Pub Date : 2025-02-01DOI: 10.1016/j.jag.2025.104417
Junwei Wang , Shuguo Chen , Shixian Hu , Linke Deng , Chaofei Ma , Hailong Peng , Qingjun Song
Ultraviolet (UV) remote sensing plays a critical role in understanding photochemical and biological processes in the global ocean. While UV radiation significantly influences the marine environment, the limited availability of global UV measurements has hindered comprehensive analyses, particularly in photochemically sensitive regions. The Ultraviolet Imager (UVI) on China’s HaiYang-1C (HY-1C) satellite, launched in 2018, offers unique data with two UV bands (355 and 385 nm), enabling the study of UV-driven oceanic processes that were previously unachievable with standard visible-only ocean color sensors. This study develops a system vicarious calibration (SVC) approach tailored for HY-1C’s UVI, integrating co-located observations from the Coastal Zone Color Scanner (COCTS) on the same satellite platform to derive accurate remote sensing reflectance (Rrs) in the UV spectrum. Using MOBY in situ measurements as reference data for SVC and ship-based measurements for validation, we demonstrate that UVI-derived Rrs achieve high accuracy, with Mean Absolute Percentage Differences (MAPD) reduced to 15.7 % and 8.4 % for the 355 and 385 nm bands, respectively, following system vicarious calibration. This enhanced accuracy provides a pathway for producing consistent UV ocean color products and contributes to a deeper understanding of marine biogeochemical cycles. The findings highlight the potential of HY-1C UVI in expanding ocean color research into the UV domain, offering valuable insights for future satellite missions.
{"title":"System vicarious calibration and ocean color retrieval from the HY-1C UVI","authors":"Junwei Wang , Shuguo Chen , Shixian Hu , Linke Deng , Chaofei Ma , Hailong Peng , Qingjun Song","doi":"10.1016/j.jag.2025.104417","DOIUrl":"10.1016/j.jag.2025.104417","url":null,"abstract":"<div><div>Ultraviolet (UV) remote sensing plays a critical role in understanding photochemical and biological processes in the global ocean. While UV radiation significantly influences the marine environment, the limited availability of global UV measurements has hindered comprehensive analyses, particularly in photochemically sensitive regions. The Ultraviolet Imager (UVI) on China’s HaiYang-1C (HY-1C) satellite, launched in 2018, offers unique data with two UV bands (355 and 385 nm), enabling the study of UV-driven oceanic processes that were previously unachievable with standard visible-only ocean color sensors. This study develops a system vicarious calibration (SVC) approach tailored for HY-1C’s UVI, integrating co-located observations from the Coastal Zone Color Scanner (COCTS) on the same satellite platform to derive accurate remote sensing reflectance (<em>R<sub>rs</sub></em>) in the UV spectrum. Using MOBY <em>in situ</em> measurements as reference data for SVC and ship-based measurements for validation, we demonstrate that UVI-derived <em>R<sub>rs</sub></em> achieve high accuracy, with Mean Absolute Percentage Differences (<em>MAPD</em>) reduced to 15.7 % and 8.4 % for the 355 and 385 nm bands, respectively, following system vicarious calibration. This enhanced accuracy provides a pathway for producing consistent UV ocean color products and contributes to a deeper understanding of marine biogeochemical cycles. The findings highlight the potential of HY-1C UVI in expanding ocean color research into the UV domain, offering valuable insights for future satellite missions.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"136 ","pages":"Article 104417"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143402961","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}
Pub Date : 2025-02-01DOI: 10.1016/j.jag.2024.104349
Xuanguang Liu , Yujie Li , Chenguang Dai , Zhenchao Zhang , Lei Ding , Mengmeng Li , Hanyun Wang
Building extraction from very high-resolution remote-sensing images still faces two main issues: (1) small buildings are severely omitted and the extracted building shapes have a low consistency with ground truths. (2) supervised deep-learning methods have poor performance in few-shot scenarios, limiting the practical application of these methods. To address the first issue, we propose an asymmetric Siamese multitask network integrating adversarial edge learning called ASMBR-Net for building extraction. It contains an efficient asymmetric Siamese feature extractor comprising pre-trained backbones of convolutional neural networks and Transformers under pre-training and fine-tuning paradigms. This extractor balances the local and global feature representation and reduces training costs. Adversarial edge-learning technology automatically integrates edge constraints and strengthens the modeling ability of small and complex building-shaped patterns. Aiming to overcome the second issue, we introduce a self-training framework and design an instance transfer strategy to generate reliable pseudo-samples. We examined the proposed method on the WHU and Massachusetts (MA) datasets and a self-constructed Dongying (DY) dataset, comparing it with state-of-the-art methods. The experimental results show that our method achieves the highest F1-score of 96.06%, 86.90%, and 84.98% on the WHU, MA, and DY datasets, respectively. Ablation experiments further verify the effectiveness of the proposed method. The code is available at: https://github.com/liuxuanguang/ASMBR-Net
