High-spatial-resolution remote sensing imagery provides a data foundation for fine-grained land use classification. However, due to long revisit cycles and susceptibility to cloud cover, large-area imagery often suffers from temporal inconsistency, which severely limits the classification accuracy of traditional unified models. To address this issue, this study proposes a geographic entity-oriented, spatiotemporally coupled land use classification method for high-resolution remote sensing imagery, with agricultural land (including paddy fields, dry farmland and gardens) as an example for validation. In this method, the study area is first divided into multiple sub-regions based on image acquisition time, ensuring temporal consistency within each sub-region. A dedicated deep texture feature extraction model is then constructed for each sub-region. This model is adapted from the advanced CAPTN texture recognition network: its classification head is removed, and a multi-scale feature fusion module is introduced, transforming it into an encoder focused on extracting spatial texture feature maps. Additionally, a self-supervised loss function combining masked feature reconstruction and cross-view consistency is designed to improve the quality of the learned texture features. During the prediction stage, the corresponding feature extractor is invoked based on the temporal phase of the imagery to generate a full-region texture feature map. This feature map is then cropped using land parcel vectors, and statistical feature vectors describing the texture attributes of each parcel are formed by calculating the mean and standard deviation of the features within each parcel. Finally, a Random Forest classifier is employed to determine the land parcel categories. This study uses the Jiangjin District of Chongqing City as the experimental area. The results show that, compared to training a unified deep learning model directly on full-region multi-temporal imagery or using traditional texture features, the proposed spatiotemporally coupled classification framework achieves significant improvements in overall accuracy and Kappa coefficient, reaching 92.3% and 0.89, respectively.

High-accuracy parcel-level agricultural mapping is fundamental to precision agriculture. However, in fragmented agricultural regions of the Yangtze River Delta, identifying cropping patterns at the parcel level faces two compounding challenges: asynchronous multi-source observations and mixed-pixel effects in small parcels. When historical archive records are used as training labels, inter-annual cropping pattern changes further introduce label noise that undermines model reliability. To address these challenges and the label noise issue, we propose PAST (Parcel-level Asynchronous SpatioTemporal), a parcel-level cropping pattern classification framework comprising three stages: K-Shape-based label quality control, parallel dual-branch classification, and decision-level fusion. PAST employs a dual-branch architecture: the temporal branch achieves interpolation-free cross-modal phenological fusion of Sentinel-1 and Sentinel-2 data, while the image branch extracts canopy texture features from 0.8 m high-resolution imagery to partially address mixed-pixel interference. Experiments in a typical fragmented agricultural region of the Yangtze River Delta demonstrate that PAST achieves an overall F1 score of 0.926 and a small-parcel F1 score of 0.906, outperforming mainstream time-series baselines. These results confirm that combining K-Shape label quality control at the data level with a dual-branch interference-robust architecture at the model level provides a complete integrated three-stage pipeline for fine-grained crop mapping under weakly supervised historical archive label conditions.

To address the limitations of hyperspectral imaging systems, super-resolution (SR) techniques that fuse low-resolution hyperspectral image (HSI) with high-resolution multispectral image (MSI) are applied. Due to the significant modal difference between HSI and MSI, and the insufficient consideration of HSI band correlation in previous works, issues arise such as spectral distortions and loss of fine texture and boundaries. In this article, an unsupervised spectral correlation-based fusion network (SCFN) is proposed to address the above challenges. A new dense spectral convolution module (DSCM) is proposed to capture the intrinsic similarity dependence between spectral bands in HSI to effectively extract spectral domain features and mitigate spectral aberrations. To preserve the rich texture details in MSI, a global-local aware block (GAB) is designed for joint global contextual information and emphasize critical regions. To address the cross-modal disparity problem, new joint losses are constructed to improve the preservation of high-frequency information during image reconstruction and effectively minimize spectral disparity for more precise and accurate image reconstruction. The experimental results on three hyperspectral remote sensing datasets demonstrate that SCFN outperforms other methods in both qualitative and quantitative comparisons. Results from the fusion of real hyperspectral and multispectral remote sensing data further confirm the applicability and effectiveness of the proposed network.

