Pub Date : 2024-10-17DOI: 10.1016/j.jag.2024.104220
Yunbo Zhang , Wenjie Chen , Bingshu Huang , Zongran Zhang , Jie Li , Ruishan Gao , Ke Wang , Chuli Hu
Effective geographic environment observation planning is the key to obtain disaster monitoring and warning information. The previous researches can only make observation plans for a single disaster at some specific stages. They are difficult to apply to the dynamic evolution of the disaster chain. Timely and comprehensive geographic environment observation planning is urgently needed to provide high-value monitoring data for the identification and response of secondary disaster chains. Event logic graph (ELG) shows great potential in evolutionary law expression and chain event reasoning. Therefore, this study proposed an observation ELG (OELG), in which events and their logical relationships are modeled as nodes and edges to express the occurrence and development motivation of observation events. The disaster chain observation planning can be transformed into the reasoning of potential continuous observation events. Subsequently, an OELG-based geographic environment observation planning framework was proposed, which realizes the construction, instantiation, and plan reasoning of OELG. The observation planning experiment was carried out taking the flood disaster chain that occurred in Beijing, China and Nordrhein-Westfalen, Germany as examples. The results show that OELG can generate disaster chain observation plan more timely, comprehensively, and continuously than other models, thus providing support for disaster chain risk monitoring and emergency response.
{"title":"An event logic graph for geographic environment observation planning in disaster chain monitoring","authors":"Yunbo Zhang , Wenjie Chen , Bingshu Huang , Zongran Zhang , Jie Li , Ruishan Gao , Ke Wang , Chuli Hu","doi":"10.1016/j.jag.2024.104220","DOIUrl":"10.1016/j.jag.2024.104220","url":null,"abstract":"<div><div>Effective geographic environment observation planning is the key to obtain disaster monitoring and warning information. The previous researches can only make observation plans for a single disaster at some specific stages. They are difficult to apply to the dynamic evolution of the disaster chain. Timely and comprehensive geographic environment observation planning is urgently needed to provide high-value monitoring data for the identification and response of secondary disaster chains. Event logic graph (ELG) shows great potential in evolutionary law expression and chain event reasoning. Therefore, this study proposed an observation ELG (OELG), in which events and their logical relationships are modeled as nodes and edges to express the occurrence and development motivation of observation events. The disaster chain observation planning can be transformed into the reasoning of potential continuous observation events. Subsequently, an OELG-based geographic environment observation planning framework was proposed, which realizes the construction, instantiation, and plan reasoning of OELG. The observation planning experiment was carried out taking the flood disaster chain that occurred in Beijing, China and Nordrhein-Westfalen, Germany as examples. The results show that OELG can generate disaster chain observation plan more timely, comprehensively, and continuously than other models, thus providing support for disaster chain risk monitoring and emergency response.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"134 ","pages":"Article 104220"},"PeriodicalIF":7.6,"publicationDate":"2024-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142445205","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}
Land cover change detection and classification, including both inter-class changes (land cover conversion, LCC) and intra-class changes (land cover modification, LCM), is critical for understanding the Earth’s dynamic processes and promoting sustainability. However, previous studies have predominantly focused on LCC, with less emphasis on LCM. Land cover classification remains challenging, and its mapping results are often affected by salt and pepper noise. Here, we propose a hybrid approach for continuous change detection and classification of LCC and LCM using Jinchang City, China, as a case study. Firstly, we combined the Continuous Change Detection and Classification (CCDC) and the Bayesian Estimator of Abrupt change, Seasonal change, and Trend (BEAST) algorithms to identify LCC and LCM using all available Landsat time series (TS) data from 2000 to 2020. Then, the harmonic regression coefficients and RMSE values derived from CCDC (hereafter called CCDC features) were fed into the DCNN model for LCC classification. Our findings indicate: (1) For LCC and LCM accuracy assessment, the CCDC and BEAST ensemble achieved a spatial F1 score of 82.7% and an average temporal F1 score of 79.7%. (2) In LCC classification, the DCNN model with CCDC features, particularly DeepLabV3+, outperformed the pixel-based XGBoost and other multi-year land cover products, with frequency-weighted intersection over union (FWIoU), overall accuracy, and Kappa scores of 88.7%, 94%, and 0.87, respectively. (3) Seasonal LCM showed a more concentrated distribution than trend LCM. (4) In Jinchang City, LCM larger than LCC in area, and grassland and cultivated land are the most distributed. Our approach can be contributed to wall-to-wall land surface monitoring and enhance land management capabilities.