从极高分辨率遥感图像中提取建筑物仍然面临两个主要问题:(1)小型建筑物被严重遗漏,提取的建筑物形状与地面实况的一致性较低。(2) 有监督的深度学习方法在少镜头场景下性能较差,限制了这些方法的实际应用。针对第一个问题,我们提出了一种集成对抗边缘学习的非对称连体多任务网络,称为 ASMBR-Net,用于建筑物提取。它包含一个高效的非对称连体特征提取器,由预先训练的卷积神经网络骨干和预训练和微调范式下的变形器组成。这种提取器平衡了局部和全局特征表示,降低了训练成本。对抗边缘学习技术可自动整合边缘约束,增强对小型和复杂建筑形态的建模能力。为了克服第二个问题,我们引入了一个自我训练框架,并设计了一种实例转移策略来生成可靠的伪样本。我们在 WHU 和 Massachusetts(MA)数据集以及自建的东营(DY)数据集上检验了所提出的方法,并将其与最先进的方法进行了比较。实验结果表明,我们的方法在 WHU、MA 和 DY 数据集上分别取得了 96.06%、86.90% 和 84.98% 的最高 F1 分数。消融实验进一步验证了所提方法的有效性。代码见: https://github.com/liuxuanguang/ASMBR-Net
{"title":"Extraction buildings from very high-resolution images with asymmetric siamese multitask networks and adversarial edge learning","authors":"Xuanguang Liu , Yujie Li , Chenguang Dai , Zhenchao Zhang , Lei Ding , Mengmeng Li , Hanyun Wang","doi":"10.1016/j.jag.2024.104349","DOIUrl":"10.1016/j.jag.2024.104349","url":null,"abstract":"<div><div>Building extraction from very high-resolution remote-sensing images still faces two main issues: (1) small buildings are severely omitted and the extracted building shapes have a low consistency with ground truths. (2) supervised deep-learning methods have poor performance in few-shot scenarios, limiting the practical application of these methods. To address the first issue, we propose an asymmetric Siamese multitask network integrating adversarial edge learning called ASMBR-Net for building extraction. It contains an efficient asymmetric Siamese feature extractor comprising pre-trained backbones of convolutional neural networks and Transformers under pre-training and fine-tuning paradigms. This extractor balances the local and global feature representation and reduces training costs. Adversarial edge-learning technology automatically integrates edge constraints and strengthens the modeling ability of small and complex building-shaped patterns. Aiming to overcome the second issue, we introduce a self-training framework and design an instance transfer strategy to generate reliable pseudo-samples. We examined the proposed method on the WHU and Massachusetts (MA) datasets and a self-constructed Dongying (DY) dataset, comparing it with state-of-the-art methods. The experimental results show that our method achieves the highest F1-score of 96.06%, 86.90%, and 84.98% on the WHU, MA, and DY datasets, respectively. Ablation experiments further verify the effectiveness of the proposed method. The code is available at: <span><span>https://github.com/liuxuanguang/ASMBR-Net</span><svg><path></path></svg></span></div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"136 ","pages":"Article 104349"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142974852","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}
Pub Date : 2025-02-01DOI: 10.1016/j.jag.2025.104355
Zhe Zhao , Jiangluqi Song , Dong Zhao , Jiajia Zhang , Huixin Zhou , Jun Zhou
Hyperspectral anomaly detection (HAD) holds significant importance in remote sensing image processing and has been recently empowered by deep learning-based methods. Despite achieving good performance by reconstructing the background and suppressing abnormal targets, these methods often fail to distinguish the attributes of latent features during background reconstruction, resulting in unsatisfactory reconstruction quality. This occurs because the background usually has low-frequency properties and the anomalies have high-frequency properties, so these issues should be considered during background reconstruction. To overcome this weakness, a multi-scale frequency-guided two-stream network (MFTNet) is proposed for HAD. Firstly, considering the different frequency attributes of background and anomalies, we use an encoder to extract features and then design a multi-scale frequency decomposition module to decompose the features into high-frequency (HF) part for anomalies and the low-frequency (LF) part for background. Furthermore, different from anomaly suppression-based methods, we introduce a high-frequency enhancement module to highlight abnormal targets in the HF part, making them stand out in the background. Meanwhile, the features from the encoder contain more background information, so we fuse the encoded features and LF part features by a self-attention module to accurately recover the background. Finally, we perform anomaly detection on the reconstructed anomaly and background parts, respectively, and the average of those two results is the final detection map. Comparative analysis on six datasets against thirteen baseline methods, including seven effective deep learning-based methods, demonstrates that our proposed MFTNet delivers competitive detection results. The code of this work will be released at: https://github.com/xautzhaozhe/MFTNet.
