The images captured from SAR sensors are inherently weakened by speckle noise. The SAR image processing community targeted this problem with many feature-based filters. Since SAR images are low-contrast images, edge retention is the most crucial aspect to consider. This helps in the efficient retrieval of information. This paper provides a two-step edge-preserving homomorphic SAR image despeckling technique that implements a guided filter as the first step, and a modified method of noise thresholding using the bivariate shrinkage rule and canny edge operator in the Discrete Orthonormal Stockwell Transform (DOST) domain as the second step. The use of a canny edge operator improves overall edge preservation after despeckling. The use of noise thresholding delivers the highest level of speckle reduction in the DOST domain. The detected edges are added to the residual part obtained after removing the noise to produce more informative content. According to several qualitative and quantitative criteria, the suggested approach is compared to some of the newest despeckling methods. The execution time of the proposed method is around 7.2679 seconds. Upon conducting qualitative and quantitative analysis, it has been determined that the proposed method surpasses all other despeckling methods that were compared.
{"title":"An Algorithmic Approach towards Remote Sensing Imagery Data Restoration Using Guided Filters in Real-Time Applications","authors":"Prabhishek Singh, Manoj Diwakar, Debjani Ghosh, Ankit Vidyarthi, Deepak Gupta, Punit Gupta","doi":"10.1080/07038992.2023.2257323","DOIUrl":"https://doi.org/10.1080/07038992.2023.2257323","url":null,"abstract":"The images captured from SAR sensors are inherently weakened by speckle noise. The SAR image processing community targeted this problem with many feature-based filters. Since SAR images are low-contrast images, edge retention is the most crucial aspect to consider. This helps in the efficient retrieval of information. This paper provides a two-step edge-preserving homomorphic SAR image despeckling technique that implements a guided filter as the first step, and a modified method of noise thresholding using the bivariate shrinkage rule and canny edge operator in the Discrete Orthonormal Stockwell Transform (DOST) domain as the second step. The use of a canny edge operator improves overall edge preservation after despeckling. The use of noise thresholding delivers the highest level of speckle reduction in the DOST domain. The detected edges are added to the residual part obtained after removing the noise to produce more informative content. According to several qualitative and quantitative criteria, the suggested approach is compared to some of the newest despeckling methods. The execution time of the proposed method is around 7.2679 seconds. Upon conducting qualitative and quantitative analysis, it has been determined that the proposed method surpasses all other despeckling methods that were compared.","PeriodicalId":48843,"journal":{"name":"Canadian Journal of Remote Sensing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136363828","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-10DOI: 10.1080/07038992.2023.2247095
Sergey V. Samsonov, Wanpeng Feng
A fully automated processing system for measuring long-term ground deformation time series and deformation rates frame-by-frame using DInSAR processing technique was developed at the Canada Center for Remote Sensing. Ground deformation rates from 2017 to 2023 were computed over a large territory of North America and Eurasia from more than 220,000 readily available Sentinel-1 images, and the performance and shortcomings of the developed processing system were analyzed. Here, we present the processing methodology and several examples of deformation rate maps and time series produced with this automated system. Examples include the deformation of slow- moving deep-seated landslides in two regions of Canada, subsidence at the Komsomolskoe oil field in the Russian Arctic, the Tengiz oil field in Kazakhstan, multiple large subsiding regions and landslides in northwestern Iran, and two large subsiding regions in the Yellow River Delta and Xinjiang, China. Many deformation processes observed in these deformation rate maps, including large landslides, have previously been unknown to the research community. Systematic radar penetration depth changes were observed in multiple regions and were investigate in detail for 1 Eurasian region. Computed deformation rates for North America and Eurasia are available to the research community and can be downloaded from the data repository.
