236 results about "Land cover" patented technology

The invention discloses an image multi-modal feature extraction method, a ground feature classification and recognition method and a GIS image generation method. Comprising the following steps: firstly, extracting texture features from a target image by adopting a multi-directional statistical method fusing rotation invariant coding of a local binary pattern and a gray-level co-occurrence matrix, and extracting color features from the target image by adopting an LAB-HSV dual-color space collaborative analysis method to obtain features of different modes of the target image; and then according to the information values of the extracted color features and texture features, adjusting the weight ratio corresponding to the color features and the texture features so as to optimize the recognition precision of the classification model on complex ground features. And finally, according to a weight ratio corresponding to the color feature and the texture feature, performing weighted fusion on the color feature and the texture feature to obtain a corresponding multi-modal feature vector.

The invention discloses a multi-temporal remote sensing image automatic change detection method based on deep learning, and the method comprises the following steps: collecting remote sensing image data obtained in different periods in a target region, carrying out the standardization processing of the remote sensing image data, and obtaining a standardized multi-temporal remote sensing image data set; based on the standardized multi-temporal remote sensing image data set, carrying out ground feature category segmentation by adopting a semantic segmentation network to obtain a ground feature semantic priori graph; inputting the standardized multi-temporal remote sensing image data set into the improved U-shaped twin network, and outputting a change detection probability graph and a binary change mask; based on the change detection probability graph and the binary change mask, attribute discrimination is carried out on the change area in combination with the ground feature semantic prior graph, and a change area result is output; and integrating the change area result to a geographic information service platform. The improved U-shaped twin network is adopted, and automatic change detection of the multi-temporal remote sensing image is achieved.

The disclosure provides an improved canopy height correction method. The method includes: obtaining GEDI LiDAR data, airborne canopy height data, GDEM with high resolution and land cover product within the selected target area and timeframe; performing quality filtering and spatial-scale filtering on GEDI footprints; extracting laser pointing parameters and waveform parameters; extracting reference canopy height from airborne data for each footprint; extracting laser pointing parameters and waveform parameters; preprocessing the GDEM and calculating topographic parameters, including topographic variability index (TVI); constructing the Laser Pointing and Topographic Index (LPTI) according to the 3D forest-ground geometry model; inputting the waveform parameters, even topographic parameters, TVI and LPTI as independent variables, and the reference canopy height as the dependent variable to modeling an improved forest canopy height extraction, and utilizing the improved canopy height extraction model to correct the twofold influence of topographic on GEDI canopy height extraction.

The invention discloses a forest degeneration degree, degeneration type and degeneration process identification method, and relates to the field of forest resource monitoring and ecological environment evaluation.The method comprises the steps that multi-source data obtained in a research area is preprocessed, and the preprocessed multi-source data is determined; selecting multi-dimensional indexes from three aspects of forest degradation structure, composition and function, and constructing a multi-dimensional index system according to the preprocessed multi-source data; the multi-dimensional indexes comprise forest coverage rate, crushing degree, tree variety diversity, aboveground biomass and net primary productivity; generating a forest degradation index according to each index in the multi-dimensional index system; determining a forest degeneration degree according to the forest degeneration index, generating a forest degeneration type graph in combination with land coverage data to reveal degeneration differences of different forest types, analyzing a dynamic change track of a forest degeneration region in combination with a normalized combustion index time sequence, and identifying a forest degeneration process; according to the method, the multi-dimensional characteristics of degradation can be comprehensively revealed.

The invention discloses a land coverage classification method based on inference segmentation, and relates to the technical field of remote sensing image inference segmentation and deep learning. Comprising the steps of training sample set establishment, multi-scale feature extraction submodule design, cross-modal feature fusion module design, land coverage classification model structure design based on reasoning segmentation, land coverage classification model training based on reasoning segmentation, model performance evaluation and index analysis. According to the method, effective combination of remote sensing image features and semantic information is realized, and the differentiated cognitive ability for different ground object targets is improved. Meanwhile, a remote sensing image land cover intelligent classification technology with practical value is obtained, the professional technical threshold is remarkably reduced, the intelligence and universality of remote sensing image interpretation are realized, and a convenient remote sensing information acquisition method is provided for users in various fields.