{"title":"Multi-scale frequency-guided two-stream network for hyperspectral anomaly detection","authors":"Zhe Zhao , Jiangluqi Song , Dong Zhao , Jiajia Zhang , Huixin Zhou , Jun Zhou","doi":"10.1016/j.jag.2025.104355","DOIUrl":"10.1016/j.jag.2025.104355","url":null,"abstract":"<div><div>Hyperspectral anomaly detection (HAD) holds significant importance in remote sensing image processing and has been recently empowered by deep learning-based methods. Despite achieving good performance by reconstructing the background and suppressing abnormal targets, these methods often fail to distinguish the attributes of latent features during background reconstruction, resulting in unsatisfactory reconstruction quality. This occurs because the background usually has low-frequency properties and the anomalies have high-frequency properties, so these issues should be considered during background reconstruction. To overcome this weakness, a multi-scale frequency-guided two-stream network (MFTNet) is proposed for HAD. Firstly, considering the different frequency attributes of background and anomalies, we use an encoder to extract features and then design a multi-scale frequency decomposition module to decompose the features into high-frequency (HF) part for anomalies and the low-frequency (LF) part for background. Furthermore, different from anomaly suppression-based methods, we introduce a high-frequency enhancement module to highlight abnormal targets in the HF part, making them stand out in the background. Meanwhile, the features from the encoder contain more background information, so we fuse the encoded features and LF part features by a self-attention module to accurately recover the background. Finally, we perform anomaly detection on the reconstructed anomaly and background parts, respectively, and the average of those two results is the final detection map. Comparative analysis on six datasets against thirteen baseline methods, including seven effective deep learning-based methods, demonstrates that our proposed MFTNet delivers competitive detection results. The code of this work will be released at: <span><span>https://github.com/xautzhaozhe/MFTNet</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"136 ","pages":"Article 104355"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142975202","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}
Frequent algal blooms pose a serious threat to the marine ecosystem of the East China Sea. The Geostationary Ocean Color Imager-Ⅱ (GOCI-Ⅱ), a second-generation geostationary satellite sensor, is crucial for monitoring marine environmental dynamics. To evaluate the potential of GOCI-II for identifying and monitoring the diurnal variation of algal blooms in the East China Sea, we combined a coupled ocean–atmosphere model with the eXtreme Gradient Boosting (XGBoost) method to develop an atmospheric correction algorithm for coastal waters (XGB-CW). Validation showed that this algorithm derived remote sensing reflectance (Rrs) from GOCI-Ⅱ with higher accuracy than those provided by the National Ocean Satellite Center of South Korea (NOSC). To further evaluate GOCI-Ⅱ’s potential for algal bloom types identification, we compared three identification algorithms’ (Bloom Index (BI), Diatom Index (DI), and Rslope) results with Rrs data derived by XGB-CW. And the BI algorithm performed best in distinguishing the diatoms and dinoflagellates blooms, while Rslope was effective under high biomass conditions. The DI algorithm was good for diatoms blooms but less effective for dinoflagellates. Using Photosynthetically Available Radiation (PAR) and Sea Surface Temperature (SST) data, we analyzed the influence of these factors on the daily variations and characteristics of Akashiwo sanguinea (Dinoflagellate) and Chaetoceros curvisetus (Diatom). The results showed more pronounced daily variations in A. sanguinea compared to C. curvisetus. GOCI-Ⅱ, combined with accurate atmospheric correction and identification algorithms, plays a crucial role in algal bloom monitoring.
频繁的藻华对东海的海洋生态系统构成严重威胁。地球同步海洋彩色成像仪-Ⅱ(GOCI-Ⅱ)是第二代地球同步卫星传感器,对监测海洋环境动态至关重要。为了评估GOCI-II识别和监测东海藻华日变化的潜力,我们将海洋-大气耦合模型与极端梯度增强(XGBoost)方法相结合,开发了一种沿海水域大气校正算法(XGB-CW)。验证结果表明,该算法获得GOCI-Ⅱ遥感反射率(Rrs)的精度高于韩国国家海洋卫星中心(NOSC)提供的遥感反射率。为了进一步评价GOCI-Ⅱ对藻华类型识别的潜力,我们将三种识别算法(bloom Index (BI)、硅藻Index (DI)和Rslope)的结果与XGB-CW获得的Rrs数据进行了比较。BI算法在区分硅藻和鞭毛藻华方面效果最好,而Rslope算法在高生物量条件下效果最好。DI算法对硅藻华效果较好,但对鞭毛藻效果较差。利用光合有效辐射(PAR)和海温(SST)资料,分析了这些因素对赤潮赤藻(Akashiwo sanguinea, Dinoflagellate)和弯角毛藻(Chaetoceros curvisetus, Diatom)的日变化和特征的影响。结果显示,与C. curvisetus相比,A. sanguinea的每日变化更为明显。GOCI-Ⅱ结合精确的大气校正和识别算法,在藻华监测中起着至关重要的作用。
{"title":"Identifying algal bloom types and analyzing their diurnal variations using GOCI-Ⅱ data","authors":"Renhu Li, Fang Shen, Yuan Zhang, Zhaoxin Li, Songyu Chen","doi":"10.1016/j.jag.2025.104377","DOIUrl":"10.1016/j.jag.2025.104377","url":null,"abstract":"<div><div>Frequent algal blooms pose a serious threat to the marine ecosystem of the East China Sea. The Geostationary Ocean Color Imager-Ⅱ (GOCI-Ⅱ), a second-generation geostationary satellite sensor, is crucial for monitoring marine environmental dynamics. To evaluate the potential of GOCI-II for identifying and monitoring the diurnal variation of algal blooms in the East China Sea, we combined a coupled ocean–atmosphere model with the eXtreme Gradient Boosting (XGBoost) method to develop an atmospheric correction algorithm for coastal waters (XGB-CW). Validation showed that this algorithm derived remote sensing reflectance (<em>R</em><sub>rs</sub>) from GOCI-Ⅱ with higher accuracy than those provided by the National Ocean Satellite Center of South Korea (NOSC). To further evaluate GOCI-Ⅱ’s potential for algal bloom types identification, we compared three identification algorithms’ (Bloom Index (BI), Diatom Index (DI), and R<sub>slope</sub>) results with <em>R</em><sub>rs</sub> data derived by XGB-CW. And the BI algorithm performed best in distinguishing the diatoms and dinoflagellates blooms, while R<sub>slope</sub> was effective under high biomass conditions. The DI algorithm was good for diatoms blooms but less effective for dinoflagellates. Using Photosynthetically Available Radiation (PAR) and Sea Surface Temperature (SST) data, we analyzed the influence of these factors on the daily variations and characteristics of <em>Akashiwo sanguinea</em> (Dinoflagellate) and <em>Chaetoceros curvisetus</em> (Diatom). The results showed more pronounced daily variations in <em>A. sanguinea</em> compared to <em>C. curvisetus</em>. GOCI-Ⅱ, combined with accurate atmospheric correction and identification algorithms, plays a crucial role in algal bloom monitoring.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"136 ","pages":"Article 104377"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142990289","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}