{"title":"A fast hybrid approach for continuous land cover change monitoring and semantic segmentation using satellite time series","authors":"Wenpeng Zhao , Rongfang Lyu , Jinming Zhang , Jili Pang , Jianming Zhang","doi":"10.1016/j.jag.2024.104222","DOIUrl":"10.1016/j.jag.2024.104222","url":null,"abstract":"<div><div>Land cover change detection and classification, including both inter-class changes (land cover conversion, LCC) and intra-class changes (land cover modification, LCM), is critical for understanding the Earth’s dynamic processes and promoting sustainability. However, previous studies have predominantly focused on LCC, with less emphasis on LCM. Land cover classification remains challenging, and its mapping results are often affected by salt and pepper noise. Here, we propose a hybrid approach for continuous change detection and classification of LCC and LCM using Jinchang City, China, as a case study. Firstly, we combined the Continuous Change Detection and Classification (CCDC) and the Bayesian Estimator of Abrupt change, Seasonal change, and Trend (BEAST) algorithms to identify LCC and LCM using all available Landsat time series (TS) data from 2000 to 2020. Then, the harmonic regression coefficients and RMSE values derived from CCDC (hereafter called CCDC features) were fed into the DCNN model for LCC classification. Our findings indicate: (1) For LCC and LCM accuracy assessment, the CCDC and BEAST ensemble achieved a spatial F1 score of 82.7% and an average temporal F1 score of 79.7%. (2) In LCC classification, the DCNN model with CCDC features, particularly DeepLabV3+, outperformed the pixel-based XGBoost and other multi-year land cover products, with frequency-weighted intersection over union (FWIoU), overall accuracy, and Kappa scores of 88.7%, 94%, and 0.87, respectively. (3) Seasonal LCM showed a more concentrated distribution than trend LCM. (4) In Jinchang City, LCM larger than LCC in area, and grassland and cultivated land are the most distributed. Our approach can be contributed to wall-to-wall land surface monitoring and enhance land management capabilities.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"134 ","pages":"Article 104222"},"PeriodicalIF":7.6,"publicationDate":"2024-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142445209","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 : 2024-10-17DOI: 10.1016/j.jag.2024.104190
Zequan Chen , Jianping Li , Qusheng Li , Zhen Dong , Bisheng Yang
Automated Aerial Triangulation (AAT), aiming to restore image poses and reconstruct sparse points simultaneously, plays a pivotal role in earth observation. AAT has evolved into a fundamental process widely applied in large-scale Unmanned Aerial Vehicle (UAV) based mapping. However classic AAT methods still face challenges like low efficiency and limited robustness. This paper introduces DeepAAT, a deep learning network designed specifically for AAT of UAV imagery. DeepAAT considers both spatial and spectral characteristics of imagery, enhancing its capability to resolve erroneous matching pairs and accurately predict image poses. DeepAAT marks a significant leap in AAT’s efficiency, ensuring thorough scene coverage and precision. Its processing speed outpaces incremental AAT methods by hundreds of times and global AAT methods by tens of times while maintaining a comparable level of reconstruction accuracy. Additionally, DeepAAT’s scene clustering and merging strategy facilitate rapid localization and pose determination for large-scale UAV images, even under constrained computing resources. The experimental results demonstrate that DeepAAT substantially improves over conventional AAT methods, highlighting its potential for increased efficiency and accuracy in UAV-based 3D reconstruction tasks. To benefit the photogrammetry society, the code of DeepAAT will be released at: https://github.com/WHU-USI3DV/DeepAAT.
