90 results about "Remote sensing application" patented technology

The invention discloses a space-time spectrum combined super-resolution reconstruction method based on a giant remote sensing satellite group. The method comprises the following steps: acquiring time series, multi-view and multi-spectral data of the same area from a plurality of heterogeneous satellites, and performing radiometric calibration and atmospheric correction; sub-pixel-level alignment of the multi-source data is realized by adopting a joint registration model; extracting time change features by using three-dimensional convolution, extracting space structure and texture features by using two-dimensional convolution, and extracting and reducing the dimension of spectral features by using one-dimensional convolution; performing adaptive weighted fusion on time, space and spectral features through an attention mechanism to generate a joint feature tensor; and carrying out super-resolution reconstruction to obtain a target image with high spatial resolution, high time resolution and high spectral fidelity. The method gives consideration to both resolution improvement and spectrum authenticity, and is suitable for high-precision remote sensing application scenes such as fine urban mapping, agricultural monitoring, ecological environment assessment, disaster emergency and battlefield situation awareness, remote reconnaissance, target change detection and damage assessment.

The invention relates to a multi-modal image matching method and system based on learning features and epipolar geometric constraints. The method comprises the following steps: carrying out edge enhancement processing on an input image through wavelet transform; extracting a multi-scale dense feature map based on the transformed convolutional neural network, and generating a feature descriptor with rotation and scale invariance in combination with principal direction normalization; adopting an FLANN algorithm and dynamic distance constraint to realize preliminary feature matching; and introducing a basic matrix construction and epipolar geometric consistency verification mechanism, and eliminating mismatching point pairs in combination with an RANSAC affine constraint model. According to the method, image enhancement, deep learning and geometric verification strategies are fused, the problems of radiation nonlinearity and geometric distortion caused by imaging mechanism differences among multi-modal images are effectively solved, the matching precision and robustness are improved, and the method is suitable for remote sensing application scenes such as optical-SAR registration, multi-source image fusion and earth surface change detection.

The invention provides a multi-source data fusion variable fertilization method and device, and belongs to the field of agricultural production technology and agricultural remote sensing application, and the method comprises the following steps: data acquisition and preprocessing: collecting historical data, ground sampling data, ground sensor data and remote sensing images; spatial data processing and consistency correction; variable fertilization model construction: constructing a fertilization amount and yield effect curve, constructing a yield and growth vigor index VI distribution diagram, and further constructing a variable fertilization model representing the relationship between the fertilization amount and the growth vigor index VI; the spatial heterogeneity quantification and zoning comprises the following steps: quantifying the spatial heterogeneity of environmental factors on fertilization requirements by using a geographic detector, and dividing a region into a plurality of sub-regions; and differential fertilization strategy optimization: generating a dynamic adjustment coefficient, and determining a final fertilization amount in combination with the fertilization amount recommended by the variable fertilization model. The crop growth condition can be described more comprehensively and accurately, and accurate agricultural management can be guided.

The invention relates to the technical field of remote sensing image processing, and particularly discloses a remote sensing image target detection method based on a multi-scale cross-stage network model, and the method comprises the following steps: obtaining a to-be-detected remote sensing image; the method comprises the following steps of: inputting a backbone network after passing through an input layer, extracting multi-scale and multi-level semantic features step by step, and obtaining feature maps of different sizes through multi-scale down-sampling; inputting the feature maps of different sizes into a feature integration network, and performing cross-stage fusion and lightweight optimization to obtain a plurality of fused feature maps; and inputting the plurality of fusion feature maps into a prediction head, and carrying out classification detection to obtain detection results of multiple scales. According to the method, the brand new multi-scale cross-stage network model is constructed, multi-scale feature extraction, fusion and target recognition capabilities are enhanced, accurate detection of multi-scale targets, especially small targets and complex-form targets, in the remote sensing image can be realized, detection precision and reasoning efficiency are both considered, and the method is suitable for remote sensing application scenes such as ocean monitoring and urban planning.

