and device for the light emitting color of the LED lamp beads and the LED lamp beads, and the method comprises the steps that LED wafers are placed in a triangular annular mode, and the LED lamp beads are prepared through centrifugal treatment of double-layer packaging glue; acquiring an initial optical parameter matrix after applying rated working parameters, and establishing a standard data model; monitoring the LED lamp beads in real time to obtain a monitoring data set; calculating a multi-dimensional deviation vector by using a spectral similarity algorithm, an Euclidean distance algorithm and a ratio algorithm; a comprehensive deviation evaluation value is obtained through weighted summation; and generating a correction amount based on the evaluation value to carry out closed-loop control until a control termination condition is reached. According to the method, accurate control over the light emitting colors of the LED lamp beads is achieved, and the uniformity and stability of the light emitting colors of the LED lamp beads are improved.

A light source module, a lighting system and a lamp. The light source module comprises a first light-emitting unit (100), a second light-emitting unit (200) and a third light-emitting unit (300) which are electrically independent of each other, light emitted from the light-emitting units being mixed to form white light. By means of the selection of the light colors and spectral characteristics of the light-emitting units, the white light obtained by the light source module has a wide color temperature coverage range and a color deviation Duv value less than 0.003, thereby solving the problem of white light color deviation. In addition, the white light spectrum of the light source module simulates a target spectrum, when the correlated color temperature is 4000K or above, the target spectrum being the solar spectrum, and when the correlated color temperature is lower than 4000K, the target spectrum being the blackbody radiation spectrum. Because the spectrum simulates the solar or blackbody radiation spectrum, the obtained white light spectrum has a good color rendering index and spectral similarity, such that the color of the white light of the light source module is closer to that of daylight, thereby being more adapted to the field of general lighting.

The invention provides a remote sensing image space-time fusion color correction method based on virtual low-resolution reconstruction, and the method comprises the steps: obtaining a to-be-reconstructed original remote sensing image, and determining a corresponding reference image according to the features of the original remote sensing image; performing virtual low-resolution image reconstruction by using the original remote sensing image and the reference image to obtain a virtual low-resolution image; and based on the virtual low-resolution image, performing color correction processing on the original remote sensing image by adopting a space-time fusion strategy to obtain a uniform color result image. According to the method, a weighted space-time filtering mechanism is introduced into color correction processing, a virtual low-resolution image is generated as auxiliary data, time correlation, spatial proximity and spectral similarity among the virtual low-resolution image, an original image and a reference bottom image element are comprehensively considered, spectral similar pixels of a central pixel in a filtering window are utilized to participate in result operation, and the color correction accuracy is improved. And generating a result image with the detail features of the original image and the spectral features of the reference image.

The invention relates to the field of electromagnetic super technology, in particular to a metasurface structure design method based on a joint discriminant generative adversarial network, which comprises the following steps: designing periodic arrangement of an electromagnetic metasurface unit structure and related parameters, inputting the periodic arrangement into electromagnetic simulation software to obtain electromagnetic response, and constructing a forward prediction data set and a reverse design data set; respectively segmenting the forward prediction data set and the reverse design data set into a training set and a test set; a forward prediction network model is constructed, a forward prediction data set is sent to a simulator network simulator for forward prediction for training, and parameters of the forward prediction network model are fixed after training is completed. According to the method, the pre-trained forward prediction network and the discriminator network for discriminating the spectrum similarity are added into the reverse design network, so that the network can well understand the relationship between the structural design and the overall optical response, more attention is paid to the generation of the absorption spectrum of the structure, a new design can be generated, and the design efficiency is improved. And the method can also be expanded to analysis of other metasurface structures.

The invention provides an ore recognition method and system based on a large model algorithm, and is applied to the technical field of computer vision and data processing, and the method comprises the steps: obtaining an ore microscopic image, and generating a feature point location candidate set through a convolutional neural network model; calculating a spectral value and a morphological parameter of each feature point location to determine a multi-dimensional data point; clustering the multi-dimensional data points to obtain a feature point cluster; and when the spatial distance between the adjacent feature point position clusters is smaller than a first preset threshold value and the spectral similarity is higher than a second preset threshold value, combining the adjacent feature point position clusters into the same mineral region. The segmentation precision and the type identification accuracy of the complex symbiotic mineral area can be obviously improved.