{"title":"Deformation Retrievals for North America and Eurasia from Sentinel-1 DInSAR: Big Data Approach, Processing Methodology and Challenges","authors":"Sergey V. Samsonov, Wanpeng Feng","doi":"10.1080/07038992.2023.2247095","DOIUrl":"https://doi.org/10.1080/07038992.2023.2247095","url":null,"abstract":"A fully automated processing system for measuring long-term ground deformation time series and deformation rates frame-by-frame using DInSAR processing technique was developed at the Canada Center for Remote Sensing. Ground deformation rates from 2017 to 2023 were computed over a large territory of North America and Eurasia from more than 220,000 readily available Sentinel-1 images, and the performance and shortcomings of the developed processing system were analyzed. Here, we present the processing methodology and several examples of deformation rate maps and time series produced with this automated system. Examples include the deformation of slow- moving deep-seated landslides in two regions of Canada, subsidence at the Komsomolskoe oil field in the Russian Arctic, the Tengiz oil field in Kazakhstan, multiple large subsiding regions and landslides in northwestern Iran, and two large subsiding regions in the Yellow River Delta and Xinjiang, China. Many deformation processes observed in these deformation rate maps, including large landslides, have previously been unknown to the research community. Systematic radar penetration depth changes were observed in multiple regions and were investigate in detail for 1 Eurasian region. Computed deformation rates for North America and Eurasia are available to the research community and can be downloaded from the data repository.","PeriodicalId":48843,"journal":{"name":"Canadian Journal of Remote Sensing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-08-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135597818","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-01-02DOI: 10.1080/07038992.2023.2264395
Wesley Van Wychen, Courtney Bayer, Luke Copland, Erika Brummell, Christine Dow
Here we use high resolution (5 m) Radarsat Constellation Mission (RCM) imagery acquired in winters 2022 and 2023 to determine motion across glaciers of the St. Elias Icefield in Yukon/Alaska. Our regional velocity mapping largely conforms with previous studies, with faster motion (>600 m/yr) for the glaciers originating in the Yukon that drain southward and westward to the coast of Alaska and relatively slower motion (100–400 m/yr) for the land terminating glaciers that drain eastward and northeastward and stay within the Yukon. We also identify two new glacier surges within the icefields: the surge of Nàłùdäy (Lowell) Glacier in Winter 2022, and Chitina Glacier in Winter 2023, and track the progression of each surge from January to March utilizing ∼4-day repeat RCM imagery. To evaluate the quality of RCM-derived velocities, we compare our results with 50 simultaneous measurements at three on-ice dGPS stations located on two Yukon glaciers and find the average absolute difference between measurements to be 6.6 m/yr. Our results demonstrate the utility of RCM data to determine glacier motion across large regions with complex topography, to support process-based studies of fast flowing and surge-type glaciers and continue the legacy of velocity products derived from the Radarsat-2 mission.
{"title":"Radarsat Constellation Mission Derived Winter Glacier Velocities for the St. Elias Icefield, Yukon/Alaska: 2022 and 2023","authors":"Wesley Van Wychen, Courtney Bayer, Luke Copland, Erika Brummell, Christine Dow","doi":"10.1080/07038992.2023.2264395","DOIUrl":"https://doi.org/10.1080/07038992.2023.2264395","url":null,"abstract":"Here we use high resolution (5 m) Radarsat Constellation Mission (RCM) imagery acquired in winters 2022 and 2023 to determine motion across glaciers of the St. Elias Icefield in Yukon/Alaska. Our regional velocity mapping largely conforms with previous studies, with faster motion (>600 m/yr) for the glaciers originating in the Yukon that drain southward and westward to the coast of Alaska and relatively slower motion (100–400 m/yr) for the land terminating glaciers that drain eastward and northeastward and stay within the Yukon. We also identify two new glacier surges within the icefields: the surge of Nàłùdäy (Lowell) Glacier in Winter 2022, and Chitina Glacier in Winter 2023, and track the progression of each surge from January to March utilizing ∼4-day repeat RCM imagery. To evaluate the quality of RCM-derived velocities, we compare our results with 50 simultaneous measurements at three on-ice dGPS stations located on two Yukon glaciers and find the average absolute difference between measurements to be 6.6 m/yr. Our results demonstrate the utility of RCM data to determine glacier motion across large regions with complex topography, to support process-based studies of fast flowing and surge-type glaciers and continue the legacy of velocity products derived from the Radarsat-2 mission.","PeriodicalId":48843,"journal":{"name":"Canadian Journal of Remote Sensing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-01-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135798449","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-01-02DOI: 10.1080/07038992.2023.2257331
Yan Duan, Shaojie Bai, Limin Liu, Guangwei Wang
Polarimetric synthetic aperture radar (PolSAR) images are classified mainly according to the backscattering information of ground objects. For regions with complex backscattering information, misclassification is easy to occur, which leads to challenges in improving the classification accuracy of the PolSAR image. Given this situation, this paper combines the Deep Learning Model and traditional classifiers to classify PolSAR image. First, the Convolution Neural Network (CNN) was used to classify the PolSAR image and according to the category prediction probability of pixels, the key pixels easily misclassified are located. Then, the adaptive boosting (AdaBoost) algorithm combined the three shallow classifiers (the Support Vector Machine (SVM), the Wishart and the Decision Tree classifier) into strong classifiers to reclassify the key pixels. Finally, the labels of key pixels and other pixels are output as the final classification result. Experiments on two PolSAR images show that the proposed method can improve classification performance and obtain better classification results.