The invention relates to an edge-guided cross-modal interactive fusion land cover classification method, which is particularly suitable for collaborative semantic segmentation of an optical image and a synthetic aperture radar (SAR) image. The method comprises the following steps: constructing a pseudo twin multi-stream encoder to respectively extract multi-scale features of optical and SAR images, embedding a cross-modal adaptive feature interaction module between encoding stages, realizing dynamic alignment and re-calibration of features between modals through a channel-level and space-level interaction mechanism, and relieving fusion deviation caused by heterogeneity and spatial dislocation. Neuron-level attention and multi-scale depth separable convolution are further fused through a lightweight feature fusion module, noise is suppressed, and context information is aggregated. And meanwhile, an edge auxiliary module is introduced to extract multi-scale boundary information from low-level features, and the multi-scale boundary information and semantic features are jointly decoded, so that the boundary detail and small target recognition capability is improved. According to the method, the precision and robustness of land coverage segmentation in a complex scene are remarkably improved on the premise of ensuring the calculation efficiency.

The invention relates to the technical field of geographic information measurement, in particular to a landform surveying method and system for territorial space planning, and the method comprises the steps: 1, carrying out the fusion preprocessing of multi-source data; 2, constructing a planning constraint dynamic rule base; 3, space-time constraint coupling analysis: taking the digital elevation model as a reference, and correcting the space offset of the land cover map through a thin-plate spline function; comparing the planning threshold value with the current landform parameter in real time, and marking a conflict coordinate set of gradient overrun or settlement overspeed; 4, intelligently segmenting the landform units; and 5, planning suitability output: setting a dynamic buffer distance according to the fault activity grade and the rock-soil body type, and generating an early warning layer when a town unit exists in a buffer area. According to the method, remote sensing images, laser radar point clouds and geological survey data are fused, space-time dislocation among various data sources is eliminated, and the precision of space analysis is improved.

The invention provides a coastal zone land coverage type classification method based on a random forest and a topology rule, and belongs to the technical field of coastal zone land coverage type classification. The method comprises the following steps: based on a GEE cloud platform, integrating a multispectral remote sensing image and elevation data, extracting spectral features and geographic features, and realizing pixel-level classification of random forest machine learning; generating a clustering unit through a simple non-iterative clustering method, and performing object-oriented special classification on the water body and the aquaculture pond in combination with spectral and morphological characteristics of the clustering unit; spatial filtering is introduced to remove salt and pepper noise, a coastal zone ground feature topology rule base is established to carry out spatial logic verification, and coastal zone land coverage classification with high precision and reasonable spatial layout is realized. According to the method, the multi-source remote sensing data and the multi-level classification strategy are fused, the problems of coastal zone ground feature spectrum confusion, boundary fuzziness and space logic inconformity are effectively solved, the classification precision and practicability are remarkably improved, and the method is suitable for land coverage monitoring and management of the complex coastal zone environment.

The invention discloses a farmland protection forest area windproof effect evaluation method based on a land-gas coupling model, and relates to the technical field of agricultural science, high-precision topographic data of a target farmland protection forest area is collected and preprocessed, and spatial distribution information of a farmland protection forest is obtained by combining remote sensing image interpretation; on the basis of topographic data and spatial distribution information, topographic relief and earth surface coverage features are analyzed, and a refined underlying surface classification system is constructed; and configuring high-resolution grid parameters of a WRF mode, and coupling a Noah-MP land surface process model. According to the farmland protection forest area wind-proof effect evaluation method based on the land-gas coupling model, through combination of high-resolution topographic data and a multi-source remote sensing image, topographic relief and surface heterogeneity of a farmland protection forest area can be accurately depicted, and a WRF mode is coupled with a Noah-MP land surface process model; fine simulation of complex terrains and various underlying surfaces is realized, the limitation of traditional uniform underlying surface hypothesis is overcome, and the accuracy of windproof effect evaluation is remarkably improved.