{"title":"DeepAAT: Deep Automated Aerial Triangulation for Fast UAV-based mapping","authors":"Zequan Chen , Jianping Li , Qusheng Li , Zhen Dong , Bisheng Yang","doi":"10.1016/j.jag.2024.104190","DOIUrl":"10.1016/j.jag.2024.104190","url":null,"abstract":"<div><div>Automated Aerial Triangulation (AAT), aiming to restore image poses and reconstruct sparse points simultaneously, plays a pivotal role in earth observation. AAT has evolved into a fundamental process widely applied in large-scale Unmanned Aerial Vehicle (UAV) based mapping. However classic AAT methods still face challenges like low efficiency and limited robustness. This paper introduces DeepAAT, a deep learning network designed specifically for AAT of UAV imagery. DeepAAT considers both spatial and spectral characteristics of imagery, enhancing its capability to resolve erroneous matching pairs and accurately predict image poses. DeepAAT marks a significant leap in AAT’s efficiency, ensuring thorough scene coverage and precision. Its processing speed outpaces incremental AAT methods by hundreds of times and global AAT methods by tens of times while maintaining a comparable level of reconstruction accuracy. Additionally, DeepAAT’s scene clustering and merging strategy facilitate rapid localization and pose determination for large-scale UAV images, even under constrained computing resources. The experimental results demonstrate that DeepAAT substantially improves over conventional AAT methods, highlighting its potential for increased efficiency and accuracy in UAV-based 3D reconstruction tasks. To benefit the photogrammetry society, the code of DeepAAT will be released at: <span><span>https://github.com/WHU-USI3DV/DeepAAT</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"134 ","pages":"Article 104190"},"PeriodicalIF":7.6,"publicationDate":"2024-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142445206","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 : 2024-10-17DOI: 10.1016/j.jag.2024.104214
Dingyi Zhou , Zhifang Zhao
Low scattering terrain areas introduce complex phase interference, which reduces the accuracy of deformation signal estimation in InSAR(Interferometric Synthetic Aperture Radar) techniques. Existing covariance matrix-based InSAR phase calculation methods often fail to account for translational offset relations between scatterers leading to inaccuracies, and pixels with zero spatial coherence exist. To address this issue, this paper proposes a distributed scatterer InSAR phase estimation method based on the Cross-Correlation complex coherence matrix. The effectiveness and superiority of the algorithm are verified through simulation and actual data. The results show that: (i) The simulation analysis shows that, compared to the traditional covariance matrix method, the optimal Cross-Correlation matrix improves the interferometric phase, coherence, and accuracy by 21.51%, 15.24%, and 6.52%, respectively. (ii) The actual experimental data show that the interferometric phase optimal by the Cross-Correlation matrix can effectively overcome the pseudo-signal caused by spatial hopping and make the phase more continuous. Compared with the traditional covariance matrix, the average a posteriori coherence and average coherence of arbitrary interference combinations in the Cross-Correlation matrix are improved by 18.12% and 58.10%, respectively. (iii) The number of DS points selected by the Cross-Correlation matrix algorithm is more than that of the covariance matrix algorithm. PS-InSAR (Persistent Scatterer Interferometric Synthetic Aperture Radar) achieved more accurate deformation rates compared to the covariance and correlation matrices, with errors of 9.34, 17.21, and 16.28 when compared against GNSS data, respectively. (iv) The Cross-Correlation matrix reduces the deformation rate error by 5.43 % relative to the covariance matrix. The algorithm provides reliable phase estimation for accurate monitoring of surface deformation in low-scattering regions, supporting geological disaster early warning and resource and environmental management.
{"title":"Optimal algorithm for distributed scatterer InSAR phase estimation based on cross-correlation complex coherence matrix","authors":"Dingyi Zhou , Zhifang Zhao","doi":"10.1016/j.jag.2024.104214","DOIUrl":"10.1016/j.jag.2024.104214","url":null,"abstract":"<div><div>Low scattering terrain areas introduce complex phase interference, which reduces the accuracy of deformation signal estimation in InSAR(Interferometric Synthetic Aperture Radar) techniques. Existing covariance matrix-based InSAR phase calculation methods often fail to account for translational offset relations between scatterers leading to inaccuracies, and pixels with zero spatial coherence exist. To address this issue, this paper proposes a distributed scatterer InSAR phase estimation method based on the Cross-Correlation complex coherence matrix. The effectiveness and superiority of the algorithm are verified through simulation and actual data. The results show that: (i) The simulation analysis shows that, compared to the traditional covariance matrix method, the optimal Cross-Correlation matrix improves the interferometric phase, coherence, and accuracy by 21.51%, 15.24%, and 6.52%, respectively. (ii) The actual experimental data show that the interferometric phase optimal by the Cross-Correlation matrix can effectively overcome the pseudo-signal caused by spatial hopping and make the phase more continuous. Compared with the traditional covariance matrix, the average a posteriori coherence and average coherence of arbitrary interference combinations in the Cross-Correlation matrix are improved by 18.12% and 58.10%, respectively. (iii) The number of DS points selected by the Cross-Correlation matrix algorithm is more than that of the covariance matrix algorithm. PS-InSAR (Persistent Scatterer Interferometric Synthetic Aperture Radar) achieved more accurate deformation rates compared to the covariance and correlation matrices, with errors of 9.34, 17.21, and 16.28 <span><math><mrow><mi>m</mi><mi>m</mi><mo>∙</mo><msup><mrow><mi>a</mi></mrow><mrow><mo>-</mo><mn>1</mn></mrow></msup></mrow></math></span> when compared against GNSS data, respectively. (iv) The Cross-Correlation matrix reduces the deformation rate error by 5.43 % relative to the covariance matrix. The algorithm provides reliable phase estimation for accurate monitoring of surface deformation in low-scattering regions, supporting geological disaster early warning and resource and environmental management.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"134 ","pages":"Article 104214"},"PeriodicalIF":7.6,"publicationDate":"2024-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142445208","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 : 2024-10-17DOI: 10.1016/j.jag.2024.104213