The light source module comprises a first light-emitting unit, a second light-emitting unit, a third light-emitting unit and a fourth light-emitting unit which are electrically independent from one another, and light emitted by the light-emitting units is mixed to form white light. By selecting the light color and spectral characteristics of the light emitting unit, the white light color temperature coverage range obtained by the light source module is wide, the color deviation Duv value is smaller than 0.003, and therefore the white light color deviation problem is solved. Meanwhile, the white light spectrum of the light source module simulates a target spectrum, and when the relative color temperature is more than 4000 K, the target spectrum is a sunlight spectrum; and when the relative color temperature is lower than 4000 K, the target spectrum is a blackbody radiation spectrum. Due to the fact that the spectrum simulates sunlight or a black body radiation spectrum, the obtained white light spectrum has good spectrum similarity and color rendering, the light color of the white light of the module is closer to sunlight, the light quality performance of the white light source is obviously superior to that of an existing double-color-temperature light mixing white light source, and the module is more suitable for the field of general illumination.

The invention relates to a self-supervised meta transfer learning hyperspectral target detection method and system, and belongs to the technical field of deep learning. The method comprises the following steps: S1, preparing data of an initial source domain and a target domain of a hyperspectral image; s2, carrying out self-supervision pre-training on the GLSL network by utilizing the source data; s3, migrating the GLSL to a target detection data set to carry out spectrum similarity detection; and S4, performing learning constraint by using a spatial constraint module to obtain a target detection result. The performance of the method is superior to that of other hyperspectral image target detection methods, the method can be effectively migrated to different target detection tasks, and the method has advantages in the aspect of background suppression compared with other methods.

The invention discloses a remote sensing image semantic change detection method, which belongs to the field of remote sensing image semantic change detection, and comprises the following steps: carrying out double-temporal feature extraction by using a residual network, sending the double-temporal features into a constructed difference enhancement module to model difference information, and obtaining a change weight of the double-temporal features; the change weight and the double-temporal feature are sent to a feature interaction module, interaction among the three branches is completed, the interacted double-temporal feature and a differential feature rich in semantic information are obtained, a binary change graph is obtained from a segmentation head to a semantic segmentation graph and through a change detection head, the semantic segmentation graph is superposed on the obtained binary change graph, and the binary change graph is obtained. And finally generating a semantic change graph. Compared with other existing semantic change detection methods, the method provided by the invention is more excellent, can reduce leak detection conditions under different spectrums similar in semantics and error detection conditions under the same spectrums change semantics, and also has a good effect in detail and edge detection.

The invention relates to the technical field of high-molecular compound preparation, in particular to a high-molecular compound preparation method and system.The method comprises the steps that distilled water and boron nitride are added into a container containing silane monomers, stirring reaction is conducted for 3-5 h, and a hyperspectral image of a mixed solution in the container in the stirring reaction process is collected; the method comprises the following steps: acquiring spectral similarity factors of matched pixels in a hyperspectral image, judging the fullness degree of a stirring reaction according to the distribution characteristics of the spectral similarity factors of the pixels, completing the stirring reaction, mixing the prepared boron nitride modified organic silicon resin, silanol, polyester resin and a crosslinking catalyst, continuously reacting for 2-4 hours at 110-130 DEG C, adding an organic solvent, and uniformly stirring to obtain the boron nitride modified organic silicon resin. The solid content of the modified polyester resin is 62-66%. The invention aims to improve the performance and prolong the service life of the prepared modified polyester resin, and realizes the preparation method of the high-molecular compound.