The invention belongs to the technical field of physical ocean and remote sensing application, and particularly relates to a sub-mesoscale signal extraction method based on Ku / Ka dual-frequency SAR (Synthetic Aperture Radar) height measurement, which comprises the following steps of: acquiring a Ku / Ka dual-frequency sea surface height observation original data set and IMU (Inertial Measurement Unit) attitude and trajectory data; registering the average sea surface and tide height to a space-time sampling position of Ku / Ka synchronous observation and deducting from the original height to obtain an initial sea surface height anomaly sequence of the two channels, and processing the initial sea surface height anomaly data of the two channels to obtain a fused height anomaly sequence; performing attitude coupling mode extraction on the fused height anomaly sequence to obtain a height anomaly sequence after platform error elimination; meanwhile, smoothing processing is carried out to obtain a smoothed sequence; and interpolating the smoothed sequence into a two-dimensional regular grid, performing small-scale secondary smoothing denoising on grid data, interpolating a mesoscale background field at the same time into the same grid, and deducting the mesoscale background field to obtain a sea surface height anomaly data system with sub-mesoscale as a main part.

The invention discloses a Chla monitoring method based on visible-near infrared spectrum and machine learning, which is a modeling method for performing chlorophyll a concentration parameter inversion by using visible-near infrared hyperspectral data, and combines primary screening of spectral characteristic wave bands, training sample expansion based on GAN, spectral characteristic wave band fine screening based on CARS and a regression modeling technology. The problems of high dimension of hyperspectral data and insufficient samples are solved, the overall Chl-a modeling precision is improved, and the method is suitable for water eutrophication monitoring, marine ecological assessment and environment remote sensing application.

The invention belongs to the technical field of remote sensing super-resolution images, and particularly relates to a remote sensing image super-resolution reconstruction method based on an adaptive gating Transform. The method comprises the following steps: S1, constructing a training set; s2, constructing a self-adaptive gated Transform super-resolution reconstruction network, and training the self-adaptive gated Transform super-resolution reconstruction network by using the training set, so as to obtain a self-adaptive gated Transform super-resolution reconstruction model; and S3, inputting a remote sensing low-resolution image to be reconstructed into the adaptive gating Transform super-resolution reconstruction model to obtain a super-resolution reconstruction image. The invention aims to break through the limitation of an existing method in the aspects of efficiency, detail recovery and multi-scale fusion, provides a super-resolution solution with high fidelity and high efficiency for remote sensing images, and meets the urgent demand of refined remote sensing application for high-quality images.

The invention provides a space-space integrated crop growth monitoring point space-time design method and device, and belongs to the field of agricultural application technology and remote sensing application. Segmenting the monitoring area by grid units to obtain space monitoring units; performing time decomposition on the crop growth phenological period in the monitoring area to obtain a time monitoring unit; calculating a satellite monitoring space-time coverage probability matrix based on the time monitoring unit and the space monitoring unit; calculating a fixed ground monitoring space-time coverage probability matrix based on the time monitoring unit and the space monitoring unit; layering space monitoring units of the monitoring area based on the satellite monitoring space-time coverage probability matrix and the fixed ground monitoring space-time coverage probability matrix, and determining a space-time coverage area and a space-time missing area; and through multi-target space-time optimization, space-time layout of monitoring points and supplementary monitoring points is carried out. According to the invention, optimization is carried out and monitoring space-time point locations of different platforms are arranged based on multi-factor similarity, so that an integrated monitoring scheme is formed.