Wenliang Chen , Kun Shang , Yibo Wang , Wenchao Qi , Songtao Ding , Xia Zhang
Effective management of corn straw and stubble is critical in conservation tillage, as it impacts soil health and productivity. However, accurate classification of different types of straw cover has been hindered by their similar spectral and spatial characteristics and the low spatial resolution of hyperspectral satellite imagery. Moreover, traditional convolution neural network (CNN)-based methods, which rely on first-order statistics for feature extraction, often struggle to extract distinguishable features of highly similar objects effectively, thereby reducing classification accuracy. In this study, a second-order statistical-feature extraction algorithm based on CNN that uses fused multispectral and hyperspectral data was tested for its ability to classify types of straw cover. In the first step, coupled non-negative matrix factorization (CNMF) was used to fuse hyperspectral and multispectral images effectively, thereby enhancing the spatial resolution of the hyperspectral data. In this study, we integrated pointwise convolution (PWC), depthwise convolution (DWC), and a distance covariance matrix (DCM) to form a mixed convolution and DCM (MCDCM) network; we used this to extract and integrate deep spectral–spatial features of the hyperspectral images. Our experimental results show that the MCDCM network significantly improved classification accuracy compared to traditional methods, with accuracy rates for the different straw-cover types exceeding 90% and overall accuracy reaching 98.26%. The fused image also exhibited better preservation of feature edges and contours. The accurate identification of corn-straw-cover types achieved with the proposed MCDCM method is a major step in optimizing conservation-farming practices, improving soil fertility and farm productivity, and supporting sustainable ecological development.
{"title":"A mixed convolution and distance covariance matrix network for fine classification of corn straw cover types with fused hyperspectral and multispectral data","authors":"Wenliang Chen , Kun Shang , Yibo Wang , Wenchao Qi , Songtao Ding , Xia Zhang","doi":"10.1016/j.jag.2024.104213","DOIUrl":"10.1016/j.jag.2024.104213","url":null,"abstract":"<div><div>Effective management of corn straw and stubble is critical in conservation tillage, as it impacts soil health and productivity. However, accurate classification of different types of straw cover has been hindered by their similar spectral and spatial characteristics and the low spatial resolution of hyperspectral satellite imagery. Moreover, traditional convolution neural network (CNN)-based methods, which rely on first-order statistics for feature extraction, often struggle to extract distinguishable features of highly similar objects effectively, thereby reducing classification accuracy. In this study, a second-order statistical-feature extraction algorithm based on CNN that uses fused multispectral and hyperspectral data was tested for its ability to classify types of straw cover. In the first step, coupled non-negative matrix factorization (CNMF) was used to fuse hyperspectral and multispectral images effectively, thereby enhancing the spatial resolution of the hyperspectral data. In this study, we integrated pointwise convolution (PWC), depthwise convolution (DWC), and a distance covariance matrix (DCM) to form a mixed convolution and DCM (MCDCM) network; we used this to extract and integrate deep spectral–spatial features of the hyperspectral images. Our experimental results show that the MCDCM network significantly improved classification accuracy compared to traditional methods, with accuracy rates for the different straw-cover types exceeding 90% and overall accuracy reaching 98.26%. The fused image also exhibited better preservation of feature edges and contours. The accurate identification of corn-straw-cover types achieved with the proposed MCDCM method is a major step in optimizing conservation-farming practices, improving soil fertility and farm productivity, and supporting sustainable ecological development.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"134 ","pages":"Article 104213"},"PeriodicalIF":7.6,"publicationDate":"2024-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142445207","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 : 2024-10-17DOI: 10.1016/j.jag.2024.104215
Feng Lin , Xie Hu , Yiling Lin , Yao Li , Yang Liu , Dongmei Li
Knowledge about the impact craters on rocky planets is crucial for understanding the evolutionary history of the universe. Compared to traditional visual interpretation, deep learning approaches have improved the efficiency of crater detection. However, single-source data and divergent data quality limit the accuracy of crater detection. In this study, we focus on valuable features in multi-modal remote sensing data from Chang’e lunar exploration mission and propose an Attention-based Dual-branch Segmentation Network (ADSNet). First, we use ADSNet to extract the multi-modal features via a dual-branch encoder. Second, we introduce a novel attention for data fusion where the features from the auxiliary modality are weighted by a scoring function and then being fused with those from the primary modality. After fusion, the features are transferred to the decoder through skip connection. Lastly, high-accuracy crater detection is achieved based on the learned multi-modal data features through semantic segmentation. Our results demonstrate that ADSNet outperforms other baseline models in many metrics such as IoU and F1 score. ADSNet is an effective approach to leverage multi-modal remote sensing data in geomorphological feature detection on rocky planets in general.
