Pub Date : 2022-09-13DOI: 10.1080/19479832.2022.2121767
Xiangguo Lin, W. Xie, Libo Zhang, H. Sang, Jing Shen, S. Cui
ABSTRACT Semi-automatic extraction of roads is greatly needed to accelerate the acquisition and updating of road maps. However, road surfaces are frequently disturbed on very high spatial resolution (VHSR) remotely sensed satellite imagery, which bothers the road trackers using least-squares-based template matching. This paper presents a novel semi-automatic framework for road tracking from VHSR satellite imagery. First, a human operator inputs three seed points. Second, the computer automatically tracks the road by the template matching using Kullback–Leibler divergence as a similarity measure. At the same time, a human operator is retained in the tracking process to supervise the extracted results, to response to the program’s prompts. Once the failure or error happens, the human operator will correct the results and restart the automatic tracking. The above procedure is repeated until a whole road is tracked. Four satellite images with different complexities are used to perform experiments. The results show that our proposed road trackers is capable of automatically, accurately and fast extracting the long and high-level roads from the VHSR satellite images.
{"title":"Semi-automatic road extraction from high resolution satellite images by template matching using Kullback–Leibler divergence as a similarity measure","authors":"Xiangguo Lin, W. Xie, Libo Zhang, H. Sang, Jing Shen, S. Cui","doi":"10.1080/19479832.2022.2121767","DOIUrl":"https://doi.org/10.1080/19479832.2022.2121767","url":null,"abstract":"ABSTRACT Semi-automatic extraction of roads is greatly needed to accelerate the acquisition and updating of road maps. However, road surfaces are frequently disturbed on very high spatial resolution (VHSR) remotely sensed satellite imagery, which bothers the road trackers using least-squares-based template matching. This paper presents a novel semi-automatic framework for road tracking from VHSR satellite imagery. First, a human operator inputs three seed points. Second, the computer automatically tracks the road by the template matching using Kullback–Leibler divergence as a similarity measure. At the same time, a human operator is retained in the tracking process to supervise the extracted results, to response to the program’s prompts. Once the failure or error happens, the human operator will correct the results and restart the automatic tracking. The above procedure is repeated until a whole road is tracked. Four satellite images with different complexities are used to perform experiments. The results show that our proposed road trackers is capable of automatically, accurately and fast extracting the long and high-level roads from the VHSR satellite images.","PeriodicalId":46012,"journal":{"name":"International Journal of Image and Data Fusion","volume":"13 1","pages":"316 - 336"},"PeriodicalIF":2.3,"publicationDate":"2022-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"42488127","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-26DOI: 10.1080/19479832.2022.2116492
Li Yuan, Wenbo Wu, Shuli Dong, Q. He, Feiran Zhang
ABSTRACT For a camera with automatic gain control, two images with high and low optical gain can be output at the same exposure time. Due to the small gain value, most of target details are hidden in the dark pixels for the low gain image, and the brightness saturation usually appears in high gain image for the high luminance areas. To obtain the essential information from the dual gain images, a generation method of high dynamic range image based on dual gain image was developed. The method is composed of five parts, including enhancement of image detail, establishment of Laplacian pyramid, selection of fusion operator, reconstruction of fusion pyramid and adjustment of image contrast. Results showed that combination of the gradient operator for N-1 layer and the neighbourhood filter operator for the Nth layer had better fusion effect. Moreover, based on the analysis of image information entropy and clarity, the fusion efficiency was calculated, and the fusion efficiency of Mertens’s method, Jiang’s method, Zhang’s method, Goshtasby’s method and the presented method was 30.5%, 33.5%, 39.5%, 51% and 99%, indicating that the HDR fusion method based on dual gain image is reliable.