The invention relates to a hyperspectral image classification method based on a bidirectional interactive fusion space-spectrum multi-order gating aggregation network, and belongs to the field of remote sensing image processing. The method comprises the steps of determining a data set; data preprocessing: performing sample block extraction on the hyperspectral data set, and dividing the hyperspectral data set into a training set, a verification set and a test set; network construction: constructing a space-spectrum multi-order gating aggregation network based on bidirectional interactive fusion, wherein the space-spectrum multi-order gating aggregation network is used for hyperspectral image classification; the hyperspectral samples in the training set are input into the constructed network in batches for training, and after each training batch is completed, the classification performance is evaluated by using the verification set samples; and sample classification: inputting the hyperspectral samples in the test set into the trained classification network to obtain a final classification result. According to the method, efficient feature extraction can be realized, high-accuracy classification is performed on the hyperspectral image, and the method can be widely applied to remote sensing application fields such as hyperspectral image surface feature category detection and recognition.

The invention relates to the technical field of fire remote sensing application, and particularly discloses a method suitable for wind cloud satellite fire remote sensing application, and the method specifically comprises the following steps: 1, data preprocessing: obtaining wind cloud satellite fire original data, and carrying out the comprehensive preprocessing of the wind cloud satellite fire original data; step 2, identification model construction: based on radiation transmission and fire point spectral characteristics, establishing an identification model in combination with multi-spectral channel data; and inversion: inversing parameters of fire point temperature and area based on a thermal radiation transmission theory and multi-channel data. All-weather and high-time-efficiency fire monitoring is achieved, dynamic fire monitoring data are obtained in real time, the monitoring range is wide, the time frequency is high, the accuracy is high, the core area of a fire scene is accurately positioned through the hyperspectral monitoring technology, key data support is provided for rescue, and the rescue efficiency is improved. According to the recognition model, wind cloud satellite fire remote sensing is applied to fire dynamic monitoring and evaluation, fire emergency drilling, fire dynamic early warning and fire plan.

The invention discloses a multi-band multi-polarization SAR (synthetic aperture radar) image fusion classification method for modal missing and non-registration scenes. The method comprises the following steps: step 1, data preprocessing; step 2, feature extraction; step 3, carrying out cross-modal constraint; step 4, carrying out modal Dropout; 5, performing feature fusion; and step 6, performing classification decision. According to the method, spatial alignment or resampling does not need to be carried out on the original image, semantic consistency and modal fusion robustness among different frequency bands can be effectively improved, high-reliability and high-adaptability polarized SAR image intelligent classification in a complex remote sensing scene is realized under the condition that the frequency bands are incomplete or the resolutions are inconsistent, and the method is suitable for popularization and application. The method solves the problem that the classification precision of the existing multi-band multi-polarization SAR image is reduced under the conditions of inconsistent image size, incapability of aligning frequency bands, mode loss and the like, and is suitable for various remote sensing application scenes such as disaster monitoring, land coverage identification, military reconnaissance and the like.

The invention relates to the field of image data processing, and provides an unmanned aerial vehicle remote sensing image real-time seamless splicing method which comprises the following steps: preprocessing an unmanned aerial vehicle remote sensing image to obtain a geometric correction image; performing feature matching processing on the geometric correction image through a LightGlue deep learning model to obtain sub-pixel-level matching point pairs; performing image registration on the unmanned aerial vehicle remote sensing image according to the sub-pixel-level matching point pair to obtain a geometric registration image; performing splicing line search on the geometric registration image through an intelligent path-finding splicing line algorithm to obtain an optimal splicing path; and performing seamless splicing on the geometric registration image according to the optimal splicing path to obtain a seamless spliced image. According to the method, the overall splicing efficiency is improved, the unmanned aerial vehicle image can be processed in real time or near real time, and a more efficient splicing technology is provided for large-scale remote sensing application.

The invention relates to the field of remote sensing application, and particularly discloses a crop planting structure extraction method based on hierarchical extraction and multi-feature integration. The method comprises the steps of data acquisition and preprocessing, vegetation region extraction, classification model training and crop structure classification, and performs vegetation and non-vegetation division on a target region by adopting an OTSU algorithm and an NDVI hierarchical extraction method, so that the sample demand quantity is reduced, and the crop structure classification efficiency is improved. And the random forest classification model is trained by using the sample data, the spectral features, the vegetation features and the texture features, so that the precision and reliability of model classification are improved.