For large-scale mapping applications, cross-domain hyperspectral image classification (HSIC) has emerged as a highly promising research area. However, the classification accuracy decreased significantly when unseen classes emerged. Few shot learning (FSL) methods are adopted in cross-domain HSIC methods to address this problem. Despite this, existing cross-domain HSIC methods still have three key issues that hamper their classification capabilities: 1) previous works struggle to balance incorporating distinctive intradomain knowledge and managing model complexity in the face of significant domain representation differences; 2) previous works inadequately consider the limited capture capacity of interdomain intrinsic mutually invariant structures; and 3) previous works fail to capture the distinct characteristics of both head categories (e.g., urban buildings) and tail categories (e.g., urban corn) simultaneously when applying FSL to deal with unseen classes problem. In this article, we propose a cycle-resemblance few-shot transformation (CF-Trans) network to effectively handle the aforementioned challenges by integrating intradomain distinctiveness with interdomain invariance. To facilitate efficient intradomain feature aggregation for HSI, a novel lightweight intradomain attentive network is introduced. Different from previous works, to reduce the negative impact caused by inaccurate classifier predictions, from the perspective of interdomain knowledge transformation, a cycle-resemblance adversarial network is designed to capture the intrinsic mutually invariant structures. A dynamic label expansion mechanism is designed to capture the distinctive intradomain features of the head and tail classes. Experimental results on six HSI datasets including agricultural, rural-urban and urban datasets show the remarkably performance of our network.

Rice is a crucial global food crop,and the accurate,timely delineation of rice cultivation areas is vital for food security assessment and sustainable agricultural planning.Satellite remote sensing,integrated with advanced information technologies,has emerged as a pivotal tool for mapping the spatial distribution patterns and monitoring the growth dynamics of rice.However,in the hilly and mountainous regions of southwest China,frequent cloud cover,fragmented cultivation patterns,and high field sampling costs hinder efficient and reliable remote sensing-based rice mapping.To address these limitations,this study,focusing on Tongnan District in Chongqing Municipality,proposes a novel cropland parcel-scale rice automated sample generation and mapping framework,leveraging the complementary strengths of multisource remote sensing data.The framework operates through several key stages:First,with cropland parcels as the basic unit,an optical-SAR time-series feature set is constructed by utilizing multiperiod Sentinel-1 and Sentinel-2 data.Second,adaptive identification of rice growth stages is achieved by analyzing temporal VH polarization features from SAR data,which are sensitive to vegetation structure and moisture variations.Third,high-quality parcel-scale training samples are automatically generated by analyzing the rice seasonal pattern across different imaging modes,bypassing labor-intensive manual sampling.Given the above samples,automated rice mapping is implemented by combining feature optimization and random forest.Finally,we evaluate the reliability of the generated samples and the rice mapping results.Results indicate four critical findings:(1)The method generates spectrally representative samples,achieving high consistency with field data(spectral correlation similarity=0.987;dynamic time warping distance=4.719).These samples exhibit spatial heterogeneity and broad coverage,addressing inefficiencies and biases inherent in traditional manual sampling.(2)Feature selection can effectively reduce model complexity while ensuring classification accuracy.The contribution of transplantation period features is substantial,and the importance of SAR features is higher than that of optical features.The two types of data have complementary advantages and can synergistically improve the classification robustness under cloudy conditions.(3)The automated rice mapping achieves an overall accuracy of 89％,an F1 score of 89％,and a total area extraction error of-7.5％,validating its reliability in spatially heterogeneous landscapes.In hilly mountainous areas with significant spatial heterogeneity,a moderate number of samples with uniform spatial distribution can help improve the stability and generalization ability of the model.In addition,rice exhibits significant clustering in flat,well-irrigated areas.(4)The overall uncertainty of rice mapping in the study area is low(0.29),indicating that the mapping results are reliable.Uncertainty is mainly affected by the combination of environmental and agricultural management conditions,parcel morphology,and sample characteristics.In areas with flat topography,favorable irrigation conditions,regular parcels,and sufficient samples,the spectral characteristics of rice are significant,and the uncertainty of mapping is low.This study provides a reliable automated sample generation approach for rapid and precise rice mapping in hilly and mountainous areas and offers scientific foundation for developing sampling strategies,selecting optimal feature bands,and assessing uncertainty.Hence,it lays a credible basis for precision agriculture planning and food security assessment.