{"title":"A High Dynamic Range Image Fusion Method Based on Dual Gain Image","authors":"Li Yuan, Wenbo Wu, Shuli Dong, Q. He, Feiran Zhang","doi":"10.1080/19479832.2022.2116492","DOIUrl":"https://doi.org/10.1080/19479832.2022.2116492","url":null,"abstract":"ABSTRACT For a camera with automatic gain control, two images with high and low optical gain can be output at the same exposure time. Due to the small gain value, most of target details are hidden in the dark pixels for the low gain image, and the brightness saturation usually appears in high gain image for the high luminance areas. To obtain the essential information from the dual gain images, a generation method of high dynamic range image based on dual gain image was developed. The method is composed of five parts, including enhancement of image detail, establishment of Laplacian pyramid, selection of fusion operator, reconstruction of fusion pyramid and adjustment of image contrast. Results showed that combination of the gradient operator for N-1 layer and the neighbourhood filter operator for the Nth layer had better fusion effect. Moreover, based on the analysis of image information entropy and clarity, the fusion efficiency was calculated, and the fusion efficiency of Mertens’s method, Jiang’s method, Zhang’s method, Goshtasby’s method and the presented method was 30.5%, 33.5%, 39.5%, 51% and 99%, indicating that the HDR fusion method based on dual gain image is reliable.","PeriodicalId":46012,"journal":{"name":"International Journal of Image and Data Fusion","volume":"14 1","pages":"15 - 37"},"PeriodicalIF":2.3,"publicationDate":"2022-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"43077309","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-07-31DOI: 10.1080/19479832.2022.2106317
A. Ben Abbes, N. Jarray
ABSTRACT Here, we present a novel unsupervised self-training method (USTM) for SM estimation. First, a ML model is trained using the labeled and unlabeled data. Then, the pseudo-labeled data are generated employing the second model by adding a proxy labeled data. Eventually, SM is estimated applying the third model by pseudo-labeled data generated by the second model and unlabeled data. The final SM estimation result is obtained by training the third model. Subsequently, in-situ measurements are performed to validate our method. The final model is an unsupervised learning model. Experiments were carried out at two different sites located in southern Tunisia using Sentinel-1A and Sentinel-2A data. The input data include the backscatter coefficient in two-mode polarization ( and ), derived from Sentinel-1A, as well as the Normalized Difference Vegetation Index (NDVI) and the Normalized Difference Infrared Index (NDII) for Sentinel-2A and in-situ data. The USTM method based on (Random Forest (RF)- Convolutional neural network (CNN)-CNN) combination allowed obtaining the best performance and precision rate, compared to other combinations (Artificial Neural Network (ANN)-CNN-CNN) and (eXtreme Gradient Boosting (XGBoost)-CNN-CNN).
{"title":"Unsupervised self-training method based on deep learning for soil moisture estimation using synergy of sentinel-1 and sentinel-2 images","authors":"A. Ben Abbes, N. Jarray","doi":"10.1080/19479832.2022.2106317","DOIUrl":"https://doi.org/10.1080/19479832.2022.2106317","url":null,"abstract":"ABSTRACT Here, we present a novel unsupervised self-training method (USTM) for SM estimation. First, a ML model is trained using the labeled and unlabeled data. Then, the pseudo-labeled data are generated employing the second model by adding a proxy labeled data. Eventually, SM is estimated applying the third model by pseudo-labeled data generated by the second model and unlabeled data. The final SM estimation result is obtained by training the third model. Subsequently, in-situ measurements are performed to validate our method. The final model is an unsupervised learning model. Experiments were carried out at two different sites located in southern Tunisia using Sentinel-1A and Sentinel-2A data. The input data include the backscatter coefficient in two-mode polarization ( and ), derived from Sentinel-1A, as well as the Normalized Difference Vegetation Index (NDVI) and the Normalized Difference Infrared Index (NDII) for Sentinel-2A and in-situ data. The USTM method based on (Random Forest (RF)- Convolutional neural network (CNN)-CNN) combination allowed obtaining the best performance and precision rate, compared to other combinations (Artificial Neural Network (ANN)-CNN-CNN) and (eXtreme Gradient Boosting (XGBoost)-CNN-CNN).","PeriodicalId":46012,"journal":{"name":"International Journal of Image and Data Fusion","volume":"14 1","pages":"1 - 14"},"PeriodicalIF":2.3,"publicationDate":"2022-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"48358633","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-21DOI: 10.1080/19479832.2022.2086304
Deepika Kamboj, Sharda Vashisth, Sumeet Saurav
ABSTRACT With advancements in autonomous driving, demand for stringent and computationally efficient traffic sign detection systems has increased. However, bringing such a system to a deployable level requires handling critical accuracy and processing speed issues. A focal loss-based single-stage object detector, i.e RetinaNet, is used as a trade-off between accuracy and processing speed as it handles the class imbalance problem of the single-stage detector and is thus suitable for traffic sign detection (TSD). We assessed the detector’s performance by combining various feature extractors such as ResNet-50, ResNet-101, and ResNet-152 on three publicly available TSD benchmark datasets. Performance comparison of the detector using different backbone includes evaluation parameters like mean average precision (mAP), memory allocation, running time, and floating-point operations. From the evaluation results, we found that the RetinaNet object detector using the ResNet-152 backbone obtains the best mAP, while that using ResNet-101 strikes the best trade-off between accuracy and execution time. The motivation behind benchmarking the detector on different datasets is to analyse the detector’s performance on different TSD benchmark datasets. Among the three feature extractors, the RetinaNet model trained using the ResNet-50 backbone is an excellent model in memory consumption, making it an optimal choice for low-cost embedded devices deployment.