The invention discloses a knowledge graph semantic path guided remote sensing data recommendation method, which introduces a representation learning mechanism based on knowledge graph semantic path guidance, and realizes modeling of structured semantic association between nodes by constructing Meta-Path with task semantics. According to the method, a path-oriented random walk strategy is designed by combining a Meta-Path2Vec embedding method, semantic representation in an embedding space is obtained, and a sorting recommendation result is calculated and generated based on path similarity. According to the method, computable modeling of the semantic chain in the knowledge graph structure is realized, and path controllability, semantic interpretability and structural universality in the recommendation process are enhanced. According to the method, a complex semantic relationship among a natural disaster type, a remote sensing application task and remote sensing data can be modeled through meta-path constraint in a multi-source heterogeneous environment. In combination with a Meta-Path2Vec embedding method and a random walk strategy, high-order semantic association modeling between tasks and remote sensing resources is realized, and potential relationships between entities are effectively captured.

The invention discloses a multi-temporal optical remote sensing image robust registration method based on multi-scale joint similarity measurement, and the method comprises the steps: 1, inputting a dual-temporal image, and obtaining uniformly distributed control point pairs through SIFT extraction and interactive manual auditing; 2, constructing a global dense displacement field by using a thin-plate spline, completing coarse registration and partitioning according to an overlapping strategy; 3, performing three-stage coarse-to-fine template matching on each image block, inheriting an initial displacement value step by step, and realizing radiation difference and geometric deformation synchronous self-adaption by taking a product of normalized mutual information and a normalized cross correlation coefficient as joint similarity measurement; and 4, performing sub-pixel-level deformation correction on the reliable displacement field by using a local thin plate spline, and outputting a registration image and metadata. According to the method, a global sparse-local dense framework and a multi-scale progressive + joint measurement + quality constraint mechanism are coupled, and a high-robustness and high-precision registration solution is provided for quantitative remote sensing application such as change detection and disaster assessment.

The invention discloses a prompt tuning method during testing based on uncertainty perception and cache driving, and belongs to the technical field of computer vision. The method comprises the following steps: firstly, extracting visual features from an input remote sensing image, and retrieving a most relevant prompt vector as initialization in a prompt cache according to feature similarity; then performing model reasoning by utilizing initialization prompt, judging the uncertainty of the sample based on an entropy value of a prediction result, executing prompt optimization on a high-confidence sample, optimizing prompt parameters through entropy minimization, and skipping updating on a low-confidence sample; reasoning again to obtain a final classification result after prompt tuning is completed; meanwhile, the prompt cache is updated according to the sample confidence, and the representativeness and timeliness of prompts in the cache are kept. The method improves the adaptability and prediction stability of the model to the distribution change remote sensing image, and is suitable for the remote sensing application scene with lack of annotation and uncertain distribution.

This application relates to the fields of SAR signal processing and marine remote sensing applications, and provides a method and apparatus for sea surface current field inversion based on electromagnetic vortex wave radar data. The method determines the Doppler spectrum using electromagnetic vortex wave radar image data, obtains the measured frequency position of the trough feature point in the Doppler spectrum, and determines the theoretical frequency position using the position and angle of the two main lobes and the angle of the trough feature point in the radar antenna pattern. Then, the difference between the measured frequency position and the theoretical frequency position is used to obtain the anomalous frequency offset. Next, the Doppler frequency offset caused by wind and wave motion is removed from this anomalous frequency offset to obtain the target Doppler frequency offset. Finally, the radial velocity of the sea surface is calculated using the Doppler frequency offset, thus realizing sea surface current field inversion. This effectively expands the applicability of the Doppler frequency shift-based ocean current inversion method and improves the application potential of electromagnetic vortex wave radar in sea surface current field detection.