Using satellite image time series (SITS) and a deep learning model to classify crops in planting parcel has become the frontier technology of precision agriculture. However, the technology's reliance on training labels remains a major challenge for widespread application at this stage. We propose a novel Instance-Temporal-Context Contrastive Semi Supervised Learning method called ITCC-SSL, to learn time-series representation from unlabeled data. The core of this method is hierarchical grouped contrastive learning and confidence threshold filtering of pseudo labels. Specifically, we first perform temporal augmentation on unlabeled data, generating scaling and masking time series. These augmented time series, along with the original time series, form two contrastive groups. We then use instance, temporal, and contextual contrastive learning separately for each group. The encoder is then fine-tuned using the limited labeled data for contrasting learning and filters for high-quality pseudo labels based on a confidence threshold. Finally, we perform semi-supervised learning using both the limited labeled data and the pseudo-labeled data to obtain the final crop classification results. We use BreizhCrops dataset and conducted experiments on ITCC-SSL. The experimental results show that ITCSSL performs better than the mainstream semi-supervised learning methods on 100 to 10,000 labeled datasets of different sizes. The classification accuracy reached 60.33% and 77.51% with 100 and 10,000 labeled samples, respectively. Therefore, we can effectively alleviate the dependence of training labels using hierarchical grouped contrastive learning and confidence threshold filtering of pseudo labels and provide a new way of crop classification using SITS with limited labels.

Automatic road extraction has gained significant attention in urban navigation, sustainable transport, and disaster response. Conventional convolutional neural networks (CNNs) operate within the local receptive field, limiting their capacity to represent potential global relations between roads and surroundings. In addition, the edge is important topological information for road targets. Several works focus on predicting precise boundaries to enhance road extraction. However, over fit edges and the course integration between features of different network layers may lead to loss of local details and incorrect road segmentation results. Therefore, the Road-detail Preserving Mapper (RoadDP-Mapper) framework is proposed. First, RoadDP-Mapper employs a hierarchical transformer as the encoder to enable local-to-global reasoning. The asymmetric upsampling layers (APLs) are introduced to enhance the model's capability to perceive and reconstruct critical road detail information. Second, a road edge-constrained branch with a detail preservation module (DPM) is devised to amplify the distinction between roads and backgrounds by extracting and preserving explicit class boundary details. The proposed joint loss inspires the transformer to capture the contextual spatial relationships while preserving the fine-grained features of the road. We evaluated our framework on the DeepGlobe dataset and self-annotated images from ten representative cities in China. The proposed framework has demonstrated its effectiveness by significantly reducing both missed detections and false alarms in road extraction. Furthermore, spatial transfer experiments have confirmed the generalizability of RoadDP-Mapper for large-scale road mapping.

The precise mapping of crop spatial distribution using remote sensing datasets is a fundamental task in precision agriculture, yet conventional methods based on pixels or segmented objects often neglect the boundary constraints from the cultivated areas and correlations between analysis units. In response, the proposed work is a novel attempt to address parcel-wise crop classification with multilevel consistency constraints, generating robust features and consequently more precise results. Pixel-level constraints enforce spectral homogeneity within parcels extracted by the neural network during feature construction. Each parcel is then correlated with its closest neighbors considering comprehensively the spatial, environmental, and temporal similarities, providing consistency information at the parcel scale. Two model structures based primarily on the graph neural network and attention mechanism were proposed for crop identification. Results show that both constraints may yield significant benefits given their enhancement to feature stability and completeness, especially in small-sample cases. The neighborhood models achieved average accuracy improvements of 3.10%, 3.47%, 4.65%, 4.34%, and 15.78%, respectively, under varying training set proportions of 0.5, 0.4, 0.3, 0.2, and 0.1, and an OA of 89.50% could still be maintained for the graph-based one even when the training set ratio was limited to 0.02. Within the calculation, various intelligent algorithms such as convolutional neural network, random forest, graph neural network (GNN), and attention mechanisms are methodically leveraged for effective pattern recognition of both spatial, sequential, and neighborhood dimensions, achieving an efficacious uncertainty reduction. Overall, the proposed algorithm may serve as a universal framework for parcel-wise type inference, facilitating the effective implementation of precision agriculture.