了解岩石行星上的撞击坑对于了解宇宙的演化历史至关重要。与传统的视觉判读相比，深度学习方法提高了陨石坑检测的效率。然而，单一来源的数据和不同的数据质量限制了陨石坑检测的准确性。在本研究中，我们聚焦嫦娥探月任务多模态遥感数据中的有价值特征，提出了基于注意力的双分支分割网络（ADSNet）。首先，我们使用 ADSNet 通过双分支编码器提取多模态特征。其次，我们为数据融合引入了一种新的注意力，即通过评分函数对来自辅助模态的特征进行加权，然后与来自主模态的特征进行融合。融合后，通过跳接将特征传输到解码器。最后，根据学习到的多模态数据特征，通过语义分割实现高精度的火山口检测。我们的研究结果表明，ADSNet 在 IoU 和 F1 分数等许多指标上都优于其他基线模型。ADSNet 是利用多模态遥感数据进行岩质行星地貌特征检测的有效方法。
{"title":"Dual-branch multi-modal convergence network for crater detection using Chang’e image","authors":"Feng Lin , Xie Hu , Yiling Lin , Yao Li , Yang Liu , Dongmei Li","doi":"10.1016/j.jag.2024.104215","DOIUrl":"10.1016/j.jag.2024.104215","url":null,"abstract":"<div><div>Knowledge about the impact craters on rocky planets is crucial for understanding the evolutionary history of the universe. Compared to traditional visual interpretation, deep learning approaches have improved the efficiency of crater detection. However, single-source data and divergent data quality limit the accuracy of crater detection. In this study, we focus on valuable features in multi-modal remote sensing data from Chang’e lunar exploration mission and propose an Attention-based Dual-branch Segmentation Network (ADSNet). First, we use ADSNet to extract the multi-modal features via a dual-branch encoder. Second, we introduce a novel attention for data fusion where the features from the auxiliary modality are weighted by a scoring function and then being fused with those from the primary modality. After fusion, the features are transferred to the decoder through skip connection. Lastly, high-accuracy crater detection is achieved based on the learned multi-modal data features through semantic segmentation. Our results demonstrate that ADSNet outperforms other baseline models in many metrics such as IoU and F1 score. ADSNet is an effective approach to leverage multi-modal remote sensing data in geomorphological feature detection on rocky planets in general.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"134 ","pages":"Article 104215"},"PeriodicalIF":7.6,"publicationDate":"2024-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142444550","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 : 2024-10-16DOI: 10.1016/j.jag.2024.104221
Gabriel Díaz-Ireland , Derya Gülçin , Aida López-Sánchez , Eduardo Pla , John Burton , Javier Velázquez
This study examines the effectiveness of five deep learning models—ViTb-19, SwinV2-t, VGG-16, ResNet-50, and DenseNet-121—in distinguishing different vegetation types in the protected grasslands of Castilla y León region, Spain, following the guidelines of the Natura 92/43/CEE directive. Among the models, ResNet-50 achieved the highest weighted overall accuracy (OA) of 0.95, closely followed by SwinV2-t with an OA of 0.94, demonstrating their strong ability to detect complex patterns in satellite imagery. DenseNet-121 also performed competitively with a weighted OA of 0.93, while ViTb-19 and VGG-16 showed slightly lower performance. SwinV2-t, a transformer-based model, outperformed traditional CNN architectures in data-rich classes but faced challenges in classifying habitats with limited representation. Consequently, this study identifies these challenges that conventional transformer architectures pose in classifying certain habitats with limited representation and intricate features. Highlighting the advantages of deep learning technologies for environmental monitoring and conservation, the study provides important insights for adjusting neural network architectures for effective habitat classification. This suggests the necessity of selecting appropriate architectures such as SwinV2-t and ResNet50 to to effectively address the intricate requirements of satellite imagery analysis.