The invention relates to the technical field of illumination, and provides an LED light source which comprises a red light bead, a green light chip, a blue light chip, a purple light chip, an amber light bead, a green light chip and a lemon light bead. The amber light lamp bead comprises a second light-emitting chip and amber fluorescent powder covering the second light-emitting chip, and the lemon light lamp bead comprises a third light-emitting chip and lemon fluorescent powder covering the third light-emitting chip. Fluorescent powder is excited to emit light, so that the light brightness is improved, the half-peak wave width is increased, and the color continuity is better; chips are adopted to emit light, so that the purity of the color light is ensured, and the high-purity color light output requirement is met; on the basis, through matching and combination of the seven-color light lines, wider color gamut coverage and finer color transition are achieved, spectral distribution closer to sunlight can be generated, and the spectral similarity index (SSI) of the sunlight can be measured to reach 90 or above.

The invention relates to the technical field of remote sensing image processing, in particular to a deep learning-based euryale ferox traditional Chinese medicinal material remote sensing recognition method, which comprises the steps of acquiring data, extracting features, constructing an optimal feature set, constructing a convolutional neural network model and training the model. According to the method, crops with high spectral similarity can be distinguished with high precision, the whole process is automatic, manual intervention is not needed in the phenological feature extraction, feature optimization and intelligent interpretation processes, subjective deviation caused by manual experience is avoided, and compared with a manual assistance method, the overall processing efficiency is also improved, and the method is suitable for popularization and application. The method can be quickly popularized to automatic interpretation in a large area; meanwhile, optical and radar multi-source data are compatible, high classification precision can still be kept under the condition that optical images are interfered by clouds and rains or crop plots are seriously broken, and the applicability in complex scenes is remarkably improved.

The invention relates to a remote sensing image space-time fusion method for reducing time sequence heterogeneous interference, and belongs to the technical field of remote sensing image processing. Modeling and identifying the land cover type by using the remote sensing image based on a deep learning network; dividing the remote sensing image into a plurality of space subblocks at corresponding positions, and then performing quality screening; a window is arranged in each block, pixels similar to the central pixel are searched in the same land class range, a weight is allocated to each similar pixel in the window, and the calculation of the weights combines the spectral similarity and the spatial distance and introduces a dynamic factor of a time dimension; calculating a conversion coefficient of the known image and the predicted image; predicting a high-space low-time-resolution image of the prediction time; and setting a confidence interval to evaluate the credibility of the prediction result, and embedding the obtained high-quality image to obtain a final prediction result. The method not only improves the accuracy of pixel matching and the stability of a fusion result, but also improves the automation degree and efficiency of model operation.

The invention discloses a sea fog three-dimensional distribution detection method based on satellite-borne active and passive remote sensing fusion. The high-precision vertical section detection capability of a satellite-borne active remote sensing laser radar and the large-range long-time continuous observation capability of passive remote sensing of a stationary satellite are fully exerted, an optimal band combination is selected through a reconstruction verification method, the multi-band spectrum similarity is evaluated, the most suitable active remote sensing profile is selected, and the multi-band spectrum similarity is evaluated. And expanding the sea fog to a certain distance outside an active remote sensing detection range, and constructing three-dimensional distribution of sea fog levels. Compared with a sea fog level detection result of an existing method, the method has the advantages that the satellite three-dimensional detection range is effectively expanded, the method has the advantages of being high in spatial resolution, capable of detecting vertical-direction levels and the like, is suitable for multiple scenes, facilitates understanding of sea fog formation and dissipation mechanisms and can also be used as a high-precision three-dimensional initial field for sea fog forecasting.

The invention discloses a hyperspectral and multispectral image fusion method and system based on dictionary extension, and the method comprises the steps: obtaining a high-resolution multispectral image Ym and a low-resolution hyperspectral image Yh in an incomplete overlapping scene, dividing Ym into an overlapped high-resolution multispectral image # imgabs0 # and a non-overlapped high-resolution multispectral image # imgabs1 # to solve an overlapped region spectrum dictionary Ds and an overlapped region sparse coefficient As, and fusing and reconstructing an overlapped region high-resolution hyperspectral image Xs; the common spectrum # imgabs4 # of the # imgabs2 # and the # imgabs3 # is extracted, the common spectrum # imgabs5 # is used for carrying out spectrum dictionary expansion, an expanded non-overlapped region spectrum dictionary Dr and a non-overlapped sparse coefficient Ar are obtained, and a non-overlapped region high-resolution hyperspectral image Xr is reconstructed; and synthesizing Xs and Xr to obtain a high-resolution hyperspectral image X of a full coverage area of the high-resolution multispectral image Ym. According to the method, the sparse coefficient matrix of the non-overlapping area is solved by utilizing the spectral similarity and the spatial information of the multispectral image, and the image fusion of the full coverage area of the multispectral image in the incomplete overlapping scene is realized.