The precise extraction of crop type information on agricultural land supports applications such as agricultural information statistics and planning. It is also a crucial foundation for improving agricultural production efficiency and promoting agricultural informatization. In smallholder agricultural regions, such as the southern agricultural areas of China, a significant number of small parcels exist. These small parcels often exhibit deficiencies and discrepancies in feature representation for time series classification of crop types, leading to considerable classification challenges. To achieve more precise crop type differentiation in smallholder agricultural systems, this study designs a parcel-based classification framework, PITT (Parcel-level Integration of Time series and Texture). The PITT framework categorizes small parcels in smallholder systems by area into small parcels and micro parcels, which are then separately used as inputs for time series classification methods and high-resolution texture classification methods. During the process, the time series classification results guide the high-resolution texture classification method. Finally, the results from the texture classification are fused with the time series classification results, achieving more accurate crop classification outcomes. The study focuses on the Jiang area of Zongyang County, Tongling city, Anhui Province. Experimental validations using Pearson correlation coefficients and TWDTW similarity comparisons reveal that larger parcels have time series features that more strongly represent the features of typical samples. Additionally, when the PITT framework was compared with other time series classification models using real labels, the F1 scores for small parcels of approximately 0.1–0.5 hectares increased for rapeseed and wheat, reaching 0.93 and 0.94, respectively. For micro parcels (less than 0.1 ha), the F1 scores improved by at least 4.11% and 17.05%, respectively. This demonstrates the ability to achieve high crop classification performance with minimal labelling in smallholder systems, advancing the informatization of smallholder agriculture.

With the rapid development of the social economy,digitization,informatization,and intelligence have become important trends in promoting national construction.In the era of big data,the essence of intelligent applications in various industries is to correlate ubiquitous information and solve required parameters in their respective spaces.How to intelligently analyze large-scale spatial parameters on the complex land surface,which is dependent on natural resources,has become an important proposition for digital driven high-quality development in the new era.As a new trend in the development of Artificial Intelligence(AI),the revolutionary influence of Large Models(LMs)on scientific research paradigms,production methods,and industrial models cannot be underestimated.Investing in LM research is an inevitable choice.In the field of geographic AI,a significant gap remains between the scientific design and practical application of LMs.This article adheres to the principle of deconstructing complex land surface systems and solving precise land parameters.It proposes to conduct land spatial object-oriented modeling supported by multisource and multimodal observation data.The article proposes an object-oriented modeling approach for land surface space,integrating basic geographic data to build an object-oriented base,and transmitting remote sensing data to collaborators' knowledge in a signal manner to systematically analyze complex land spaces.On this basis,a land spatial parameter system and a solution framework are outlined via the integration of five land parameters from land use,land cover change,land soil,land resource,and land type/application.Furthermore,an intelligent computing remote sensing LM is designed for large-scale parameter solving via integrating three core systems,namely,symbol,perception,and control systems.This model deploys heterogeneous deep learning algorithms to break through the bottlenecks in mapping,transforming,and transmitting relationships on key nodes.It is worth emphasizing that in deep learning algorithms,we introduce attention and external incremental information to solve the ordered decomposition and step-by-step simplification of complex problems,thereby achieving large-scale,accurate,and fast solution of land parameters.A preliminary experiment is conducted using the solution of land use parameters in agricultural production spaces as an application case.Results show that the proposed framework has great potential in improving the accuracy of large-scale parameter calculation in land space.The experiment has shown that the remote sensing model constructed in this article has good performance,revealing that the land spatial information generated by this research model has five advantages of measurability,detectability,verifiability,optimizability,and customizability,and has broad application potential in the comprehensive service of human,land,money,and matter.The proposed model helps serve the intelligent customization of refined land information products and deepen the understanding of land space.Finally,prospects for LM research on land spatial parameter calculation are presented from the perspectives of model adaptability/robustness and interpretability/credibility of results.This study is based on the existing research of the authors'team and presents the spiral evolution from remote sensing regression to geography,from big data to big model research in recent years.It is another milestone in theoretical development and practical application.It should be noted that the LM framework established in this article is more of an intelligent computing strategy proposed for solving large-scale land parameter problems in complex geographic systems,and there is still room for optimization and adjustment in specific implementation stages.

