International journal of applied earth observation and geoinformation : ITC journal最新文献

{"title":"A novel hyperspectral remote sensing estimation model for surface soil texture using AHSI/ZY1-02D satellite image","authors":"Qiang Shen ,&nbsp;Kun Shang ,&nbsp;Chenchao Xiao ,&nbsp;Hongzhao Tang ,&nbsp;Taixia Wu ,&nbsp;Changkun Wang","doi":"10.1016/j.jag.2025.104453","DOIUrl":"10.1016/j.jag.2025.104453","url":null,"abstract":"<div><div>Soil texture is an essential attribute of soil structure, which plays an important role in evaluating soil fertility and carrying out agricultural production. This study developed a novel soil texture estimation model using ZiYuan-1-02D (ZY1-02D) satellite Advanced Hyperspectral Imager (AHSI), based on the mechanism of soil spectral mixing, that enables simultaneous estimation of the three soil texture attributes (clay, silt, and sand). Study area is located in the north-eastern region of China covering 1683.31 km². To reduce data redundancy, we used correlation analysis and Competitive Adaptive Reweighted Sampling (CARS) algorithms to select sensitive spectral features of soil texture, and excluded spectral bands that are strongly influenced by other soil physicochemical properties. Finally, the spatial distribution map and classification map of soil texture have been generated for the study area. We also used AHSI/GaoFen-5 (GF-5) satellite images to further validate the generalizability of the model. The results suggest that the model can be used in the estimation of soil texture, and the developed novel model can effectively reflect the spatial distribution characteristics of surface soil texture attributes. The <em>R</em>² values of all outcomes for inverting three texture attributes were larger than 0.5, with silt exhibiting the best estimation effect (<em>R</em>² = 0.79, RMSE = 6.46 %, RPD = 2.19). The Max-divergence between the estimated surface soil texture attributes based on the two satellite images (AHSI/ZY1-02D and AHSI/GF-5) and the measured data were less than 4 %. The novel spectral mixture model of soil texture is suitable for spaceborne remote sensing data and has broad application prospects in surface soil texture mapping.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"138 ","pages":"Article 104453"},"PeriodicalIF":7.6,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143548158","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An operational Airborne-Ground Integrate observation scheme for validating land surface temperature over heterogeneous surface","authors":"Yajun Huang ,&nbsp;Wenping Yu ,&nbsp;Xujun Han ,&nbsp;Jianguang Wen ,&nbsp;Qing Xiao ,&nbsp;Xufeng Wang ,&nbsp;Jiayuan Lin ,&nbsp;Zengjing Song ,&nbsp;Dandan Li ,&nbsp;Xiangyi Deng","doi":"10.1016/j.jag.2025.104450","DOIUrl":"10.1016/j.jag.2025.104450","url":null,"abstract":"<div><div>At present, there are more than 30 satellite remote sensing Land Surface Temperature (LST) products from kilometers to hectometers resolutions. The accuracy of these products is the key issue for further application. The validation of LST products is mainly achieved through ground observations on homogeneous surfaces, but the accuracy of satellite products on heterogeneous surfaces is also an important factor in the performance of satellite products. We proposed an integrated airborne-ground observation scheme to validate the accuracy of hectometers Landsat LST product. Firstly, in this scheme, the optimal deployment of ground observations is constructed by the prior knowledge, which is the brightness temperature from an unmanned aerial vehicle(UAV). Secondly, UAV flight which synchronization with satellite transit to obtain brightness temperature. Thirdly, the atmospheric effect between the UAV and the ground observations is corrected by the radiative transfer equation. Finally, the LST over the heterogenous land surface is validated by upscaled UAV LST. The results showed that the error between the UAV LST and the ground observations could be reduced from 3.2 K to about 0.5 K by calibrating the near-surface atmospheric effect. Besides, the validation of the LST satellite product by upscaling the UAV LST as “true values”, the results showed that the accuracy was about 1.17 K of Landsat product in heterogeneous surface, the bias was more observably with more big heterogeneity of surface which might cause by adjacent effect in Landsat products. This paper has achieved integrated airborne-space-ground observation and provided a better solution for satellite product validation on heterogeneous surfaces.