Pub Date : 2025-02-01DOI: 10.1016/j.jag.2025.104387
Jianao Cai , Dongping Ming , Feng Liu , Xiao Ling , Ningjie Liu , Liang Zhang , Lu Xu , Yan Li , Mengyuan Zhu
Interferometric Synthetic Aperture Radar (InSAR) techniques are commonly used approach for identifying Slow-moving Landslide (SML). However, most SML boundary identification with deep learning are based on single-source InSAR data, which cannot fully explore the dynamic process of destabilization, and are inefficient due to high model complexity. Meanwhile, research on automatic procession with multi-source InSAR data is few. To enhance efficiency in geohazard monitoring, this paper proposed an automatic framework for Boundary-Changed Slow-moving Landslide (BCSML) detection by integrating multi-source Small Baseline Subset InSAR (SBAS-InSAR), Convolutional Neural Network (CNN), and change detection methodologies. Firstly, surface deformation was estimated using multi-source SBAS-InSAR. Then, a novel and effective Light-U2Net was constructed with decreased complexity to identify Significant Deformation Zone (SDZ) and locate SML candidate. Finally, BCSMLs were identified using a change detection approach based on newly defined geometric measurements. Two study areas were selected to test the model’s performance: Zayu County and the Nu-Lancang River parallel flow (NLPF) area (in China). The proposed Light-U2Net model achieves high Precision (80.1 %), Recall (80.2 %), and F1-scores (80.1 %) in Zayu County. Additionally, the model’s complexity has reduced by 42.4 % without compromising identification accuracy compared to the original model. The pre-trained model was then applied to the NLPF area, and a total of 273 BCSMLs were detected, with 176 identified as expanding and 97 as shrinking. BCSML identification accuracy can reach to 90.47 %. The results have proved that the proposed framework with the Light-U2Net model is effective and practically potential in landslide disaster prevention.
{"title":"Change detection of slow-moving landslide with multi-source SBAS-InSAR and Light-U2Net","authors":"Jianao Cai , Dongping Ming , Feng Liu , Xiao Ling , Ningjie Liu , Liang Zhang , Lu Xu , Yan Li , Mengyuan Zhu","doi":"10.1016/j.jag.2025.104387","DOIUrl":"10.1016/j.jag.2025.104387","url":null,"abstract":"<div><div>Interferometric Synthetic Aperture Radar (InSAR) techniques are commonly used approach for identifying Slow-moving Landslide (SML). However, most SML boundary identification with deep learning are based on single-source InSAR data, which cannot fully explore the dynamic process of destabilization, and are inefficient due to high model complexity. Meanwhile, research on automatic procession with multi-source InSAR data is few. To enhance efficiency in geohazard monitoring, this paper proposed an automatic framework for Boundary-Changed Slow-moving Landslide (BCSML) detection by integrating multi-source Small Baseline Subset InSAR (SBAS-InSAR), Convolutional Neural Network (CNN), and change detection methodologies. Firstly, surface deformation was estimated using multi-source SBAS-InSAR. Then, a novel and effective Light-U<sup>2</sup>Net was constructed with decreased complexity to identify Significant Deformation Zone (SDZ) and locate SML candidate. Finally, BCSMLs were identified using a change detection approach based on newly defined geometric measurements. Two study areas were selected to test the model’s performance: Zayu County and the Nu-Lancang River parallel flow (NLPF) area (in China). The proposed Light-U<sup>2</sup>Net model achieves high Precision (80.1 %), Recall (80.2 %), and F1-scores (80.1 %) in Zayu County. Additionally, the model’s complexity has reduced by 42.4 % without compromising identification accuracy compared to the original model. The pre-trained model was then applied to the NLPF area, and a total of 273 BCSMLs were detected, with 176 identified as expanding and 97 as shrinking. BCSML identification accuracy can reach to 90.47 %. The results have proved that the proposed framework with the Light-U<sup>2</sup>Net model is effective and practically potential in landslide disaster prevention.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"136 ","pages":"Article 104387"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143050022","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}
Pub Date : 2025-02-01DOI: 10.1016/j.jag.2024.104304
Xin Shen , Lin Cao , Yisheng Ma , Nicholas C. Coops , Evan R. Muise , Guibin Wang , Fuliang Cao
Bamboo forests are natural habitat for the giant panda which is one of the most vulnerable mammal species. In structurally complex natural forests, bamboos are normally located under the canopy of taller trees, which makes them difficult to be quantified accurately. Although Light Detection and Ranging (LiDAR) technologies have been well established as the effective tool for forest structure assessment, the use of LiDAR to assess understory bamboo in structurally complex natural forests is less well known. We present a novel vertical vegetation classification (VVC) approach to map the structure of understory bamboos for giant panda forage in natural forests. An optimized demarcation point identification (DPI) model was developed for stratifying different vertical layers from coarse to fine scales. Three-dimensional understory bamboo point clouds were successfully isolated from the forest point cloud, then bamboo structure predictive models were developed through understory bamboo point cloud metrics and applied over the entire study area to generate spatially continuous maps of understory bamboo structure. Our results indicate that the isolation of the understory bamboo point cloud using the developed VVC approach performs well and has small bias, the extracted maximum height is close to field-measured maximum height (R2 = 0.77, rRMSE = 15.02 %). Height-related metrics have higher correlations with bamboo structure (mean natural and true height, basal diameter, and total aboveground biomass) than other metrics (r > 0.8), and understory bamboo structures are estimated with relatively high accuracy (R2 = 0.84 – 0.91, rRMSE = 10.87 – 29.41 %). We also find varying effects of topography on the spatial distribution of different understory bamboo species. This study demonstrates the benefits of utilizing LiDAR data to ascertain fine-scale understory bamboo resources, providing critical supports for giant panda habitat assessment and conservation.