The invention belongs to the technical field of remote sensing image processing, computer vision and deep learning, and particularly relates to a remote sensing image erosion gully semantic segmentation method based on improved OfficientNet-UNet. Comprising the following steps of 1, data preparation and data preprocessing; 2, constructing and enhancing a data set; step 3, model construction and strategy training; and 4, performing contrast experiment and result evaluation. According to the method, detail features of ground features can be more accurately captured, an overfitting phenomenon caused by too high model complexity is reduced, so that the classification precision and boundary recognition accuracy of land coverage data are effectively improved, weights of different types of samples can be automatically adjusted in the training process, and the training efficiency is improved. Particularly, the contribution of background pixels to a loss function is reduced, so that the problem of dominant training of the background pixels is effectively relieved; the method has good expansibility.

The invention provides a global biodiversity hot spot area threat condition remote sensing monitoring method and system, and relates to the technical field of ecological environment monitoring, spatial analysis and remote sensing and information processing, and the method comprises the steps: obtaining multi-temporal global land coverage data and auxiliary GIS data of a biodiversity hot spot area, and carrying out the preprocessing; extracting forest coverage layers of a reference period and a monitoring period from global land coverage data, and extracting pixels converted from a forest to a non-forest through a layer difference method to generate a forest loss layer; based on the forest loss map layer and the preprocessed auxiliary GIS data, calculating various ecological pressure indexes including a forest loss rate, a protection failure rate and a population pressure index, and performing normalization; and calculating a comprehensive threat index based on each ecological pressure index, and determining an ecological threat level. According to the method, automatic identification, quantitative evaluation and grading of the ecological threats of the hot spot region are realized, and technical support is provided for global ecological protection planning and management.

The invention relates to a remote sensing interpretation method and system integrating multi-source spatio-temporal spectrum characteristics and a visual model. The method comprises the following steps: firstly, acquiring and preprocessing an optical image, harmonic and synthetic aperture radar data of a target area; time spectrum features of the optical image and the harmonic data are extracted respectively, and fusion features are generated through cross attention mechanism fusion; carrying out image serialization, time coding and space coding processing on the fused features and synthetic aperture radar data, and extracting spatio-temporal features by using a self-attention mechanism; and finally, decoding the spatio-temporal features and fusing a multi-source feature convolution result to generate a pixel-level land coverage classification result. According to the method, harmonic data are introduced to capture the time change trend of the ground features, a visual Transform self-attention mechanism is utilized to capture the space-time interdependence relationship, multi-source data are combined to capture the multi-dimensional features of the ground features, the limitation of a single data source is avoided, the multi-source time sequence features are fused to enable the model to adapt to different geographical environments, and the generalization performance is improved.

The invention provides a method and system for determining appropriate granularity for landscape pattern analysis, and the method comprises the steps: enabling high-resolution land cover or land utilization vector data of a target region to generate a first multi-granularity sequence with a coarse step length according to a maximum area value rule (Rule of Maximum Area, RMA), and determining a granularity upper threshold value and a scale domain according to three conservation criteria of landscape composition, area and pattern form; a second multi-granularity sequence in the scale domain is generated with a fine step length according to the RMA rule, and a landscape pattern index, sensitive to the scale effect, of the second multi-granularity sequence is calculated; the method comprises the following steps: performing mutation analysis by adopting a non-parametric rank test (Mann-Kendall and Pettitt test), and taking an effective mutation point with the highest frequency as an appropriate granularity; according to the method, information loss and nonparametric mutation inspection are quantified through three conservation criteria, the problems that a traditional method neglects landscape pattern form loss, granularity judgment is high in subjectivity, data distribution depends and the like are solved, and the objectivity of granularity decision and ecological application precision are improved.