Karst mountain areas, as complex geological systems formed by carbonate rock development, possess unique three-dimensional spatial structures and hydrogeological processes that fundamentally influence regional ecosystem evolution, land resource assessment, and sustainable development strategy formulation. In recent years, through the implementation of systematic ecological restoration projects, the ecological degradation of karst mountain areas in Southwest China has been significantly curbed. However, the research on the fine-grained land use mapping and quantitative characterization of spatial heterogeneity in karst mountain areas is still insufficient. This knowledge gap impedes scientific decision-making and precise policy formulation for regional ecological environment management. Hence, this paper proposes a novel methodology for land use mapping in karst mountain areas using very high resolution (VHR) remote sensing (RS) images. The innovation of this method lies in the introduction of strategies of geographical zoning and stratified object extraction. The former divides the complex mountain areas into manageable subregions to provide computational units and introduces a priori data for providing constraint boundaries, while the latter implements a processing mechanism with a deep learning (DL) of hierarchical semantic boundary-guided network (HBGNet) for different geographic objects of building, water, cropland, orchard, forest-grassland, and other land use features. Guanling and Zhenfeng counties in the Huajiang section of the Beipanjiang River Basin, China, are selected to conduct the experimental validation. The proposed method achieved notable accuracy metrics with an overall accuracy (OA) of 0.815 and a mean intersection over union (mIoU) of 0.688. Comparative analysis demonstrated the superior performance of advanced DL networks when augmented with priori knowledge in geographical zoning and stratified object extraction. The approach provides a robust mapping framework for generating fine-grained land use data in karst landscapes, which is beneficial for supporting academic research, governmental analysis, and related applications.

Accurately determining the spatial position and distribution structure of agricultural cultivation parcels (ACPs) is essential for regional agricultural planning and food security. Currently, utilizing deep learning technology based on very high resolution remote sensing imagery has proven effective for intelligent parcel extraction. However, relying solely on the model output, especially from single-task models in mountainous regions with complex, heterogeneous, and fragmented smallholder agriculture, remains questionable. To address this challenge, leveraging geographical prior knowledge is critical. This article proposes using the deep semantic segmentation algorithm in conjunction with comprehensive prior strategies. An improved densely connected link network (D-LinkNet) is employed to delineate the parcels, while geographical zoning, coarse spatial scope, stratification strategy, and homogeneity checking are exerted to understand regions, facilitate samples, reduce interferences, decompose objects, and identify undersegmentation. The proposed framework was validated in Jiangjin district, Chongqing of China, using Gaofen-2 images as the vital data. Compared to the method relying solely on deep learning, our method achieved superior performance with an overall accuracy of 0.924, Kappa coefficient of 0.847, $F1$ score of 0.921, and IoU exceeding 0.8. Moreover, the results demonstrated high accuracy in the individual geometric precision of parcel. Over 1.23 million parcels were identified, comprising 77% cultivated land and 23% garden land. The areal proportion of paddy fields, drylands, and pepper gardens approximated 1:1:1, consistent with statistical data. This method offers a feasible approach for finely extracting agricultural parcels.