{"title":"Classification of protected grassland habitats using deep learning architectures on Sentinel-2 satellite imagery data","authors":"Gabriel Díaz-Ireland , Derya Gülçin , Aida López-Sánchez , Eduardo Pla , John Burton , Javier Velázquez","doi":"10.1016/j.jag.2024.104221","DOIUrl":"10.1016/j.jag.2024.104221","url":null,"abstract":"<div><div>This study examines the effectiveness of five deep learning models—ViTb-19, SwinV2-t, VGG-16, ResNet-50, and DenseNet-121—in distinguishing different vegetation types in the protected grasslands of Castilla y León region, Spain, following the guidelines of the Natura 92/43/CEE directive. Among the models, ResNet-50 achieved the highest weighted overall accuracy (OA) of 0.95, closely followed by SwinV2-t with an OA of 0.94, demonstrating their strong ability to detect complex patterns in satellite imagery. DenseNet-121 also performed competitively with a weighted OA of 0.93, while ViTb-19 and VGG-16 showed slightly lower performance. SwinV2-t, a transformer-based model, outperformed traditional CNN architectures in data-rich classes but faced challenges in classifying habitats with limited representation. Consequently, this study identifies these challenges that conventional transformer architectures pose in classifying certain habitats with limited representation and intricate features. Highlighting the advantages of deep learning technologies for environmental monitoring and conservation, the study provides important insights for adjusting neural network architectures for effective habitat classification. This suggests the necessity of selecting appropriate architectures such as SwinV2-t and ResNet50 to to effectively address the intricate requirements of satellite imagery analysis.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"134 ","pages":"Article 104221"},"PeriodicalIF":7.6,"publicationDate":"2024-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142442781","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 : 2024-10-16DOI: 10.1016/j.jag.2024.104202
Ayushi Gupta , Prashant K. Srivastava , Karuna Shanker , K. Chandra Sekar
Precise spatial mapping of individual species using hyperspectral data is crucial for effective forest management and policy-making. This study focuses on Rhododendron arboreum, known for its medicinal properties attributed to the flavonoid Quercitrin. Sample data and spectroradiometer data were collected from the complex terrain of the Kumaon region in the Himalayas. Hyperspectral data, which includes signal variations based on biophysical and biochemical properties along with noise, were preprocessed using filtering techniques to enhance signal clarity by removing noise. Smoothing techniques were applied to remove noisy bands from the spectra, such as the Savitzky-Golay filter for reduced least square fit complexity and the Average Mean filter for taking mean spectral values. Subsequently, Spectral Analysis (SA) techniques, including first derivative, second derivative, and continuum removal, were employed. These mathematical transformations highlighted absorption troughs and determined the effect of Quercitrin on spectral wavelengths. Principal Component Analysis (PCA) was used to identify the most relevant bands related to Quercitrin. Additionally, regression analysis was applied on resampled spectral data, selected significant wavelengths based on variable importance values, pinpointing the most prominent wavelengths: 1196, 1229, 1328, 1383, 1425, 1636, 1661, 1699, 1785, and 1715 nm. Over 50 two-band combination indices were tested, and those with p-values less than 0.05 were deemed significant. For the development of prediction model, Machine Learning (ML) algorithms, including Support Vector Machine (SVM), Relevance Vector Machine (RVM), Random Forest (RF), and Artificial Neural Network (ANN), were applied. The Random Forest model, which splits input data into trees to simulate the best model based on observed values, demonstrated high effectiveness in predicting Quercitrin levels, achieving a training correlation of 0.864 and a testing correlation of 0.570. Hence RF proved to be a best technique of band selection as well as robust for Quercitrin prediction. This methodological approach highlights the importance of advanced data processing and analysis techniques in remote sensing applications for forest phytochemical prediction.