The invention discloses a habitat quality influence assessment method based on an InVEST model. The habitat quality influence assessment method comprises the steps of S1, acquiring land cover classification data and threat factor data of an assessment area; s2, on the basis of the land cover classification data and the threat factor data, a habitat quality spatial distribution diagram of the evaluation area is calculated by using a habitat quality module in an InVEST model; s3, based on the habitat quality spatial distribution map, combining selected driving factor data, and utilizing a geographic detector model to quantitatively analyze the interpretation force of each driving factor on the habitat quality spatial diversity; s4, performing spatial overlay analysis based on the habitat quality spatial distribution map and ecological protection red line range data of the evaluation area; the method has the beneficial effects that the independent interpretation force of multiple driving factors such as natural geography and human activities on the habitat quality spatial diversity is quantitatively revealed by introducing the geography detector model, and the interaction effect between the factors can be further detected.

The invention discloses a space remote sensing image analysis method for detecting surface water and underground water exchange in a river region based on surface temperature, and belongs to the field of remote sensing detection surface temperature anomaly analysis. According to the method, background field construction is carried out on pixels in a space by combining single-scene high-resolution surface temperature data and land cover classification data, and temperature thermal anomaly of a river main stream is extracted by adopting an adjacent pixel temperature anomaly analysis method, namely, the temperature thermal anomaly of the river main stream is extracted by analyzing subtle change of surface temperature in a remote sensing image; and identifying abnormal pixels which are obviously different from the temperature characteristics of the surrounding area. The abnormal pixels often correspond to areas where surface water and underground water exchange is active, such as permeable zones, spring water exposure points and the like. The method can be applied to drainage basin water resource management and ecological protection, and the conversion path and intensity between surface water and underground water can be deduced by further analyzing and processing the abnormal pixels extracted by the method and combining background information such as geology and hydrology.

Example methods and systems for geological interpretation in hydrocarbon exploration are disclosed. One example method includes obtaining a geological sketch of a first region, where the geological sketch includes a color-labeled image of the first region, and each color in the geological sketch represents a respective type of geological land cover in the first region. A first machine learning model is applied to convert the geological sketch into a first virtual satellite image of the first region. The first virtual satellite image is provided for geological interpretation of the first region.

The invention relates to a global land comprehensive improvement monitoring method based on artificial intelligence and remote sensing images, and belongs to the technical field of environment management. According to the method, multi-band high-resolution remote sensing images of a target area in two time phases before and after improvement are obtained for radiation and atmospheric correction to ensure data quality, and land coverage classification maps of the two time phases are generated respectively by using a pre-trained deep convolutional neural network land coverage classification model. Through pixel-level change detection and confidence analysis, a reliable land coverage change area is identified, vectorization and spatial filtering are carried out, multi-dimensional change features are extracted, ecological influence and a spatial distribution mode of the multi-dimensional change features are comprehensively evaluated, and a comprehensive monitoring result including a spatial distribution diagram, a statistical report and trend prediction is generated. And the information is integrated to an interactive geographic information system platform for visual display and early warning. According to the invention, automatic, precise and systematic monitoring and evaluation of the global land comprehensive remediation process are realized.

The invention provides a water environment non-point source pollution recognition method and device based on hyperspectral scanning, and belongs to the field of non-point source pollution recognized.The method comprises the steps that a target point location is obtained, and the target point location is the point location with the non-point source characteristic pollution index concentration exceeding the preset quality control concentration; the non-point source pollution confidence coefficient of the target point location is determined according to the hyperspectral information of the target point location, the multiple ground feature types and the weights corresponding to the ground feature types, the weight corresponding to each ground feature type is in positive correlation with the distance between the ground feature type and the target point location, and the multiple ground feature types include the agricultural land type and the non-agricultural land type; and if the non-point source pollution confidence exceeds the preset threshold, calculating the change rate of the water quality parameter of the target point location, and identifying whether the non-point source pollution occurs at the target point location according to the change rate. According to the water environment non-point source pollution identification method and device based on hyperspectral scanning, the non-point source pollution hot spot area can be automatically identified based on the hyperspectral scanning technology in combination with the ground feature type.