{"title":"Estimating structure of understory bamboo for giant panda habitat by developing an advanced vertical vegetation classification approach using UAS-LiDAR data","authors":"Xin Shen , Lin Cao , Yisheng Ma , Nicholas C. Coops , Evan R. Muise , Guibin Wang , Fuliang Cao","doi":"10.1016/j.jag.2024.104304","DOIUrl":"10.1016/j.jag.2024.104304","url":null,"abstract":"<div><div>Bamboo forests are natural habitat for the giant panda which is one of the most vulnerable mammal species. In structurally complex natural forests, bamboos are normally located under the canopy of taller trees, which makes them difficult to be quantified accurately. Although Light Detection and Ranging (LiDAR) technologies have been well established as the effective tool for forest structure assessment, the use of LiDAR to assess understory bamboo in structurally complex natural forests is less well known. We present a novel vertical vegetation classification (VVC) approach to map the structure of understory bamboos for giant panda forage in natural forests. An optimized demarcation point identification (DPI) model was developed for stratifying different vertical layers from coarse to fine scales. Three-dimensional understory bamboo point clouds were successfully isolated from the forest point cloud, then bamboo structure predictive models were developed through understory bamboo point cloud metrics and applied over the entire study area to generate spatially continuous maps of understory bamboo structure. Our results indicate that the isolation of the understory bamboo point cloud using the developed VVC approach performs well and has small bias, the extracted maximum height is close to field-measured maximum height (R<sup>2</sup> = 0.77, rRMSE = 15.02 %). Height-related metrics have higher correlations with bamboo structure (mean natural and true height, basal diameter, and total aboveground biomass) than other metrics (<em>r</em> > 0.8), and understory bamboo structures are estimated with relatively high accuracy (R<sup>2</sup> = 0.84 – 0.91, rRMSE = 10.87 – 29.41 %). We also find varying effects of topography on the spatial distribution of different understory bamboo species. This study demonstrates the benefits of utilizing LiDAR data to ascertain fine-scale understory bamboo resources, providing critical supports for giant panda habitat assessment and conservation.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"136 ","pages":"Article 104304"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142825344","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}
Pub Date : 2025-02-01DOI: 10.1016/j.jag.2024.104316
Euihyun Kim , Changbin Lim , Jung Lyul Lee
Satellite images have been adopted in recent years for identifying topographical features on the Earth’s surface. Researchers have also published reports on the use of satellite images to analyze shoreline changes or to verify shoreline change in numerical models. But reports that demonstrate the reverse process of using satellite images to estimate the incident waves to a beach are rare, particularly to a place where protective coastal structures exist. This paper describes a once-thriving coastal townsite with two fishing ports in Korea which has been transformed into a typical example that relies on protective structures with occasional artificial nourishment to maintain its shoreline stability in the past 20 years. Unlike many others, this study proposes a new methodology to estimate the deepwater wave heights based on the analysis of shoreline data extracted from satellite images over 5 years (2019–2023) for Wonpyeong-Chogok Beach, its median sediment grain sizes , and the known empirical relationship between sediment and waves. The entire shoreline of 2,860 m in length is divided into 39 transects, of which one-half of it is protected by submerged and emergent detached breakwaters, where shoreline has advanced, while the rest has eroded. From the standard deviation values calculated from the extracted shoreline location data, the influence of long-term trends was excluded, and the intrinsic standard deviation is obtained by applying sediment size information, and then the incident deep-water (average annual maximum) wave height of 4.363 m was estimated. Applying this methodology to the beach area where the coastal structure was placed, the wave transmission of the coastal structure was calculated 0.91 and 0.72 for LCSs and TT-DBWs, respectively, through the reduction ratio of the standard deviation. Finally, discussions are made on how the resolution of the Sentinel-2 satellite images in affecting the standard deviation and long-term trend results in the shoreline data.