{"title":"Quantification and mapping of medicinally important Quercitrin compound using hyperspectral imaging and machine learning","authors":"Ayushi Gupta , Prashant K. Srivastava , Karuna Shanker , K. Chandra Sekar","doi":"10.1016/j.jag.2024.104202","DOIUrl":"10.1016/j.jag.2024.104202","url":null,"abstract":"<div><div>Precise spatial mapping of individual species using hyperspectral data is crucial for effective forest management and policy-making. This study focuses on <em>Rhododendron arboreum</em>, known for its medicinal properties attributed to the flavonoid Quercitrin. Sample data and spectroradiometer data were collected from the complex terrain of the Kumaon region in the Himalayas. Hyperspectral data, which includes signal variations based on biophysical and biochemical properties along with noise, were preprocessed using filtering techniques to enhance signal clarity by removing noise. Smoothing techniques were applied to remove noisy bands from the spectra, such as the Savitzky-Golay filter for reduced least square fit complexity and the Average Mean filter for taking mean spectral values. Subsequently, Spectral Analysis (SA) techniques, including first derivative, second derivative, and continuum removal, were employed. These mathematical transformations highlighted absorption troughs and determined the effect of Quercitrin on spectral wavelengths. Principal Component Analysis (PCA) was used to identify the most relevant bands related to Quercitrin. Additionally, regression analysis was applied on resampled spectral data, selected significant wavelengths based on variable importance values, pinpointing the most prominent wavelengths: 1196, 1229, 1328, 1383, 1425, 1636, 1661, 1699, 1785, and 1715 nm. Over 50 two-band combination indices were tested, and those with p-values less than 0.05 were deemed significant. For the development of prediction model, Machine Learning (ML) algorithms, including Support Vector Machine (SVM), Relevance Vector Machine (RVM), Random Forest (RF), and Artificial Neural Network (ANN), were applied. The Random Forest model, which splits input data into trees to simulate the best model based on observed values, demonstrated high effectiveness in predicting Quercitrin levels, achieving a training correlation of 0.864 and a testing correlation of 0.570. Hence RF proved to be a best technique of band selection as well as robust for Quercitrin prediction. This methodological approach highlights the importance of advanced data processing and analysis techniques in remote sensing applications for forest phytochemical prediction.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"134 ","pages":"Article 104202"},"PeriodicalIF":7.6,"publicationDate":"2024-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142442782","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 : 2024-10-16DOI: 10.1016/j.jag.2024.104203
Sergios-Anestis Kefalidis , Dharmen Punjani , Eleni Tsalapati , Konstantinos Plas , Maria-Aggeliki Pollali , Pierre Maret , Manolis Koubarakis
We present the question answering engine GeoQA2 which is able to answer geospatial questions over the union of knowledge graphs YAGO2 and YAGO2geo. We also present the dataset GeoQuestions1089 which consists of 1089 natural language questions, their corresponding SPARQL or GeoSPARQL queries and their answers over the union of the same knowledge graphs. We use this dataset to compare the effectiveness of GeoQA2 and the system of Hamzei et al. 2022 and make it publicly available to be used by other researchers. Our evaluation shows that although the engine GeoQA2 performs better than the engine of Hamzei et al. 2022, both engines have ample room for improvement in their question answering performance.
{"title":"The question answering system GeoQA2 and a new benchmark for its evaluation","authors":"Sergios-Anestis Kefalidis , Dharmen Punjani , Eleni Tsalapati , Konstantinos Plas , Maria-Aggeliki Pollali , Pierre Maret , Manolis Koubarakis","doi":"10.1016/j.jag.2024.104203","DOIUrl":"10.1016/j.jag.2024.104203","url":null,"abstract":"<div><div>We present the question answering engine GeoQA2 which is able to answer geospatial questions over the union of knowledge graphs YAGO2 and YAGO2geo. We also present the dataset <span>GeoQuestions1089</span> which consists of 1089 natural language questions, their corresponding SPARQL or GeoSPARQL queries and their answers over the union of the same knowledge graphs. We use this dataset to compare the effectiveness of GeoQA2 and the system of Hamzei et al. 2022 and make it publicly available to be used by other researchers. Our evaluation shows that although the engine GeoQA2 performs better than the engine of Hamzei et al. 2022, both engines have ample room for improvement in their question answering performance.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"134 ","pages":"Article 104203"},"PeriodicalIF":7.6,"publicationDate":"2024-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142442783","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}
This article provides a critical review of progress, challenges, emerging gaps as well as future recommendations on the remote sensing of grass quality during the senescence phenological stage. The study adopted a critical approach and analysed nineteen peer-reviewed articles which were retrieved from Scopus, Web of Science, and Institute of Electrical and Electronics Engineers using key search words. Overall, the results showed that remote sensing has been used to map the quality elements of senescent grass as determined by the concentration of macronutrients, fibre content and biochemical variables such as chlorophyll content. Successful estimation of these variables was achieved using ground-based, airborne, and spaceborne sensors. Nonetheless, this critical review demonstrates that the choice of suitable remote sensing sensor for mapping grass quality attributes during senescence depends on the trade-offs between sensing characteristics, spatial coverage, and data availability. Critical assessment of retrieved literature showed that wavebands located in the red, red-edge, and shortwave infrared regions had the highest sensitivity to senescent grass quality constituents. Remote sensing algorithms reported within the retrieved studies include multivariate analysis techniques, machine learning algorithms and radiative transfer models. Although these are associated with different performances in different settings and vary in their strengths and limitations, it is argued that there is no specific algorithm that is suitable for a specific variable in the context of characterizing grass quality during the senescence period. In this regard, there is a need to assess and ascertain based on factors such as sample size and number of explanatory variables used which affect their accuracy. It is concluded that despite the noted progress in sensor capabilities, the new generation of space borne hyperspectral sensors such as Environmental Mapping and Analysis Program provides untapped prospects to advance the scientific inquiry for remote sensing grass quality during the senescence stage. The review therefore recommends that further research in this field can also consider the utility of such sensor systems, which are readily accessible to enhance the discreet detection of grass quality attributes over space and time. Precise detection of subtle changes in grass nutritional quality during the senescence phenological stage is essential for monitoring forage provisioning ecosystem services.