The invention discloses a boundary perception gating cross-scale feature fusion method for a remote sensing image, belongs to the technical field of computer vision and remote sensing image processing, and particularly relates to the field of remote sensing target detection. Aiming at the problems of insufficient cross-layer alignment, boundary confusion, insufficient small target recall and the like in the prior art, the invention provides a boundary perception gating cross-scale feature fusion method for a remote sensing image. The method comprises the following steps: constructing an Epag module containing Scharr edge prior and logit space fusion, and realizing dual-branch feature adaptive fusion; designing an SAFM module with channel separation and multi-scale pooling, and completing spatial adaptive enhancement; the two modules are integrated at the neck of a detection network, and multi-scale features are output through top-down gating fusion and spatial modulation. According to the method, the target boundary positioning precision and the multi-scale adaptation capability are remarkably improved, multiple technical pain points are solved, mAPs on DOTA1.0 and DIOR-R data sets are improved respectively, multiple detectors are adapted, and accurate technical support is provided for remote sensing application such as urban planning and traffic monitoring.

The present invention discloses a method and system for reconstructing optical images with high spatiotemporal resolution based on the fusion of multi-source remote sensing data, aiming to solve the problem of missing spatiotemporal information of optical remote sensing images due to cloud cover, sensor failure and the like in the prior art. The method is as follows: 1. Acquisition, preprocessing and data pair construction of multi-source remote sensing data. 2. Constructing and training an optical image reconstruction model, the model comprising a generator that receives multi-source data as input and a discriminator. 3. Reconstructing optical images of missing phases using the trained generator to generate a complete continuous optical image time series with high spatiotemporal resolution. The present invention can effectively fill in the gaps in optical image data, and the generated images have high spatial and spectral accuracy. It can effectively process data of different resolutions and reduce the influence of geometric registration errors through a multi-input network structure, providing reliable data guarantee for remote sensing applications that rely on continuous optical observations.

The invention provides a multi-modal large model quantitative fine tuning method and device for remote sensing application. The method comprises the following steps: quantizing a backbone network in a trained multi-modal large model to obtain a quantized backbone network; training a side chain network in the second network model based on a data set for remote sensing semantic segmentation and remote sensing target detection to obtain a trained second network model; testing performance indexes of the trained second network model based on a test set; when the performance index of the trained second network model meets the performance index of the unmanned aerial vehicle, stopping training the side chain network in the trained second network model to obtain a trained second network model; the second network model comprises a quantized backbone network and a quantized side chain network; the side chain network is a lightweight network; the trained second network model is used for being deployed on the unmanned aerial vehicle so as to perform semantic segmentation or target detection on the to-be-processed remote sensing image.

The application discloses a canopy coverage considering aggregation index inversion method and correction method, relates to the technical field of remote sensing image processing and aggregation index inversion, collects remote sensing images of a vegetation canopy, extracts hot spot reflectivity and dark spot reflectivity respectively, calculates a normalized hot spot and dark spot index NDHD, calculates an initial value of an aggregation index CI according to a linear relationship between the NDHD and the CI, and corrects the initial value of the CI by using a canopy coverage Ccan, wherein an error between the initial value of the CI and a real value of the CI has a linear relationship with the Ccan. The application first establishes a CI correction model based on the Ccan, and solves the deviation problem of a traditional algorithm in a sample field with high or low Ccan. By improving a mainstream CI inversion algorithm based on the NDHD, the application significantly improves the precision of CI inversion, and provides more reliable technical support for forest ecological monitoring, carbon cycle research and remote sensing application.