{"title":"Utilization of Sentinel-2 satellite imagery for correlation analysis of shoreline variation and incident waves: Application to Wonpyeong-Chogok Beach, Korea","authors":"Euihyun Kim , Changbin Lim , Jung Lyul Lee","doi":"10.1016/j.jag.2024.104316","DOIUrl":"10.1016/j.jag.2024.104316","url":null,"abstract":"<div><div>Satellite images have been adopted in recent years for identifying topographical features on the Earth’s surface. Researchers have also published reports on the use of satellite images to analyze shoreline changes or to verify shoreline change in numerical models. But reports that demonstrate the reverse process of using satellite images to estimate the incident waves to a beach are rare, particularly to a place where protective coastal structures exist. This paper describes a once-thriving coastal townsite with two fishing ports in Korea which has been transformed into a typical example that relies on protective structures with occasional artificial nourishment to maintain its shoreline stability in the past 20 years. Unlike many others, this study proposes a new methodology to estimate the deepwater wave heights based on the analysis of shoreline data extracted from satellite images over 5 years (2019–2023) for Wonpyeong-Chogok Beach, its median sediment grain sizes <span><math><msub><mi>D</mi><mn>50</mn></msub></math></span>, and the known empirical relationship between sediment and waves. The entire shoreline of 2,860 m in length is divided into 39 transects, of which one-half of it is protected by submerged and emergent detached breakwaters, where shoreline has advanced, while the rest has eroded. From the standard deviation values calculated from the extracted shoreline location data, the influence of long-term trends was excluded, and the intrinsic standard deviation is obtained by applying sediment size information, and then the incident deep-water (average annual maximum) wave height of 4.363 m was estimated. Applying this methodology to the beach area where the coastal structure was placed, the wave transmission of the coastal structure was calculated 0.91 and 0.72 for LCSs and TT-DBWs, respectively, through the reduction ratio of the standard deviation. Finally, discussions are made on how the resolution of the Sentinel-2 satellite images in affecting the standard deviation and long-term trend results in the shoreline data.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"136 ","pages":"Article 104316"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142825346","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}
Pub Date : 2025-02-01DOI: 10.1016/j.jag.2024.104314
Ziying Zhou , Saini Yang , Siqin Wang , Xiaoyan Liu , Fuyu Hu , Yaqiao Wu , Yu Chen
Compound hazards caused by tropical cyclones involve interactions among multiple hazards, such as wind, rainfall, and storm surge, significantly increasing the uncertainty and destructiveness of disasters. Existing studies primarily focus on probabilistic analyses of single or dual hazards associated with tropical cyclones, revealing limitations in handling high-dimensional data and complex dependencies. This study developed the ComHazAsTC-RRE (Compound Hazard Assessment of Tropical Cyclones within Repeatable, Reproducible, and Expandable Framework) model to analyze the compound hazards of wind, rainfall, and storm surge induced by tropical cyclones and successfully applied it to China’s coast. We collected globally accessible daily records of maximum wind speed, cumulative rainfall, and maximum storm surge for China’s coastal areas from 1979 to 2018. Using a C-Vine Copula with wind speed as the core, incorporating rainfall and storm surge as branches, we accurately captured complex dependencies of tropical cyclones. Our various return period analyses underscore the importance of considering multiple hazards and their interactions. Additionally, the application of Compound Hazard Index in China reveals that southeastern coastal areas are subjected to significantly higher compound hazards, driven by high wind speeds and strong spatial–temporal consistency of hazards. An in-depth analysis of failure probabilities indicates that neglecting the interactions among hazards can result in substantial additional cost for engineering projects, especially during severe tropical cyclones. This study offers new perspectives and scientific tools for understanding and addressing compound hazards, formulating effective disaster prevention and mitigation strategies, and supporting the sustainable development of coastal regions worldwide.