本文对衰老物候期草质遥感的进展、挑战、新出现的差距以及未来建议进行了批判性评述。研究采用了批判性方法，分析了从 Scopus、Web of Science 和电气与电子工程师学会使用关键词检索到的 19 篇同行评审文章。总之，研究结果表明，遥感技术已被用于绘制衰老草的质量要素图，这些要素由常量营养素的浓度、纤维含量和叶绿素含量等生化变量决定。利用地面、机载和空间传感器成功地估算了这些变量。然而，本评论表明，选择合适的遥感传感器来绘制衰老期草地质量属性图取决于在传感特性、空间覆盖范围和数据可用性之间进行权衡。对检索到的文献进行的严格评估表明，位于红色、红边和短波红外区域的波段对衰老草质成分的敏感度最高。检索到的研究报告中提到的遥感算法包括多元分析技术、机器学习算法和辐射传递模型。虽然这些算法在不同的环境下有不同的性能，其优势和局限性也各不相同，但在描述衰老期草地质量特征的背景下，没有一种特定的算法适合特定的变量。在这方面，有必要根据影响其准确性的因素（如样本大小和所用解释变量的数量）进行评估和确定。综述认为，尽管在传感器能力方面取得了显著进步，但新一代空间超光谱传感器（如环境制图与分析计划）为推进衰老期草质遥感科学研究提供了尚未开发的前景。因此，综述建议在这一领域开展进一步研究时，也可考虑这类传感器系统的效用，因为它们可随时用于加强对草地质量属性在空间和时间上的谨慎检测。精确检测衰老物候期草地营养质量的细微变化对于监测草料供应生态系统服务至关重要。
{"title":"A critical review of literature on remote sensing grass quality during the senescence phenological stage","authors":"Anita Masenyama , Onisimo Mutanga , Mbulisi Sibanda , Timothy Dube","doi":"10.1016/j.jag.2024.104211","DOIUrl":"10.1016/j.jag.2024.104211","url":null,"abstract":"<div><div>This article provides a critical review of progress, challenges, emerging gaps as well as future recommendations on the remote sensing of grass quality during the senescence phenological stage. The study adopted a critical approach and analysed nineteen peer-reviewed articles which were retrieved from Scopus, Web of Science, and Institute of Electrical and Electronics Engineers using key search words. Overall, the results showed that remote sensing has been used to map the quality elements of senescent grass as determined by the concentration of macronutrients, fibre content and biochemical variables such as chlorophyll content. Successful estimation of these variables was achieved using ground-based, airborne, and spaceborne sensors. Nonetheless, this critical review demonstrates that the choice of suitable remote sensing sensor for mapping grass quality attributes during senescence depends on the trade-offs between sensing characteristics, spatial coverage, and data availability. Critical assessment of retrieved literature showed that wavebands located in the red, red-edge, and shortwave infrared regions had the highest sensitivity to senescent grass quality constituents. Remote sensing algorithms reported within the retrieved studies include multivariate analysis techniques, machine learning algorithms and radiative transfer models. Although these are associated with different performances in different settings and vary in their strengths and limitations, it is argued that there is no specific algorithm that is suitable for a specific variable in the context of characterizing grass quality during the senescence period. In this regard, there is a need to assess and ascertain based on factors such as sample size and number of explanatory variables used which affect their accuracy. It is concluded that despite the noted progress in sensor capabilities, the new generation of space borne hyperspectral sensors such as Environmental Mapping and Analysis Program provides untapped prospects to advance the scientific inquiry for remote sensing grass quality during the senescence stage. The review therefore recommends that further research in this field can also consider the utility of such sensor systems, which are readily accessible to enhance the discreet detection of grass quality attributes over space and time. Precise detection of subtle changes in grass nutritional quality during the senescence phenological stage is essential for monitoring forage provisioning ecosystem services.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"134 ","pages":"Article 104211"},"PeriodicalIF":7.6,"publicationDate":"2024-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142432769","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}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号