A method for identifying similar ground objects with spectral similarity in a remote sensing image based on condition number of matrix includes the steps of: obtaining sample's spectral curve of the ground object to be identified, and constructing the sample spectrum matrix A according to the number of bands of the image and the number of categories of ground objects; sequentially obtaining distance index vectors between the spectrum P of the pixel to be identified and each category of ground objects in A; calculating a distance error of each type of ground object when P is identified; then calculating a distance error threshold vector of A; classifying and identifying the pixels. This method uses linear system of equations and the condition number theory to transform ill-conditioned equations into well-conditioned equations, thereby adding significant identifiable features to the ground objects to improve the identification performance and avoid constructing large training samples.

The invention discloses a grain hyperspectral image classification method based on incremental learning, and belongs to the field of image classification, and the method comprises the following steps: S1, the construction of an experimental data set: selecting a plurality of old-class rice grains and a plurality of new-class rice grains, collecting a fixed number of samples for each class, dividing a historical training set, a historical test set, a new category training set and a new category test set, converting the hyperspectral original data into a Tensor format, and normalizing the hyperspectral original data to [0, 1]; s2, initial model training; s3, new category incremental training: extracting historical category statistical parameters from a feature memory module to construct an anchoring pool, screening Top-3 similar historical categories, embedding a diffusion model noise adding process, injecting spectrum and space priori knowledge to generate virtual samples, performing dual dynamic verification, and combining real and virtual samples to form a new category training set; and calculating the distillation weight based on the spectrum similarity, and training the model through the dual-target distillation loss.

The invention relates to the technical field of remote sensing image processing, in particular to a geographic information acquisition method and system based on remote sensing image.The method comprises the following steps that a narrow-band spectrum sensor obtains multiband image data, reflectivity peak value coordinates are extracted through peak value detection, and band peak value displacement data are calculated; the method comprises the steps of extracting peak displacement of wave bands, analyzing stability through Kalman filtering, screening a stable wave band set, constructing a standard spectrum feature library based on the set, calculating spectrum similarity values through difference values, outputting land utilization types through Bayesian classification, and generating a geographic information distribution diagram. The spectral expression is constructed based on stable reflection characteristics, the influence of imaging condition fluctuation on samples is reduced, the type distinction degree is enhanced through stable wave band difference value symbol matching, the category judgment accuracy based on likelihood inference is improved, classification keeps coherence in an illumination change scene, and the reliability and consistency of spatial distribution results are improved.

The present invention discloses a spectral smoothing method applicable to infrared spectral data. The steps of the method are as follows: (1) Determine whether the infrared spectral data is sampled at equal intervals in the spectral dimension; (2) If the infrared spectral data is non-uniformly sampled, perform uniform resampling on the infrared spectral data; (3) Perform mean filtering on the uniformly sampled infrared spectral data; (4) Calculate the spectral similarity before and after mean filtering; (5) Determine whether the spectral similarity is within the threshold range. If it is not within the threshold range, continue to perform mean filtering; if it is within the threshold range, output the smoothed infrared spectral data. The method of the present invention has the characteristics of good accuracy, easy operation, and easy determination of parameters.