{"title":"A Novel Classification Method for PolSAR Image Combining the Deep Learning Model and Adaptive Boosting of Shallow Classifiers","authors":"Yan Duan, Shaojie Bai, Limin Liu, Guangwei Wang","doi":"10.1080/07038992.2023.2257331","DOIUrl":"https://doi.org/10.1080/07038992.2023.2257331","url":null,"abstract":"Polarimetric synthetic aperture radar (PolSAR) images are classified mainly according to the backscattering information of ground objects. For regions with complex backscattering information, misclassification is easy to occur, which leads to challenges in improving the classification accuracy of the PolSAR image. Given this situation, this paper combines the Deep Learning Model and traditional classifiers to classify PolSAR image. First, the Convolution Neural Network (CNN) was used to classify the PolSAR image and according to the category prediction probability of pixels, the key pixels easily misclassified are located. Then, the adaptive boosting (AdaBoost) algorithm combined the three shallow classifiers (the Support Vector Machine (SVM), the Wishart and the Decision Tree classifier) into strong classifiers to reclassify the key pixels. Finally, the labels of key pixels and other pixels are output as the final classification result. Experiments on two PolSAR images show that the proposed method can improve classification performance and obtain better classification results.","PeriodicalId":48843,"journal":{"name":"Canadian Journal of Remote Sensing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-01-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135799414","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-01-02DOI: 10.1080/07038992.2023.2247091
Benoit Montpetit, Benjamin Deschamps, Joshua King, Jason Duffe
{"title":"Assessing the Parameterization of RADARSAT-2 Dual-polarized ScanSAR Scenes on the Accuracy of a Convolutional Neural Network for Sea Ice Classification: Case Study over Coronation Gulf, Canada","authors":"Benoit Montpetit, Benjamin Deschamps, Joshua King, Jason Duffe","doi":"10.1080/07038992.2023.2247091","DOIUrl":"https://doi.org/10.1080/07038992.2023.2247091","url":null,"abstract":"","PeriodicalId":48843,"journal":{"name":"Canadian Journal of Remote Sensing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-01-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135799634","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-01-02DOI: 10.1080/07038992.2023.2256895
Ramon Melser, Nicholas C. Coops, Michael A. Wulder, Chris Derksen
Understanding the processes driving terrestrial vegetation productivity dynamics in boreal ecosystems is critical for accurate assessments of carbon dynamics. Monitoring these dynamics typically requires a fusion of broad-scale remote sensing observations, climate information and other geospatial data inputs, which often have unknown errors, are difficult to obtain, or limit spatial and temporal resolutions of productivity estimates. The past decade has seen notable advances in technologies and the diversity of observed wavelengths from remote sensing instruments, offering new insights on vegetation carbon dynamics. In this communication, we review key current approaches for modeling terrestrial vegetation productivity, followed by a discussion on new remote sensing instruments and derived products including Sentinel-3 Land Surface Temperature, freeze & thaw state from the Soil Moisture & Ocean Salinity (SMOS) mission, and soil moisture from the Soil Moisture Active Passive (SMAP) mission. We outline how these products can improve the spatial detail and temporal representation of boreal productivity estimates driven entirely by a fusion of remote sensing observations. We conclude with a demonstration of how these different elements can be integrated across key land cover types in the Hudson plains, an extensive wetland-dominated region of the Canadian boreal, and provide recommendations for future model development.