热带气旋造成的复合灾害涉及风、雨和风暴潮等多种灾害之间的相互作用,大大增加了灾害的不确定性和破坏性。现有的研究主要集中在与热带气旋相关的单一或双重危害的概率分析上,揭示了在处理高维数据和复杂依赖关系方面的局限性。本研究建立了ComHazAsTC-RRE (Compound Hazard Assessment of Tropical cyclone within Repeatable, Reproducible, and Expandable Framework)模型,分析了热带气旋引起的风、雨、风暴潮的复合危害,并成功应用于中国沿海地区。我们收集了1979年至2018年中国沿海地区全球可获取的最大风速、累积降雨量和最大风暴潮的日记录。使用以风速为核心,以降雨和风暴潮为分支的C-Vine Copula,我们准确地捕捉了热带气旋的复杂依赖关系。我们的各种回报期分析强调了考虑多种危害及其相互作用的重要性。复合灾害指数在中国的应用表明,东南沿海地区受高风速和强时空一致性的影响,复合灾害强度显著增加。对破坏概率的深入分析表明,忽视灾害之间的相互作用可能导致工程项目的大量额外成本,特别是在严重的热带气旋期间。该研究为认识和应对复合灾害,制定有效的防灾减灾战略,支持全球沿海地区的可持续发展提供了新的视角和科学工具。
{"title":"ComHazAsTC-RRE: Compound Hazard Assessment of Tropical Cyclones within Repeatable, Reproducible, and Expandable Framework","authors":"Ziying Zhou , Saini Yang , Siqin Wang , Xiaoyan Liu , Fuyu Hu , Yaqiao Wu , Yu Chen","doi":"10.1016/j.jag.2024.104314","DOIUrl":"10.1016/j.jag.2024.104314","url":null,"abstract":"<div><div>Compound hazards caused by tropical cyclones involve interactions among multiple hazards, such as wind, rainfall, and storm surge, significantly increasing the uncertainty and destructiveness of disasters. Existing studies primarily focus on probabilistic analyses of single or dual hazards associated with tropical cyclones, revealing limitations in handling high-dimensional data and complex dependencies. This study developed the ComHazAsTC-RRE (<u>Com</u>pound <u>Haz</u>ard <u>As</u>sessment of <u>T</u>ropical <u>C</u>yclones within <u>R</u>epeatable, <u>R</u>eproducible, and <u>E</u>xpandable Framework) model to analyze the compound hazards of wind, rainfall, and storm surge induced by tropical cyclones and successfully applied it to China’s coast. We collected globally accessible daily records of maximum wind speed, cumulative rainfall, and maximum storm surge for China’s coastal areas from 1979 to 2018. Using a C-Vine Copula with wind speed as the core, incorporating rainfall and storm surge as branches, we accurately captured complex dependencies of tropical cyclones. Our various return period analyses underscore the importance of considering multiple hazards and their interactions. Additionally, the application of Compound Hazard Index in China reveals that southeastern coastal areas are subjected to significantly higher compound hazards, driven by high wind speeds and strong spatial–temporal consistency of hazards. An in-depth analysis of failure probabilities indicates that neglecting the interactions among hazards can result in substantial additional cost for engineering projects, especially during severe tropical cyclones. This study offers new perspectives and scientific tools for understanding and addressing compound hazards, formulating effective disaster prevention and mitigation strategies, and supporting the sustainable development of coastal regions worldwide.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"136 ","pages":"Article 104314"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142874814","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}
Pub Date : 2025-02-01DOI: 10.1016/j.jag.2024.104319
Daeyeol Kim , Youngkeun Song , Hansoo Kim , Ohsung Kwon , Young-Kwang Yeon , Taiyang Lim
Accurate identification and classification of tree species on a large scale are crucial for the effective management of urban green spaces; however, previous research combining airborne sensors such as LiDAR and hyperspectral imaging for tree classification has generally focused on smaller areas or specific sites, with limited studies applying this approach at the city-wide scale. This study focuses on the utilization of multi-temporal airborne light detection and ranging (LiDAR) and hyperspectral imaging (HSI) data for the classification of 10 species of urban trees at the city scale, which collectively cover over 95 % of the tree-covered areas within the city. Our objective is to evaluate the utility of metrics and indices derived from LiDAR (leaf-on/leaf-off) and HSI (peak growing season/autumn senescence) data in a 35.86 km2 urban green space in Gwacheon, Republic of Korea. A comprehensive set of 15 independent variables was extracted from preprocessed and calibrated airborne LiDAR data (footprint size: 0.46 m, density: 42.7 points/m2) and HSI data (127 bands, 400–970 nm range, spatial resolution: 0.68 m) to train seven machine learning classifiers. The model was trained on a stratified random sample of 21,826 tree crown polygon samples collected from individual trees surveyed. The results showed that the combination of airborne LiDAR and HSI data from two seasons achieved the highest classification accuracy with the light gradient boosting machine (LGBM) classifier (90.6 %; Kappa: 0.895) for all 10 major tree species across the entire city, especially for Ginkgo, American sycamore, and Yoshino cherry. Among all variables, the maximum tree height (Hmax) and the intersection symmetric difference ratio index (ISDRI) were among the top influential factors for tree species classification accuracy. Hmax, with an importance value of 0.490, is particularly effective due to the characteristics of urban green spaces. ISDRI, with an importance value of 0.336, highlights seasonal leaf volume differences, aiding in species differentiation. The spectral indices acquired during the autumn leaf senescence showed a cumulative shapley additive explanations (SHAP) importance score that was 0.374 points higher than that of the leaf-on period, highlighting the enhanced significance of hyperspectral data from the leaf senescence phase in classifying tree species. The synergistic integration of airborne LiDAR, HSI, and seasonal data gathered during key phenological periods, along with relevant indices, will contribute significantly to urban forest management at the city-wide level.