Abstract. The precise mapping of crop spatial distribution using remote sensing datasets is a fundamental task in precision agriculture, which has experienced profound development triggered by the continuous improvement of earth observation systems jointly with the innovation of machine learning theories. However, the extraction and classification of cultivated areas are typically accomplished simultaneously with single-task models at pixel scale, and the prior geographical knowledge was generally ignored, which may present accuracy limitation and significant separation from the monitor application. In response, the proposed work is a novel attempt to address successively the parcel extraction and parcel-wise crop classification over mountainous regions with heterogeneous and fragmented smallholder agriculture. Land parcels get precisely delineated utilizing an improved Densely Connected Link Network (D-LinkNet) with geo-knowledge as prior constraints. Each parcel is then correlated with its closest neighbors considering the environmental and temporal similarity, and classified subsequently by a proposed attention-based Network. Results show that ideal precision was attained in both stages. The incorporation of prior knowledge and neighborhood information has effectively enhanced the accuracy of parcel extraction and crop classification, respectively. Overall, the parcel-wise crop mapping framework may constrain the analysis range within the agricultural space and provides identification results corresponding to real geographic objects, contributing directly to the downstream applications such as crop monitoring, management decision-making, etc.

Hyperspectral image (HSI) shows great potential for application in remote sensing due to its rich spectral information and fine spatial resolution. However, the high dimensionality, nonlinearity, and complex relationship between spectral and spatial features of HSI pose challenges to its accurate classification. Traditional convolutional neural network (CNN)-based methods suffer from detail loss in feature extraction; Transformer-based methods rely too much on the quantity and quality of HSI; and graph neural network (GNN)-based methods provide a new impetus for HSI classification by virtue of their excellent ability to handle irregular data. To address these challenges and take advantage of GNN, we propose a network of parallel GNNs called PGNN-Net. The network first extracts the key spatial-spectral features of HSI using principal component analysis, followed by preprocessing to obtain two primary features and a normalized adjacency matrix. Then, a parallel architecture is constructed using improved GCN and ChebNet to extract local and global spatial-spectral features, respectively. Finally, the discriminative features obtained through the fusion strategy are input into the classifier to obtain the classification results. In addition, to alleviate the over-fitting problem, the label smoothing technique is embedded in the cross-entropy loss function. The experimental results show that the average overall accuracy obtained by our method on Indian Pines, Kennedy Space Center, Pavia University Scene, and Botswana reaches 97.35%, 99.40%, 99.64%, and 98.46%, respectively, which are better compared to some state-of-the-art methods.

Recently, high-resolution (HR) remote sensing images have attracted increasing attention in a number of tasks. Super-resolution (SR) is an efficient method to obtain high-resolution remote sensing images. Due to the influence of imaging distances and angles, remote sensing images significantly differ from natural images in terms of land cover element distribution, ground object scale and scene complexity. This poses a challenge for capturing global and local low- and high-frequency and restoring fine image details for remote sensing image SR. In this article, a local-global context-aware generative dual-region adversarial network (LGC-GDAN) is designed for remote sensing image SR. It is composed of dual region-level discriminators and a dual-path generator with a context-aware network and an edge-assisted network. To capture global and local low- and high-frequency information, the global-aware self-attention (GAS) mechanism and local-aware self-attention (LAS) mechanism are introduced into the context-aware network. The GAS mechanism combines high-pass and low-pass filtering for long-range similarity feature, while LAS uses local aggregation for fine-level feature. The LR images and the corresponding edge maps are input to the edge-assisted network to extract the detailed geometric structure. To address small ground object and complex ground scenes, conventional image-level discriminators exhibit limited performance in capturing detailed information. Unlike previous discriminator, a region-level discriminator is designed to obtain the real/fake label of each local region. Moreover, two task-driven loss functions are designed to produce diverse images for further scene classification. The experiments undertaken on several remote sensing datasets demonstrate that LGC-GDAN outperforms the other state-of-the-art methods.