{"title":"Multi-Source Remote Sensing Based Modeling of Vegetation Productivity in the Boreal: Issues & Opportunities","authors":"Ramon Melser, Nicholas C. Coops, Michael A. Wulder, Chris Derksen","doi":"10.1080/07038992.2023.2256895","DOIUrl":"https://doi.org/10.1080/07038992.2023.2256895","url":null,"abstract":"Understanding the processes driving terrestrial vegetation productivity dynamics in boreal ecosystems is critical for accurate assessments of carbon dynamics. Monitoring these dynamics typically requires a fusion of broad-scale remote sensing observations, climate information and other geospatial data inputs, which often have unknown errors, are difficult to obtain, or limit spatial and temporal resolutions of productivity estimates. The past decade has seen notable advances in technologies and the diversity of observed wavelengths from remote sensing instruments, offering new insights on vegetation carbon dynamics. In this communication, we review key current approaches for modeling terrestrial vegetation productivity, followed by a discussion on new remote sensing instruments and derived products including Sentinel-3 Land Surface Temperature, freeze & thaw state from the Soil Moisture & Ocean Salinity (SMOS) mission, and soil moisture from the Soil Moisture Active Passive (SMAP) mission. We outline how these products can improve the spatial detail and temporal representation of boreal productivity estimates driven entirely by a fusion of remote sensing observations. We conclude with a demonstration of how these different elements can be integrated across key land cover types in the Hudson plains, an extensive wetland-dominated region of the Canadian boreal, and provide recommendations for future model development.","PeriodicalId":48843,"journal":{"name":"Canadian Journal of Remote Sensing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-01-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135799420","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-01-02DOI: 10.1080/07038992.2023.2215333
Amir M. Chegoonian, Nima Pahlevan, Kiana Zolfaghari, Peter R. Leavitt, John-Mark Davies, Helen M. Baulch, Claude R. Duguay
Remote retrieval of near-surface chlorophyll-a (Chla) concentration in small inland waters is challenging due to substantial optical interferences of various water constituents and uncertainties in the atmospheric correction (AC) process. Although various algorithms have been developed to estimate Chla from moderate-resolution terrestrial missions (∼10–60 m), the production of both accurate distribution maps and time series of Chla has proven challenging, limiting the use of remote analyses for lake monitoring. Here, we develop a support vector regression (SVR) model, which uses satellite-derived remote-sensing reflectance spectra (Rrsδ) from Sentinel-2 and Landsat-8 images as input for Chla retrieval in a representative eutrophic prairie lake, Buffalo Pound Lake (BPL), Saskatchewan, Canada. Validated against in situ Chla from seven ice-free seasons (N ∼ 200; 2014–2020), the SVR model outperformed both locally tuned, Rrsδ-fed empirical models (Normalized Difference Chlorophyll Index, 2- and 3-band, and OC3) and Mixture Density Networks (MDNs) by 15–65%, while exhibiting comparable performance to a locally trained MDN, with an error of ∼35%. Comparison of Chla retrieval models, AC processors (iCOR, ACOLITE), and radiometric products (Rayleigh-corrected, surface, and top-of-atmosphere reflectance) showed that the best Chla maps and optimal time series (up to 100 mg m−3) were produced using a coupled SVR-iCOR system.
{"title":"Comparative Analysis of Empirical and Machine Learning Models for Chl<i>a</i> Extraction Using Sentinel-2 and Landsat OLI Data: Opportunities, Limitations, and Challenges","authors":"Amir M. Chegoonian, Nima Pahlevan, Kiana Zolfaghari, Peter R. Leavitt, John-Mark Davies, Helen M. Baulch, Claude R. Duguay","doi":"10.1080/07038992.2023.2215333","DOIUrl":"https://doi.org/10.1080/07038992.2023.2215333","url":null,"abstract":"Remote retrieval of near-surface chlorophyll-a (Chla) concentration in small inland waters is challenging due to substantial optical interferences of various water constituents and uncertainties in the atmospheric correction (AC) process. Although various algorithms have been developed to estimate Chla from moderate-resolution terrestrial missions (∼10–60 m), the production of both accurate distribution maps and time series of Chla has proven challenging, limiting the use of remote analyses for lake monitoring. Here, we develop a support vector regression (SVR) model, which uses satellite-derived remote-sensing reflectance spectra (Rrsδ) from Sentinel-2 and Landsat-8 images as input for Chla retrieval in a representative eutrophic prairie lake, Buffalo Pound Lake (BPL), Saskatchewan, Canada. Validated against in situ Chla from seven ice-free seasons (N ∼ 200; 2014–2020), the SVR model outperformed both locally tuned, Rrsδ-fed empirical models (Normalized Difference Chlorophyll Index, 2- and 3-band, and OC3) and Mixture Density Networks (MDNs) by 15–65%, while exhibiting comparable performance to a locally trained MDN, with an error of ∼35%. Comparison of Chla retrieval models, AC processors (iCOR, ACOLITE), and radiometric products (Rayleigh-corrected, surface, and top-of-atmosphere reflectance) showed that the best Chla maps and optimal time series (up to 100 mg m−3) were produced using a coupled SVR-iCOR system.","PeriodicalId":48843,"journal":{"name":"Canadian Journal of Remote Sensing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-01-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135655526","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-01-02DOI: 10.1080/07038992.2023.2246158