The invention discloses a remote sensing image ground object segmentation method based on a de-noising diffusion probability model, and relates to the technical field of remote sensing image processing, and the method comprises the steps: carrying out the grid division and interference feature recognition of a target remote sensing image, and carrying out the grid confidence evaluation based on a recognition result; a plurality of image sub-regions are obtained through fitting according to the plurality of interference feature types and the plurality of grid confidence coefficients; calling a sample denoising diffusion probability model library, carrying out sample model matching and integration according to the regional position environment feature and high-frequency interference feature types and the regional image confidence, and constructing a plurality of adaptive image denoising plug-ins; respectively carrying out image denoising on the plurality of image sub-regions, obtaining a plurality of denoised region images, and carrying out image splicing to obtain a denoised remote sensing image; and carrying out ground feature segmentation on the denoised remote sensing image by using a convolutional neural network, and outputting a ground feature classification map. According to the invention, the technical problem of inaccurate ground feature segmentation result caused by poor denoising effect of the remote sensing image in the prior art is solved.

The invention discloses a colored shaded relief map generation method and system fusing land coverage information. The method comprises the steps of obtaining DEM data and land coverage data of a target area, taking the preprocessed DEM data as DEM condition information, and taking the preprocessed land coverage data as land coverage condition information; the method comprises the following steps: inputting a randomly generated noise image and DEM condition information into a backbone network, inputting the noise image, the DEM condition information and land coverage condition information into a ControlNet fine tuning module to obtain a correction signal, inputting the correction signal into the backbone network to obtain predicted noise, and performing reverse denoising based on the predicted noise to obtain a colored shaded relief map. According to the method, the semantic correctness of the generated color can be ensured, and the generated colored shaded relief image better conforms to the visual effect of a real natural landscape.

The invention provides a hyperspectral and laser radar fusion classification method (GGCDM) based on gating guide condition diffusion, which is used for fusing hyperspectral image (HSI) and laser radar (LiDAR) data to realize high-precision ground feature classification. The existing method is difficult to consider deep coupling of high-dimensional spectral features and three-dimensional geometric information under the problems of insufficient multi-modal feature interaction and limited generalization ability. According to the method, by designing a gating condition modulator (GCM), HSI and LiDAR features are mapped to a unified potential space, and an interactive perception gating structure is introduced to realize dynamic weight adjustment of two modal features, so that the cross-modal cooperative characterization capability is enhanced. Meanwhile, a deep interaction enhancement module (DIEM) is embedded in a diffusion reconstruction network, cross-modal association in a potential space is explicitly modeled, and the stability and robustness of a classification result are improved. Different from a traditional diffusion model based on noise estimation, the method adopts an image reconstruction normal form, takes a classification graph as a generation target, avoids training instability, and remarkably improves generalization performance. Experimental results show that the classification precision of the method is superior to that of an existing method on multiple groups of real data sets. The method can be widely applied to the fields of remote sensing image intelligent interpretation, land cover classification, environment monitoring and the like.