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"138 ","pages":"Article 104450"},"PeriodicalIF":7.6,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143548159","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Dynamic inference for on-orbit scene classification with the scale boosting model","authors":"Kunyang Yang ,&nbsp;Naisen Yang ,&nbsp;Hong Tang","doi":"10.1016/j.jag.2025.104447","DOIUrl":"10.1016/j.jag.2025.104447","url":null,"abstract":"<div><div>Existing scene classification methods allocate the same computational resources, i.e., all model parameters in the neural network, to each remote sensing image whenever from any geographic scene. However, this might be redundant for images of certain scenes that are easy to discriminate, e.g., homogeneous scenes. This observation motivates us to propose an efficient method for on-orbit scene classification, namely, the Scale Boosting Model (SBM). Specifically, during the training process, the SBM is built as a set of different scale learners in a scale-increasing manner, each of which is used to learn and classify image features at a specific scale. During inference, the scale learners in the SBM will be selectively run in a scale-increasing manner and automatically decide when to exit early or expand the computation according to the scene complexity. In addition, by replacing the backbone of the scale learner, the SBM could provide a deployment possibility for computationally limited models for on-orbit processing, thereby reducing their computational requirements. Extensive experiments on UC Merced Land Use, NWPU-RESISC45 and RSD46-WHU datasets show that the SBM achieved a more effective classification performance more efficiently.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"138 ","pages":"Article 104447"},"PeriodicalIF":7.6,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143562069","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluating Earth observation products for Catchment-Scale operational flood monitoring and risk management in a sparsely gauged to ungauged river basin in Nigeria","authors":"Dorcas Idowu ,&nbsp;Brad G. Peter ,&nbsp;Jessica Boakye ,&nbsp;Sagy Cohen ,&nbsp;Elizabeth Carter","doi":"10.1016/j.jag.2025.104445","DOIUrl":"10.1016/j.jag.2025.104445","url":null,"abstract":"<div><div>With the persistent rise in intensity and magnitude of hydrological extremes globally, timely information from operational early flood warning systems provide lead times that translate into actionable strategies to monitor and mitigate flood risk. However, the situation is often different for flood-prone regions of the global south with sparse to no ground flood monitoring systems, where flood management practices are inadequate or lacking, and floods are further exacerbated due to cross-border water infrastructure management practices. With the unprecedented volume of data from Earth Observation Satellites (EOS), such as MODIS (Moderate Resolution Imaging Spectroradiometer) and passive microwave radiometry (PMR) river discharges, insights are being provided into different catchment hydrologic variables required for early flood detection and risk management. This work presents a geospatial and satellite-based heuristic that will be of use for operational flood risk management, which ties together PMR river discharges, the Floodwater Depth Estimation Tool (FwDET), and MODIS-based inundation detection. Case studies of the 2012, 2018, 2020, and 2022 floods in the Lower Niger River Basin in Nigeria are presented, with emphasis placed on the devastating 2012 and 2022 floods. Furthermore, using the time-series from the PMR river discharges, a flood frequency analysis was performed. The analysis was found to capture peak discharges corresponding to the flood events. The result of the flood frequency analysis shows that the return periods of all the flood events combined are approximately equal to or below 30-year, with the 2012 and 2022 catastrophic floods approximately 14- and 20-year floods at satellite gauging reach (SGR) 1438, while 14- and 30-year floods at SGR 1441. From the perspective of a 100-year flood regulatory magnitude, these return periods are an indication of possible river flooding in its natural floodplain. The predicted flow magnitude of a 100-year flood at SGR 1438 was 44% and 38% greater than the 2012 and 2022 flood magnitudes, respectively; 18% and 12% greater at SGR 1441. The results show that operational flood monitoring and risk management or assessment are possible in sparse to ungauged river basins using these products, especially with appropriate predictive algorithms to enable incorporation in an early flood warning system. Since satellite-based measurements have no regard for political boundaries, possible effects of cross-border water infrastructure management practices within the basin could also be assessed.