Jianshang Liao, Liguo Wang, Genping Zhao
Gabor filter is widely used to extract spatial texture features of hyperspectral images (HSI) for HSI classification; however, a single Gabor filter cannot obtain the complete image features. In the paper, we propose an HSI classification method that combines the Gabor filter (GF) and domain-transformation standard convolution (DTNC) filter. First, we use the Gabor filter to extract spatial texture features from the first two principal components of the dimensionality-reduction HSI with PCA. Second, we use the DTNC filter to extract spatial correlation features from HSI in all bands. Finally, the Large Margin Distribution Machine (LDM) uses the linear fusion of the two kinds of spatial features to classify HSI. The experimental results show that the classification accuracy of Indian Pines, Pavia University, and Kennedy Space Center data sets is 96.64, 98.23, and 98.95% with only 4, 3, and 6% training samples, respectively; and these accuracies are 2–20% higher than the other tested methods. Compared with the hyperspectral information based on SVM, EPF, IFRF, PCA-EPFs, LDM-FL, and GFDN method, the proposed method, GFDTNCLDM, significantly improves the accuracy of HSI classification.
{"title":"Hyperspectral Image Classification Based on the Gabor Feature with Correlation Information","authors":"Jianshang Liao, Liguo Wang, Genping Zhao","doi":"10.1080/07038992.2023.2246158","DOIUrl":"https://doi.org/10.1080/07038992.2023.2246158","url":null,"abstract":"Gabor filter is widely used to extract spatial texture features of hyperspectral images (HSI) for HSI classification; however, a single Gabor filter cannot obtain the complete image features. In the paper, we propose an HSI classification method that combines the Gabor filter (GF) and domain-transformation standard convolution (DTNC) filter. First, we use the Gabor filter to extract spatial texture features from the first two principal components of the dimensionality-reduction HSI with PCA. Second, we use the DTNC filter to extract spatial correlation features from HSI in all bands. Finally, the Large Margin Distribution Machine (LDM) uses the linear fusion of the two kinds of spatial features to classify HSI. The experimental results show that the classification accuracy of Indian Pines, Pavia University, and Kennedy Space Center data sets is 96.64, 98.23, and 98.95% with only 4, 3, and 6% training samples, respectively; and these accuracies are 2–20% higher than the other tested methods. Compared with the hyperspectral information based on SVM, EPF, IFRF, PCA-EPFs, LDM-FL, and GFDN method, the proposed method, GFDTNCLDM, significantly improves the accuracy of HSI classification.","PeriodicalId":48843,"journal":{"name":"Canadian Journal of Remote Sensing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-01-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135799624","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-01-02DOI: 10.1080/07038992.2023.2255068
Xiaolu Zhang, Zhaoshun Wang, Anlei Wei
As remote sensing images have complex backgrounds and varying object sizes, their semantic segmentation is challenging. This study proposes a multiscale cascaded network (MSCNet) for semantic segmentation. The resolutions employed with respect to the input remote sensing images are 1, 1/2, and 1/4, which represent high, medium, and low resolutions. First, 3 backbone networks extract features with different resolutions. Then, using a multiscale attention network, the fused features are input into the dense atrous spatial pyramid pooling network to obtain multiscale information. The proposed MSCNet introduces multiscale feature extraction and attention mechanism modules suitable for remote sensing land-cover classification. Experiments are performed using the Deepglobe, Vaihingen, and Potsdam datasets; the results are compared with those of the existing classical semantic segmentation networks. The findings indicate that the mean intersection over union (mIoU) of the MSCNet is 4.73% higher than that of DeepLabv3+ with the Deepglobe datasets. For the Vaihingen datasets, the mIoU of the MSCNet is 15.3%, and 6.4% higher than those of a segmented network (SegNet), and DeepLabv3+, respectively. For the Potsdam datasets, the mIoU of the MSCNet is higher than those of a fully convolutional network, Res-U-Net, SegNet, and DeepLabv3+ by 11.18%, 5.89%, 4.78%, and 3.03%, respectively.