This invention belongs to the field of soil environmental monitoring and discloses a three-dimensional prediction method for organic matter in soda saline-alkali soils. The method includes the following steps: First, multi-source remote sensing, environmental, and soil organic matter measurement data are collected, and a unified coordinate resolution is established to create a grid. Then, land cover features are extracted using UAV imagery and DSM (Digital Signal Processing), and land cover classification is completed by combining shadow restoration. Sub-pixel endmember abundance is calculated using satellite pixel aggregation data. A three-dimensional covariate dataset is constructed by fusing various features with soil depth. Finally, an ExtraTrees model is built to predict the three-dimensional distribution of organic matter layer by layer. This invention integrates multi-source remote sensing data to build a complete technical system, which can accurately characterize differences in surface composition, simultaneously represent the spatial distribution and vertical variation of organic matter, and effectively solve problems such as mixed pixel interference, poor data coordination, and insufficient three-dimensional prediction. It has strong practicality and good application prospects.

The application discloses a method for determining optimal positions of synchronous differential base stations for aerial remote sensing flight, and belongs to the technical field of aerial remote sensing. The method comprises the following steps: performing spatial analysis on vector data of a working area, quantitatively calculating distance average distance and maximum distance parameters, calculating average distances of each position from the working area within a given maximum base station distance threshold range by using python, outputting results, performing spatial analysis on the evaluation distance results and road data, extracting a range with a certain threshold distance from the road, and combining surface land cover classification data and road vector data according to base station erection requirements of different scales, surface cover, water area and forest area which are not suitable for base station erection, open source DEM data, and the optimal base station erection position or the nearest position of the base station erection is circled, so as to guarantee the differential processing precision of aerial remote sensing POS data.

The invention discloses a coastal salt marsh wetland non-photosynthetic vegetation coverage inversion method, which comprises the following steps: S1, obtaining hyperspectral reflection curves of soil, non-photosynthetic vegetation and photosynthetic vegetation of the coastal salt marsh wetland of the Yellow River Delta, simulating and measuring spectral characteristics of samples under different moisture contents (dry, low humidity, medium humidity and saturation) in a laboratory, and calculating the coverage of the non-photosynthetic vegetation in the coastal salt marsh wetland; key wave bands of non-photosynthetic vegetation and other ground features are screened and distinguished; s2, constructing a moisture-insensitive hyperspectral NPV index (MINI), and determining an optimal wave band combination through wave band traversal; and S3, constructing mixed data of soil, non-photosynthetic vegetation and photosynthetic vegetation under different moisture conditions in a laboratory, and evaluating inversion precision and stability of MINI under different moisture conditions. S4, acquiring a coastal salt marsh wetland hyperspectral image, and calculating an NDVI value and an MINI index value based on the processed image; and S5, performing spectral unmixing on the mixed pixels based on a triangular space method, and inverting the non-photosynthetic vegetation, photosynthetic vegetation and soil coverage of the research area. The method provided by the invention solves the technical problems of low precision and poor stability of the traditional NPV index inversion coverage under the condition that the coastal salt marsh wetland is influenced by tides and the moisture change is obvious.

This invention relates to a remote sensing interpretation method and system that integrates multi-source spatiotemporal spectral features and a visual model. First, optical imagery, harmonic data, and synthetic aperture radar (SAR) data of the target area are acquired and preprocessed. Next, spatiotemporal features of the optical imagery and harmonic data are extracted separately and fused using a cross-attention mechanism to generate fused features. Then, the fused features and SAR data are subjected to image serialization, temporal encoding, and spatial encoding processing, and spatiotemporal features are extracted using a self-attention mechanism. Finally, the spatiotemporal features are decoded and the multi-source feature convolution results are fused to generate pixel-level land cover classification results. This invention introduces harmonic data to capture the temporal variation trend of ground features, utilizes the visual Transformer self-attention mechanism to capture spatiotemporal interdependencies, and combines multi-source data to capture multi-dimensional features of ground features, avoiding the limitations of a single data source. Furthermore, the fusion of multi-source temporal features allows the model to adapt to different geographical environments, improving generalization performance.