{"title":"Airborne multi-seasonal LiDAR and hyperspectral data integration for individual tree-level classification in urban green spaces at city scale","authors":"Daeyeol Kim , Youngkeun Song , Hansoo Kim , Ohsung Kwon , Young-Kwang Yeon , Taiyang Lim","doi":"10.1016/j.jag.2024.104319","DOIUrl":"10.1016/j.jag.2024.104319","url":null,"abstract":"<div><div>Accurate identification and classification of tree species on a large scale are crucial for the effective management of urban green spaces; however, previous research combining airborne sensors such as LiDAR and hyperspectral imaging for tree classification has generally focused on smaller areas or specific sites, with limited studies applying this approach at the city-wide scale. This study focuses on the utilization of multi-temporal airborne light detection and ranging (LiDAR) and hyperspectral imaging (HSI) data for the classification of 10 species of urban trees at the city scale, which collectively cover over 95 % of the tree-covered areas within the city. Our objective is to evaluate the utility of metrics and indices derived from LiDAR (leaf-on/leaf-off) and HSI (peak growing season/autumn senescence) data in a 35.86 km<sup>2</sup> urban green space in Gwacheon, Republic of Korea. A comprehensive set of 15 independent variables was extracted from preprocessed and calibrated airborne LiDAR data (footprint size: 0.46 m, density: 42.7 points/m<sup>2</sup>) and HSI data (127 bands, 400–970 nm range, spatial resolution: 0.68 m) to train seven machine learning classifiers. The model was trained on a stratified random sample of 21,826 tree crown polygon samples collected from individual trees surveyed. The results showed that the combination of airborne LiDAR and HSI data from two seasons achieved the highest classification accuracy with the light gradient boosting machine (LGBM) classifier (90.6 %; Kappa: 0.895) for all 10 major tree species across the entire city, especially for Ginkgo, American sycamore, and Yoshino cherry. Among all variables, the maximum tree height (Hmax) and the intersection symmetric difference ratio index (ISDRI) were among the top influential factors for tree species classification accuracy. Hmax, with an importance value of 0.490, is particularly effective due to the characteristics of urban green spaces. ISDRI, with an importance value of 0.336, highlights seasonal leaf volume differences, aiding in species differentiation. The spectral indices acquired during the autumn leaf senescence showed a cumulative shapley additive explanations (SHAP) importance score that was 0.374 points higher than that of the leaf-on period, highlighting the enhanced significance of hyperspectral data from the leaf senescence phase in classifying tree species. The synergistic integration of airborne LiDAR, HSI, and seasonal data gathered during key phenological periods, along with relevant indices, will contribute significantly to urban forest management at the city-wide level.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"136 ","pages":"Article 104319"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142901791","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}
Pub Date : 2025-02-01DOI: 10.1016/j.jag.2024.104340
Qingcheng Pan , Zonghan Ma , Hantian Wu , Nana Yan , Weiwei Zhu , Yixuan Wang , Bingfang Wu
Land surface temperature (LST) serves as a crucial indicator of the thermal state and environmental changes on the Earth’s surface, and it can be retrieved effectively from satellite thermal infrared sensors. Although algorithms for retrieving LSTs have been developed successfully for many satellites, the newly launched Sustainable Development Science Satellite 1 (SDGSAT-1), which includes three thermal infrared bands, does not yet include effective LST algorithms and parameters. Here, parameters are calibrated for retrieving LSTs from SDGSAT-1 Thermal Infrared Spectrometer (TIS) data for the split-window (SW) method and the three-channel (TC) method under both daytime and nighttime conditions. In this process, the Thermodynamic Initial Guess Retrieval (TIGR) dataset and observation data from the University of Wyoming were used. Validations were conducted using in situ LST measurements at the Guantao, Turpan, and Heihe sites in China and from the Surface Radiation Budget (SURFRAD) network in North America, covering cropland, desert and bare land, and grassland. The overall accuracies of the models are fairly good, with RMSEs of 2.507 K and 2.272 K for split-window method during daytime and nighttime respectively, and 2.847 K and 1.923 K for three-channel method. Additionally, the LST retrieval models that use observation data from the University of Wyoming had higher accuracy than those using the TIGR2000 profiles. Currently, the proposed models can be applied under different atmospheric water vapor contents and underlying surface conditions both during the day and at night, paving the way for retrieving LST products from SDGSAT-1.
{"title":"Algorithm parameters for retrieving land surface temperature from the SDGSAT-1 thermal infrared spectrometer","authors":"Qingcheng Pan , Zonghan Ma , Hantian Wu , Nana Yan , Weiwei Zhu , Yixuan Wang , Bingfang Wu","doi":"10.1016/j.jag.2024.104340","DOIUrl":"10.1016/j.jag.2024.104340","url":null,"abstract":"<div><div>Land surface temperature (LST) serves as a crucial indicator of the thermal state and environmental changes on the Earth’s surface, and it can be retrieved effectively from satellite thermal infrared sensors. Although algorithms for retrieving LSTs have been developed successfully for many satellites, the newly launched Sustainable Development Science Satellite 1 (SDGSAT-1), which includes three thermal infrared bands, does not yet include effective LST algorithms and parameters. Here, parameters are calibrated for retrieving LSTs from SDGSAT-1 Thermal Infrared Spectrometer (TIS) data for the split-window (SW) method and the three-channel (TC) method under both daytime and nighttime conditions. In this process, the Thermodynamic Initial Guess Retrieval (TIGR) dataset and observation data from the University of Wyoming were used. Validations were conducted using in situ LST measurements at the Guantao, Turpan, and Heihe sites in China and from the Surface Radiation Budget (SURFRAD) network in North America, covering cropland, desert and bare land, and grassland. The overall accuracies of the models are fairly good, with RMSEs of 2.507 K and 2.272 K for split-window method during daytime and nighttime respectively, and 2.847 K and 1.923 K for three-channel method. Additionally, the LST retrieval models that use observation data from the University of Wyoming had higher accuracy than those using the TIGR2000 profiles. Currently, the proposed models can be applied under different atmospheric water vapor contents and underlying surface conditions both during the day and at night, paving the way for retrieving LST products from SDGSAT-1.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"136 ","pages":"Article 104340"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142901861","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}