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"138 ","pages":"Article 104445"},"PeriodicalIF":7.6,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143562070","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"WAPooling: An adaptive plug-and-play module for feature aggregation in point cloud classification networks","authors":"Kristin Eggen,&nbsp;Hongchao Fan","doi":"10.1016/j.jag.2025.104439","DOIUrl":"10.1016/j.jag.2025.104439","url":null,"abstract":"<div><div>Deep learning methods for classification have achieved significant advancements in processing 3D point clouds. A fundamental aspect of deep learning networks is how to best aggregate features into a global representation of the point cloud. While many existing networks rely on the traditional max-pooling for feature aggregation due to its efficiency and permutation-invariance, max-pooling has some limitations. It is a fixed operation, lacking learnability and adaptability, which restricts the network’s ability to capture the most informative global feature representation. This limitation motivates for developing a new feature aggregation method that generates a richer and more diverse feature representation. Therefore, this paper introduces a novel pooling operation, the Weighted Activation Pooling (WAP) module. WAP adds adaptability to the pooling operation by using learnable weights to dynamically adjust the importance of the pooled features. Features are further transformed through different activation functions, allowing the network to learn complex patterns and relations in the data. The WAP module is a plug-and-play module that can replace the traditional pooling operation in existing networks, with minimal computational overhead. Moreover, this paper introduces the AdaptNet classification network, where the proposed WAP module is used to obtain global features. Extensive experiments are conducted to evaluate AdaptNet using both real-world data and the ModelNet40 dataset. Results show that AdaptNet outperforms other networks, achieving higher overall accuracy on both datasets. Moreover, WAP is integrated into existing classification networks, and experiments using real-world data show an increased performance of all tested networks compared to using their original pooling strategy.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"138 ","pages":"Article 104439"},"PeriodicalIF":7.6,"publicationDate":"2025-03-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143548160","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"GeoCode-GPT: A large language model for geospatial code generation","authors":"Shuyang Hou ,&nbsp;Zhangxiao Shen ,&nbsp;Anqi Zhao ,&nbsp;Jianyuan Liang ,&nbsp;Zhipeng Gui ,&nbsp;Xuefeng Guan ,&nbsp;Rui Li ,&nbsp;Huayi Wu","doi":"10.1016/j.jag.2025.104456","DOIUrl":"10.1016/j.jag.2025.104456","url":null,"abstract":"<div><div>The increasing demand for spatiotemporal data and modeling tasks in geosciences has made geospatial code generation technology a critical factor in enhancing productivity. Although large language models (LLMs) have demonstrated potential in code generation tasks, they often encounter issues such as refusal to code or hallucination in geospatial code generation due to a lack of domain-specific knowledge and code corpora. To address these challenges, this paper presents and open-sources the GeoCode-PT and GeoCode-SFT corpora, along with the GeoCode-Eval evaluation dataset. Additionally, by leveraging QLoRA and LoRA for pretraining and fine-tuning, we introduce GeoCode-GPT-7B, the first LLM focused on geospatial code generation, fine-tuned from Code Llama-7B. Furthermore, we establish a comprehensive geospatial code evaluation framework, incorporating option matching, expert validation, and prompt engineering scoring for LLMs, and systematically evaluate GeoCode-GPT-7B using the GeoCode-Eval dataset. Experimental results reveal that GeoCode-GPT significantly outperforms existing models across multiple tasks. For multiple-choice tasks, its accuracy improves by 9.1% to 32.1%. In code summarization, it achieves superior scores in completeness, accuracy, and readability, with gains ranging from 1.7 to 25.4 points. For code generation, its performance in accuracy, readability, and executability surpasses benchmarks by 1.2 to 25.1 points. Grounded in the fine-tuning paradigm, this study introduces and validates an approach to enhance LLMs in geospatial code generation and associated tasks. These findings extend the application boundaries of such models in geospatial domains and offer a robust foundation for exploring their latent potential.