This invention relates to the field of remote sensing image processing technology, specifically disclosing a remote sensing image target detection method based on a multi-scale, cross-stage network model. The method includes the following steps: acquiring the remote sensing image to be tested; inputting the image into the backbone network after passing through an input layer, extracting multi-scale and multi-level semantic features step-by-step, and obtaining feature maps of different sizes through multi-scale downsampling; inputting the feature maps of different sizes into a feature integration network for cross-stage fusion and lightweight optimization to obtain multiple fused feature maps; and inputting the multiple fused feature maps into a prediction head for classification and detection to obtain detection results at multiple scales. This invention, by constructing a novel multi-scale, cross-stage network model, enhances the capabilities of multi-scale feature extraction, fusion, and target recognition, enabling accurate detection of multi-scale targets, especially small and complex-shaped targets, in remote sensing images. It balances detection accuracy and inference efficiency, making it suitable for remote sensing applications such as marine monitoring and urban planning.

The present invention relates to a method for calculating the water volume of a river channel compound section based on remote sensing data, and belongs to the field of river section water volume measurement and remote sensing application technology. The steps are to obtain basic survey data of the river channel compound section, obtain remote sensing images covering the required river channel section through satellite, pre-process and set the image resolution to R×R; extract the river channel water body by calculating the normalized water body index of the pre-processed image; according to the water body extraction result, calculate the number of pixels of the upstream and downstream river channel width and calculate the upstream and downstream river surface width; calculate the water depth based on the survey basic data of the upstream and downstream river channel compound section and the river surface width, and then calculate the river section water volume based on the water depth and volume formula. The present invention can not only effectively reduce the investment of manpower and material resources, but also realize the normalized monitoring of the river channel compound section and improve the efficiency of water resource management.

The invention provides a shallow sea underwater terrain inversion method and system based on a video SAR, and belongs to the synthetic aperture radar signal processing field and the ocean remote sensing application field, and the method comprises the steps: obtaining multi-frame SAR data with sea wave characteristics in a shallow sea area based on the video SAR; registering the obtained multi-frame SAR data with characteristics in the shallow sea area to obtain a video SAR image; every two frames of video SAR images are uniformly divided into sub-image blocks in an overlapped mode, and the wave lengths and the wave velocities of the sub-image blocks are estimated respectively; inverting water depths corresponding to all the sub-image blocks by using a linear dispersion relationship to obtain multiple groups of underwater terrain inversion results; and fusing multiple groups of underwater topography inversion results, and taking the result as a final underwater topography detection result of the research area. According to the invention, more refined water depth inversion can be realized.

This application provides a method, apparatus, equipment, and storage medium for identifying agricultural irrigation districts based on the irrigation-enhanced evaporation stress index (IESI), relating to the fields of agricultural water resources monitoring and management and remote sensing applications. The method includes: constructing a multi-source remote sensing and auxiliary dataset for the target area; calculating potential evapotranspiration and actual evapotranspiration and coupling them with a vegetation growth stability factor to obtain quarterly IESI data for the target area; determining irrigation identification thresholds applicable to plains and mountainous areas respectively; and discriminating the IESI of the target area to generate a binary spatial distribution map of irrigated and non-irrigated areas. This application solves the problems of existing irrigation area identification methods in large-scale applications, such as reliance on statistical data or large training samples, strong empirical thresholds, insufficient adaptability to terrain and regional differences, and complex calculation processes. It achieves irrigation behavior identification centered on evaporation stress response, ensuring computational efficiency while also considering spatial accuracy and regional adaptability.

The invention belongs to the technical field of remote sensing application, and relates to a knowledge migration type high-resolution remote sensing image building change detection method and system. The method comprises the following steps: constructing an adaptive knowledge-driven building change detection model containing multi-layer attention; obtaining a dual-time-phase high-resolution remote sensing image, and constructing a dual-time-phase high-resolution remote sensing image data set; training a building change detection model by using the double-time-phase high-resolution remote sensing image data set; storing solidification model parameters corresponding to each dual-time-phase high-resolution remote sensing image data set into a weight pool; and calling matched curing model parameters from the weight pool according to a data set scene of a to-be-detected remote sensing image, loading the parameters to the MMA-Net network, outputting a binary change graph, and completing building change detection. According to the method, a knowledge accumulation mode of adaptively modeling attention by using a weight pool is provided, model parameters obtained by training different data are stored to form the weight pool, and the weight pool is reused in a new building change detection task to realize migration application.