The invention discloses a garbage room comprehensive management method and system. The method comprises the following steps: establishing a recyclable molecular spectrum knowledge base; collecting image information of the garbage; calling interface identification; classifying for the first time; packaging the non-resourceful wastes; collecting near infrared spectrum data of the recyclable wastes; screening key characteristic wavelength points; similarity calculation is conducted, recyclable molecules of the recyclable waste are obtained, and mixed spectrum similarity measurement is adopted for similarity calculation; and performing secondary classification. The system comprises an information acquisition module, an intelligent robot classification and packaging module, a data module and a central control module. According to the method, the garbage is automatically identified, and the recyclable wastes in the garbage are divided out for independent treatment, so that the utilization rate of the wastes is increased; and meanwhile, a recyclable molecular spectrum knowledge base is established, the characteristic spectrum of the recyclable waste and the recyclable molecular spectrum are subjected to mixing similarity calculation, the classification of the recyclable waste is judged, and the subsequent recycling efficiency is improved.

The invention discloses a tree species identification method based on multi-modal spectrum and texture features of a wood cross section. The method comprises the following steps: acquiring hyperspectral image data of the wood cross section; constructing a comprehensive similarity matrix based on a plurality of spectrum similarity indexes, and selecting a representative wave band subset from the hyperspectral image data by adopting a multi-strategy wave band screening mechanism; performing multi-scale wavelet fusion on the representative wave band subset to generate a single-channel fusion image with consistent spatial resolution, and extracting points of interest and spectral features thereof from the single-channel fusion image; generating a gray-scale base map based on the hyperspectral image data, extracting various complementary texture feature maps from the gray-scale base map, and screening out significant points of interest from the various texture feature maps; and fusing the spectral features and the texture features, constructing feature vectors for representing wood tree species, and inputting the feature vectors to a classification model to complete tree species identification. According to the invention, rapid, accurate and intelligent identification of wood tree species can be realized.

This invention discloses a moonlight-spectrum LED packaging structure and method. The structure includes a bracket, a bowl, a first LED chip, a second LED chip, a third LED chip, a fourth LED chip, a fifth LED chip, a sixth LED chip, a seventh LED chip, and an eighth LED chip. The method includes steps one, fixing the chips; step two, connecting the circuits; step three, preparing the fluorescent adhesive solution; step four, dispensing the adhesive; and step five, parameter spectral dispersion. In step five, qualified products must meet the following conditions: chromaticity bandwidth controlled within a 3rd-order McAdam ellipse; light color quality satisfying Ra>93, R9>90, TM-30-18, Rg>100, Rf>95; color temperature satisfying 3800-4200K; chromaticity coordinates satisfying the 403 chromaticity standard; and moonlight spectrum SSI (spectral similarity index) coefficient>93%. Compared to existing LED lights, this invention is designed as a moonlight-spectrum LED. The light color quality of this LED is close to moonlight, and the emitted light has the soft beauty of moonlight, which can create a romantic atmosphere when used for auxiliary lighting.

The invention relates to a method for rapidly identifying degenerated strains of phlebopus portentosus. The method comprises the following steps: obtaining a normal test strain group, a degenerated test strain group and a typical degenerated strain group according to hypha morphology and growth rate of different strains; the method comprises the following steps: measuring infrared spectrum information of a strain by adopting a Fourier transform attenuated total reflection infrared spectrometer, and storing an infrared spectrogram and data of the strain in the infrared spectrometer to establish a spectrum similarity retrieval spectrum library; and performing spectrum retrieval matching on the acquired infrared spectrogram and data of the unknown sample to be detected and the spectrum in the spectrum similarity retrieval spectrum library, and identifying whether the unknown sample to be detected is degraded or not according to the matching value. According to the method disclosed by the invention, the spectrum similarity retrieval spectrum library for quickly identifying normal and degraded strains is established through the infrared spectrums of the strains of the normal test strain group, the degraded test strain group and the typical degraded strain group, so that the purpose of judging the degraded strains in advance is achieved, and a scientific basis is provided for the use of the phlebopus portentosus strains.

The present application relates to the technical field of remote sensing information extraction and snow monitoring, and particularly relates to a sub-pixel snow mapping method based on medium-resolution remote sensing data, comprising the following steps: obtaining medium-resolution remote sensing images, a digital elevation model, and ground meteorological station snow observation data, and preprocessing the remote sensing images to obtain surface reflectivity; identifying cloud-covered pixels, filling in the cloud-free dataset by temporal and spatial interpolation and ground object spectral similarity matching; selecting snow, vegetation, and water as spectral endmembers and extracting features; constructing a linear mixing model, and using the least square method to retrieve the snow proportion in the pixel to obtain preliminary data; and generating a high temporal and spatial accuracy snow distribution map through meteorological station data verification, digital elevation model terrain correction, and model parameter optimization, which solves the problems of mixed pixels, cloud coverage, terrain errors, and the lack of accuracy verification.