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"138 ","pages":"Article 104456"},"PeriodicalIF":7.6,"publicationDate":"2025-03-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143644249","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Emergency-oriented fine change detection of flood-damaged farmland from medium-resolution remote sensing images","authors":"Gang Qin,&nbsp;Shixin Wang,&nbsp;Futao Wang,&nbsp;Zhenqing Wang,&nbsp;Suju Li,&nbsp;Xingguang Gu,&nbsp;Kailong Hu,&nbsp;Longfei Liu","doi":"10.1016/j.jag.2025.104442","DOIUrl":"10.1016/j.jag.2025.104442","url":null,"abstract":"<div><div>Flood disasters are characterized by frequent and sudden occurrences and obvious chain effects, posing a major threat to agricultural production. Government disaster relief and agricultural insurance are increasingly urgent in assessing losses to flood-damaged farmland. There are many challenges in assessing flood-damaged farmland. On the one hand, the historical data needed for the farmland loss assessment model is missing. On the other hand, the farmland loss assessment conducted within the flood inundation area only focuses on the completely submerged farmland, which is not accurate enough. Semi-supervised deep learning can effectively alleviate the above problems. This study specifically designed a fine-grained change detection model for flood-damaged farmland using medium-resolution remote sensing (RS) images to achieve multi-scenario change detection of flood-damaged farmland. In order to assist in the automatic sample generation of large-scale unlabeled sample RS images of flood status, a semi-supervised sample generation framework for flood-damaged farmland using medium-resolution RS images is proposed. Based on this framework, FloodedCropland datasets is created. The experimental results show that the proposed change detection model has an F1-score of 0.9047 on flood-damaged farmland. After the semi-supervised sample generation framework optimized the model, the change detection F1-score was improved to 0.9241. Experiments have verified that the automatic generation of labels for flood-damaged farmland in medium-resolution RS images using a semi-supervised sample generation framework performs better than the scarce manual labeling model and can save a lot of manual labeling time. The consistent performance in different geographic regions and under different RS satellites imaging conditions demonstrate the practical application potential of this method for cross-regional and cross-RS satellites intelligent information extraction in natural disaster scenes with a lack of labeled samples.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"138 ","pages":"Article 104442"},"PeriodicalIF":7.6,"publicationDate":"2025-03-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143548161","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"PatchOut: A novel patch-free approach based on a transformer-CNN hybrid framework for fine-grained land-cover classification on large-scale airborne hyperspectral images","authors":"Renjie Ji ,&nbsp;Kun Tan ,&nbsp;Xue Wang ,&nbsp;Shuwei Tang ,&nbsp;Jin Sun ,&nbsp;Chao Niu ,&nbsp;Chen Pan","doi":"10.1016/j.jag.2025.104457","DOIUrl":"10.1016/j.jag.2025.104457","url":null,"abstract":"<div><div>Airborne hyperspectral systems can provide high-resolution hyperspectral images (HSIs) covering large scenes, enabling fine-grained land-cover classification. However, the most popular patch-based methods are limited by low computational efficiency and broken classification results, which hinders the full utilization of this powerful technology in Earth observation applications. Therefore, in this paper, considering the efficiency requirements for large-scale land-cover classification, a novel <strong>p</strong>atch-free <strong>a</strong>pproach based on a <strong>T</strong>ransformer-<strong>C</strong>NN <strong>h</strong>ybrid (PatchOut) framework is proposed. The proposed PatchOut framework adopts an encoder-decoder architecture, enabling rapid semantic segmentation for HSI classification. For the encoder module, we introduce a computationally efficient reduced Transformer module integrated with convolutional neural network (CNN), to leverage their complementary strengths for long-range and local feature extraction, respectively. A multi-scale spatial-spectral feature fusion (MSSSFF) module is also proposed to amalgamate the characteristics of different levels from the encoder, which enhances the overall feature representation. Then, to address the loss of semantic detail and resolution inherent in multi-level feature extraction, a novel feature reconstruction module (FRM) is applied to recover high-quality semantic features. Finally, a large-scale benchmark dataset, Qingpu-HSI, is presented, comprising airborne HSIs covering 33.91 km² with 20 land-cover classes. Experiments on the Qingpu-HSI and another public dataset demonstrate the superior accuracy and efficiency of our proposed PatchOut framework, outperforming several well-known patch-free and patch-based methods. The Qingpu HSI dataset, along with the PatchOut framework code will be released at <span><span>https://github.com/busbyjrj/PatchOut</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"138 ","pages":"Article 104457"},"PeriodicalIF":7.6,"publicationDate":"2025-03-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143548031","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"AIAM: Adaptive interactive attention model for solving p-Median problem via deep reinforcement learning","authors":"Haojian Liang ,&nbsp;Shaohua Wang ,&nbsp;Huilai Li ,&nbsp;Jie Pan ,&nbsp;Xiao Li ,&nbsp;Cheng Su ,&nbsp;Bingzhi Liu","doi":"10.1016/j.jag.2025.104454","DOIUrl":"10.1016/j.jag.2025.104454","url":null,"abstract":"<div><div>The p-Median Problem (PMP) is a classical discrete facility location problem with significant implications for optimizing the placement of urban public service facilities. Improved heuristics, a well-established method for solving the PMP, aim to iteratively enhance solution quality through efficient neighborhood exploration. In this study, we model the neighborhood exploration process as a Markov decision process and propose a novel deep reinforcement learning approach to solving the PMP, achieving higher problem-solving efficiency and quality. The proposed method introduces an encoder-decoder structure, consisting of an Interactive Attention Encoder (IAE), a Node Removal Decoder (NRD), and a Node Insertion Decoder (NID), aimed at learning an optimal strategy for node selection. The experimental results demonstrate that our approach outperforms genetic algorithms in terms of both accuracy and computational efficiency. While the solution time is slightly longer than that of the Attention Model (AM), our method achieves a reduced gap to the optimal solution. Furthermore, ablation studies confirm that the proposed adaptive interactive encoder and the two decoders significantly enhance the model performance. Finally, we applied the Adaptive Interactive Attention Model (AIAM) to a real-world scenario, demonstrating its practical utility in guiding medical facility location decisions.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"138 ","pages":"Article 104454"},"PeriodicalIF":7.6,"publicationDate":"2025-03-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143548157","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A lightweight spatiotemporal classification framework for tree species with entropy-based change resistance filter using satellite imagery","authors":"Biao Zhang ,&nbsp;Zhichao Wang ,&nbsp;Boyi Liang ,&nbsp;Liguo Dong ,&nbsp;Zebang Feng ,&nbsp;Mingyang He ,&nbsp;Zhongke Feng","doi":"10.1016/j.jag.2025.104449","DOIUrl":"10.1016/j.jag.2025.104449","url":null,"abstract":"<div><div>The spatiotemporal characteristics of remote sensing data are often time-varying, leading to significant fluctuation and instability in tree species classification results across different years, especially in regions referred to as high-variance areas. To improve the stability and accuracy of the classification results, this study proposes a lightweight spatiotemporal classification framework, with the core algorithm being the Spatiotemporal Entropy-based Change Resistance Filter (STECR-F) algorithm. The STECR-F algorithm integrates the concept of Spatiotemporal Entropy (STE) and, by applying weighted spatiotemporal neighborhood information, suppresses uncertainty in the classification process. It effectively enhances the spatiotemporal consistency of the classification results, particularly in high-variance regions, and reduces classification instability caused by spatiotemporal fluctuations. This study comprehensively evaluates the performance of STECR-F from three dimensions: STE, transfer change, and classification accuracy, and compares it with other methods. The results show that STECR-F significantly reduces the STE value, with an average decrease of 0.3876, effectively mitigating the fluctuation of the classification results. In high-variance regions, the effect of STECR-F is particularly pronounced, with an average decrease in STE value of up to 0.6847. Moreover, STECR-F significantly suppresses random transitions between classes, reducing category transitions by an average of 22.47%, with the maximum reduction reaching 46%. In terms of classification accuracy, STECR-F achieved an overall accuracy of 91.35%, representing an improvement of 8.02% compared to the results without using STECR-F. Additionally, compared to the DMSPN method using only neighborhood information and pattern filtering, STECR-F’s performance improved by 5.86% and 6.42%, respectively. Overall, the STECR-F algorithm effectively addresses the interannual dynamics and uncertainty in tree species classification results. By integrating weighted spatiotemporal neighborhood information, it significantly enhances classification stability and reduces random variability, making it particularly suitable for areas with high spatiotemporal variability and classification uncertainty.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"138 ","pages":"Article 104449"},"PeriodicalIF":7.6,"publicationDate":"2025-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143534162","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}