The invention relates to the technical field of earth space information, in particular to a super-resolution land coverage mapping method coupled with a large language model knowledge base. Comprising the following steps: acquiring a remote sensing image and geoscience text data of a target area and constructing a geoscience knowledge system; mining geoscience knowledge from the geoscience text data based on a large language model, constructing a structured geoscience knowledge base, and further generating a geoscience knowledge graph; obtaining a deep feature graph of the remote sensing image based on a convolutional neural network, and inputting a knowledge graph into a graph attention network to extract a regional geoscience semantic embedding vector; realizing geoscience semantic embedding vector and deep feature map fusion based on a global-local attention fusion mechanism; and performing up-sampling on the fused feature map to obtain a high-resolution depth feature map, and generating a super-resolution land coverage map of the remote sensing image by adopting a nonlinear activation function. According to the method, dynamic modeling and computable expression of regional global knowledge can be realized, and the problems of surface feature boundary fuzziness and category confusion are remarkably improved.

The invention relates to the technical field of electronic maps, in particular to a global earth surface coverage mapping method based on a 2-meter-resolution domestic satellite image, and aims at solving the problem that an existing 30-meter and 10-meter-resolution global earth surface coverage product is low in accuracy. And a mapping strategy combining general and regional feature modeling with layered land surface coverage is adopted. According to the image prepared by the method, the land cover accuracy is high.

This invention discloses a visual reconstruction method and system for remote sensing interpretation based on a generative diffusion model. The method includes: acquiring high-resolution and low-resolution remote sensing image data; adding different levels of Gaussian noise to the training data using a forward stochastic differential equation until pure Gaussian noise data is obtained; training a noise conditional scoring network to predict the scores corresponding to these noisy data; adding noise to the low-resolution image using a forward stochastic differential equation to finally obtain pure Gaussian noise; using a trained neural network to guide the random noise to gradually converge and generate a super-resolution remote sensing image; rapidly identifying land cover types on the generated remote sensing image; and delineating land cover patches on the original remote sensing image and assigning patch information based on the identified land cover categories. This invention achieves a super-resolution effect from low resolution without changing the land cover types and patch boundaries, thereby reducing interpretation costs and improving interpretation efficiency.

This invention relates to the fields of data processing and remote sensing information processing technology, specifically to a remote sensing carbon source and sink monitoring and assessment method based on a multimodal hybrid expert model. The method includes: feature extraction from multi-source remote sensing data (hyperspectral, visible, and infrared); multimodal fusion of the extracted features; construction of a large-scale multimodal hybrid expert carbon source and sink model using a hybrid expert partitioning sparsity processing mechanism to detect and segment carbon source and sink changes, obtaining a carbon source and sink distribution map; assessment of annual carbon emissions / absorption rates for different types of land cover using a combination of source and sink distribution and ground measurement data; and calculation of regional carbon emissions and absorption based on land cover type area and annual carbon emissions / absorption rates to obtain carbon flux assessment results. This invention can be applied to carbon metering, regional carbon emission inventory compilation, and carbon neutrality effectiveness assessment, providing technical support for achieving the "dual carbon" goal and improving the effectiveness of remote sensing carbon source and sink monitoring and assessment.

The invention discloses a construction method and a detection method of an artifact detection model of an optical remote sensing image restoration result, relates to the technical field of aerospace remote sensing and image processing, and solves the problems of single feature, weak generalization and poor detection accuracy of an artifact detection technology of an existing remote sensing image restoration result. And the method is difficult to adapt to artifact differences under different restoration algorithms and ground feature types. An artifact detection model construction method comprises the steps that a training set is obtained, and the training set comprises original images of a plurality of ground feature types and corresponding restored images; carrying out artifact region labeling on the restored image, and taking a labeling result as an artifact mask; and taking the artifact mask as a supervision signal, training the improved P2V network based on the training set, and obtaining the trained improved P2V network as an artifact detection model. The method is suitable for the fields of satellite optical remote sensing data processing, image restoration quality control, remote sensing image intelligent analysis, product quality detection and the like.