{"title":"Multiscale Cascaded Network for the Semantic Segmentation of High-Resolution Remote Sensing Images","authors":"Xiaolu Zhang, Zhaoshun Wang, Anlei Wei","doi":"10.1080/07038992.2023.2255068","DOIUrl":"https://doi.org/10.1080/07038992.2023.2255068","url":null,"abstract":"As remote sensing images have complex backgrounds and varying object sizes, their semantic segmentation is challenging. This study proposes a multiscale cascaded network (MSCNet) for semantic segmentation. The resolutions employed with respect to the input remote sensing images are 1, 1/2, and 1/4, which represent high, medium, and low resolutions. First, 3 backbone networks extract features with different resolutions. Then, using a multiscale attention network, the fused features are input into the dense atrous spatial pyramid pooling network to obtain multiscale information. The proposed MSCNet introduces multiscale feature extraction and attention mechanism modules suitable for remote sensing land-cover classification. Experiments are performed using the Deepglobe, Vaihingen, and Potsdam datasets; the results are compared with those of the existing classical semantic segmentation networks. The findings indicate that the mean intersection over union (mIoU) of the MSCNet is 4.73% higher than that of DeepLabv3+ with the Deepglobe datasets. For the Vaihingen datasets, the mIoU of the MSCNet is 15.3%, and 6.4% higher than those of a segmented network (SegNet), and DeepLabv3+, respectively. For the Potsdam datasets, the mIoU of the MSCNet is higher than those of a fully convolutional network, Res-U-Net, SegNet, and DeepLabv3+ by 11.18%, 5.89%, 4.78%, and 3.03%, respectively.","PeriodicalId":48843,"journal":{"name":"Canadian Journal of Remote Sensing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-01-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135799633","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Le processus de révision par les pairs Toutes les propositions d’article sont soumises à un processus de révision par les pairs avant leur publication. Le comité de rédaction se réserve le droit de choisir ses propres évaluateurs, mais il est utile pour les auteurs de fournir une liste d’au moins quatre spécialistes dans le domaine d’expertise de l’article proposé pouvant agir comme évaluateurs et dont les intérêts de recherche sont suffisamment éloignés du domaine de l’article à évaluer. Prière de fournir leur adresse complète, incluant le numéro de téléphone et l’adresse de courrier électronique. Les auteurs recevront un ensemble de documents comprenant la décision relative à l’évaluation, les formulaires de révision de manuscrit, les commentaires détaillés fournis par les évaluateurs et le manuscrit annoté si disponible. Il est possible qu’on demande aux auteurs de réviser et de re-soumettre leur manuscrit accompagné des réponses point par point à tous les commentaires de l’évaluateur.
{"title":"Instructions aux auteurs","authors":"Droit d’auteur","doi":"10.5589/cjrs_instruct_f","DOIUrl":"https://doi.org/10.5589/cjrs_instruct_f","url":null,"abstract":"Le processus de révision par les pairs Toutes les propositions d’article sont soumises à un processus de révision par les pairs avant leur publication. Le comité de rédaction se réserve le droit de choisir ses propres évaluateurs, mais il est utile pour les auteurs de fournir une liste d’au moins quatre spécialistes dans le domaine d’expertise de l’article proposé pouvant agir comme évaluateurs et dont les intérêts de recherche sont suffisamment éloignés du domaine de l’article à évaluer. Prière de fournir leur adresse complète, incluant le numéro de téléphone et l’adresse de courrier électronique. Les auteurs recevront un ensemble de documents comprenant la décision relative à l’évaluation, les formulaires de révision de manuscrit, les commentaires détaillés fournis par les évaluateurs et le manuscrit annoté si disponible. Il est possible qu’on demande aux auteurs de réviser et de re-soumettre leur manuscrit accompagné des réponses point par point à tous les commentaires de l’évaluateur.","PeriodicalId":48843,"journal":{"name":"Canadian Journal of Remote Sensing","volume":null,"pages":null},"PeriodicalIF":2.6,"publicationDate":"2006-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80642280","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：