Download This Paper Open PDF in Browser Add Paper to My Library Share: Permalink Using these links will ensure access to this page indefinitely Copy URL Copy DOI

The formation of shadows in very high spatial resolution (VHR) remote sensing imagery is attributed to light being blocked by objects, reducing spectral radiance in the shadow landscape. An accurate and robust shadow removal method can recover spectral and textural information and, hence, is a crucial preprocessing step for urban image analyses. In this study, we develop a KnOwledge-driven shadow progressive removal (KO-Shadow) framework with three subnets for VHR imagery using a weakly supervised manner. Specifically, the shadow preelimination subnet is proposed to initially address the large chromatic aberration between the real and shadow situations. Then, the prior knowledge-guided refinement subnet is proposed to refine the preelimination results by mining tone and texture information. Moreover, the locality feature discriminator is designed for region-specific evaluation of the generated shadow-free samples to improve the capacity of subnets. Experimental results of six typical cities in the world show that KO-Shadow is superior to the existing methods. Moreover, the generalizability analysis in complex urban scenarios validates the robustness of our method. The shadow recovery score (SRI) is proposed to evaluate the spectral similarities between the recovered area and shadow-related land-cover types (e.g., road, building, and lawn). The results show that KO-Shadow can yield more visually realistic shadow-free images and better quantitative performance. Overall, KO-Shadow provides a new perspective for VHR image shadow removal by mining the prior knowledge of the complex shadows in urban areas.

The current research on rocky desertification primarily prioritizes large-scale surveillance, with minimal attention given to internal agricultural areas. This study offers a comprehensive framework for bedrock extraction in agricultural areas, employing spatial constraints and spatio-temporal fusion methodologies. Utilizing the high resolution and capabilities of Gaofen-2 imagery, we first delineate agricultural land, use these boundaries as spatial constraints to compute the agricultural land bedrock response Index (ABRI), and apply the spatial and temporal adaptive reflectance fusion model (STARFM) to achieve spatio-temporal fusion of Gaofen-2 imagery and Sentinel-2 imagery from multiple time periods, resulting in a high-spatio-temporal-resolution bedrock discrimination index (ABRI*) for analysis. This work demonstrates the pronounced rocky desertification phenomenon in the agricultural land in the study area. The ABRI* effectively captures this phenomenon, with the classification accuracy for the bedrock, based on the ABRI* derived from Gaofen-2 imagery, reaching 0.86. The bedrock exposure area in the farmland showed a decreasing trend from 2019 to 2021, a significant increase from 2021 to 2022, and a gradual decline from 2022 to 2024. Cultivation activities have a significant impact on rocky desertification within agricultural land. The ABRI significantly enhances the capabilities for the dynamic monitoring of rocky desertification in agricultural areas, providing data support for the management of specialized farmland. For vulnerable areas, timely adjustments to planting schemes and the prioritization of intervention measures such as soil conservation, vegetation restoration, and water resource management could help to improve the resilience and stability of agriculture, particularly in karst regions.

Parcel-scale crop classification utilizing time-series satellite observations is of significant importance in precision agriculture. The prior knowledge that crop types can be organized in a hierarchical tree structure is beneficial for improving crop classification. Moreover, the crop hierarchy aligns with the coarse-to-fine cognitive process of geographic scenes. Based on the crop hierarchy, this study developed a general hierarchical classification framework for enhancing crop mapping using time-series Sentinel-1 data. Central to this method is a deep-learning-based hierarchical classification model that explores and makes use of crop hierarchical knowledge. First, preprocessed Sentinel-1 data were geometrically overlaid onto farmland parcel maps to derive parcel-scale time-series features. Second, we constructed a hierarchical crop type system for study areas based on the crop phenology of labeled crop-type samples. Third, we developed a deep-learning-based hierarchical classification model to identify crop types for each parcel, to generate final crop-type classification maps. The proposed approach was further discussed and verified through the implementation of parcel-scale time-series crop hierarchical classifications in a study area in France with farmland parcel maps and time-series Sentinel-1 data. The classification results, indicating significant improvements greater than 4.0% in overall accuracy and 5.0% in F1 score over comparative methods, demonstrated the effectiveness of the proposed method in learning multi-scale time-series features for hierarchical crop classification utilizing Sentinel-1 data sequences.