The application provides a change detection method and system based on feature interaction and space-time correlation, and relates to the technical field of remote sensing image change detection. The specific steps are as follows: first, a double-time change detection data set is obtained and preprocessed. Second, a change detection model based on a double-encoding double-decoding structure is constructed. The model includes a mixed effective channel module, which mainly enables the double-time features to maintain spatial information similarity and retain more effective channel information when interacting; a spatial denoising attention module, which mainly strengthens position information such as details and edges on the shallow layer features; and a feature fusion up-sampling module, which is mainly used for model training and data transmission in the up-sampling stage. Finally, the effect of the trained model is verified, and the result is saved for use. The application can effectively solve the challenges faced in remote sensing application tasks due to the limitations of remote sensing images themselves, such as fog, shadow, complex targets, etc., and improve the accuracy of the change detection task.

The application discloses a global average sea clutter near real-time estimation method, system, device and storage medium, including the following steps: obtaining multi-source satellite-borne radar historical backscattering coefficient data, and performing time and space matching and binning on the ERA5 wind field; an average sea clutter estimation model is established according to a geophysical model function fitting; the satellite-borne microwave scatterometer wind field is corrected in deviation and fused with a numerical weather prediction wind field, so that a high-coverage, continuous near real-time global sea surface wind field is obtained; the relative azimuth is calculated in combination with radar observation parameters and fused wind direction, and global average sea clutter data is obtained by using the model; the application improves the precision and application range of sea clutter estimation under the condition of medium and small incident angles, provides a high-quality input wind field through a multi-source wind field fusion technology, realizes rapid, stable and near real-time estimation of global average sea clutter, has high timeliness and business feasibility, and can meet the needs of near real-time sea clutter monitoring and ocean remote sensing application.

The application focuses on the field of agricultural informatization and remote sensing application technology, and particularly relates to a kind of alfalfa yield estimation method based on multispectral and physical constraint sample enhancement. The method first acquires unmanned aerial vehicle multispectral image data corresponding to the alfalfa growth period, and lays out field quadrats with equal area in the image coverage area, collects measured data of alfalfa hay yield, leaf area index, chlorophyll content and leaf equivalent water thickness, and constructs an initial sample set. Then, input each parameter data within a reasonable range, generate predicted multispectral samples that meet the physical consistency constraint using the radiation transfer model, to expand the sample space. The measured samples and predicted samples are fused to construct an alfalfa hay yield estimation training data set. Based on the training data set, a mapping model between unmanned aerial vehicle multispectral features and alfalfa hay yield is established, and the alfalfa hay yield is estimated through machine learning.

The invention provides a remote sensing recognition method, device, medium and system for rice planting area distribution, and belongs to the field of agricultural remote sensing, the remote sensing recognition method adopts a hierarchical classification algorithm for extraction, does not depend on actual measurement sample points, and obtains high-precision agricultural irrigation area rice spatial distribution rasterized data after automatic training; specifically, an agricultural area and a non-agricultural area are classified through a supervised learning algorithm classification regression tree (CART), a support vector machine (SVM) and a random forest (RF), and then a decision tree algorithm is adopted to further classify and discriminate a rice planting area based on obtained agricultural area classification raster data. The method does not depend on actual measurement data, classification time and economic cost are greatly reduced, and negative effects of the quality of a sample training set of a traditional machine algorithm on a classification result are effectively avoided; meanwhile, on the premise of ensuring high classification efficiency, the method meets the rice drawing precision required by agricultural remote sensing application, and is beneficial to popularization and application.