The invention discloses a remote sensing image crop classification method based on space-spectrum joint enhancement contrast learning. According to the method, a three-dimensional space-time spectrum feature encoder is designed to perform preliminary extraction of crop features. Secondly, extracting spatial and spectral features of the remote sensing image in a multi-view joint contrast learning mode based on the features of the remote sensing image, and introducing an optimal index method into a spectrum enhancement module to perform feature optimization from the multi-spectral image to construct a spectrum enhancement view so as to improve the utilization efficiency of spectral information; the spatial enhancement module utilizes the spatial distribution relationship of crops and combines spectral similarity modeling to construct a spatial enhancement view for enhancing feature association between adjacent farmland samples. Compared with an existing most advanced self-supervised comparative learning method, the method has the advantages that effective spatial-spectral joint semantic features can be mined from unlabeled remote sensing data, the performance of the model in crop classification tasks is improved, and the method has good application value.

The application relates to the field of electromagnetic super technology, in particular to a metasurface structure design method based on a joint discriminant generative adversarial network, a periodic arrangement of a metasurface unit structure and related parameters is designed, and is input into electromagnetic simulation software to obtain electromagnetic response, a forward prediction data set and a reverse design data set are constructed; the forward prediction data set and the reverse design data set are respectively divided into a training set and a test set; a forward prediction network model is constructed, the forward prediction data set is sent into a simulator network Simulator used for forward prediction to perform training, and the forward prediction network model parameters are fixed after the training is completed. The application adds a pre-trained forward prediction network and a discriminator network for discriminating spectral similarity into a reverse design network, so that the network can well understand the relationship between structure design and overall optical response, more attention is paid to the absorption spectrum of the generated structure, new design can be generated, and the application can be extended to analyze other metasurface structures.

The present application relates to the technical field of material testing, in particular to a spectral temperature brightness degradation prediction method of polyurethane functional material component. It constructs a weighted joint regression network prediction model for predicting the polyurethane functional material component; the spectral similarity, temperature similarity and brightness similarity of the polyurethane functional material component are predicted through the weighted joint regression network prediction model. A functional material temperature, brightness and spectral feature degradation characterization model based on the weighted joint regression network is proposed, and through the weighted joint regression of the temperature, brightness and spectral feature image detection results, compared with the traditional one regression fitting algorithm, the situation of falling into local minimum solution and overfitting can be avoided, so that the prediction result is closer to the actual situation.

This invention discloses a collaborative classification method for hyperspectral imagery and lidar data, comprising: a spectral spatial feature interaction enhancement module, a graph convolution module, a multi-head self-attention module, and a progressive feature fusion module for spectral spatial information. Specifically, firstly, deep spectral spatial information between hyperspectral images and lidar data is mined through weighted feature fusion and cross-modal convolution, extracting spatial element-level and spectral channel-level interaction features respectively. Then, a graph convolution network is used to extract spatial information from the spatial element-level and spectral channel-level interaction features, calculate spectral similarity, and construct an adjacency matrix. Next, the spatial features are learned through graph convolution operations to obtain new spatial feature representations. Then, a multi-head self-attention mechanism is used to enhance the global dependency of spectral features, fusing the extracted spatial and spectral features. Finally, a maximum-based decision fusion and progressive feature fusion strategy are employed to integrate features at different levels, obtaining the final fused features, which are then used for final classification and output. This invention acquires rich spatial-spectral features simultaneously through a spectral spatial feature interaction enhancement module, a graph convolution module, and a multi-head self-attention module. It also incorporates a spectral spatial information progressive feature fusion module to regulate the information interaction at different perception levels, thereby leveraging the synergistic advantages of multimodal information and significantly improving the accuracy and stability of land cover classification tasks.