{"title":"Evaluation of focal loss based deep neural networks for traffic sign detection","authors":"Deepika Kamboj, Sharda Vashisth, Sumeet Saurav","doi":"10.1080/19479832.2022.2086304","DOIUrl":"https://doi.org/10.1080/19479832.2022.2086304","url":null,"abstract":"ABSTRACT With advancements in autonomous driving, demand for stringent and computationally efficient traffic sign detection systems has increased. However, bringing such a system to a deployable level requires handling critical accuracy and processing speed issues. A focal loss-based single-stage object detector, i.e RetinaNet, is used as a trade-off between accuracy and processing speed as it handles the class imbalance problem of the single-stage detector and is thus suitable for traffic sign detection (TSD). We assessed the detector’s performance by combining various feature extractors such as ResNet-50, ResNet-101, and ResNet-152 on three publicly available TSD benchmark datasets. Performance comparison of the detector using different backbone includes evaluation parameters like mean average precision (mAP), memory allocation, running time, and floating-point operations. From the evaluation results, we found that the RetinaNet object detector using the ResNet-152 backbone obtains the best mAP, while that using ResNet-101 strikes the best trade-off between accuracy and execution time. The motivation behind benchmarking the detector on different datasets is to analyse the detector’s performance on different TSD benchmark datasets. Among the three feature extractors, the RetinaNet model trained using the ResNet-50 backbone is an excellent model in memory consumption, making it an optimal choice for low-cost embedded devices deployment.","PeriodicalId":46012,"journal":{"name":"International Journal of Image and Data Fusion","volume":"14 1","pages":"122 - 144"},"PeriodicalIF":2.3,"publicationDate":"2022-06-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"44159531","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-17DOI: 10.1080/19479832.2022.2076160
H. H. Handayani, Arizal Bawasir, A. Cahyono, T. Hariyanto, H. Hidayat
ABSTRACT Digital Elevation Model (DEM) is the most vital data to generate drainage networks and to provide critical terrain factors and hydrologic derivatives, such as slope, aspect, and streamflow. The accuracy of generated drainage features is extensively dependent on the quality and resolution of DEM, such as LiDAR-derived DEM. Contrary, it has a high level of roughness and complexity. Thus, smoothing methods are sometimes employed to conquer the roughness. This paper presents feature-preserving DEM smoothing (FPDEM-S) and edge-preserving DEM smoothing (EPDEM-S) approaches to smooth surface complexity in kind of preserving small drainage features using the 0.5 m – resolution LiDAR DEM of the Kedungbener River area in Kebumen Regency, Indonesia. Entangling linear morphometric factors, those smoothing approaches delivered a slight difference of stream number, with the FPDEM-S stream length ratio performing 7% better tendencies. The FPDEM-S method perormed better than EPDEM-S in this study area to provide an optimal smoothed LiDAR DEM at certain parameter values. Summarising that two smoothing methods approaches performed similar characteristics of watershed as an oval structure close to the circular shape. Also, it can be revealed that the watershed did not reach maturity phase.
{"title":"Surface drainage features identification using LiDAR DEM smoothing in agriculture area: a study case of Kebumen Regency, Indonesia","authors":"H. H. Handayani, Arizal Bawasir, A. Cahyono, T. Hariyanto, H. Hidayat","doi":"10.1080/19479832.2022.2076160","DOIUrl":"https://doi.org/10.1080/19479832.2022.2076160","url":null,"abstract":"ABSTRACT Digital Elevation Model (DEM) is the most vital data to generate drainage networks and to provide critical terrain factors and hydrologic derivatives, such as slope, aspect, and streamflow. The accuracy of generated drainage features is extensively dependent on the quality and resolution of DEM, such as LiDAR-derived DEM. Contrary, it has a high level of roughness and complexity. Thus, smoothing methods are sometimes employed to conquer the roughness. This paper presents feature-preserving DEM smoothing (FPDEM-S) and edge-preserving DEM smoothing (EPDEM-S) approaches to smooth surface complexity in kind of preserving small drainage features using the 0.5 m – resolution LiDAR DEM of the Kedungbener River area in Kebumen Regency, Indonesia. Entangling linear morphometric factors, those smoothing approaches delivered a slight difference of stream number, with the FPDEM-S stream length ratio performing 7% better tendencies. The FPDEM-S method perormed better than EPDEM-S in this study area to provide an optimal smoothed LiDAR DEM at certain parameter values. Summarising that two smoothing methods approaches performed similar characteristics of watershed as an oval structure close to the circular shape. Also, it can be revealed that the watershed did not reach maturity phase.","PeriodicalId":46012,"journal":{"name":"International Journal of Image and Data Fusion","volume":"14 1","pages":"182 - 203"},"PeriodicalIF":2.3,"publicationDate":"2022-05-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46974091","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-01DOI: 10.1080/19479832.2022.2069160
Yousef Salem, H. Ghrefat, R. Sankaran
ABSTRACT Spatial variability of grain sizes and mapping of aeolian dunes is important to study the sand erosion, transport, and dune movement and to understand the dune encroachment and land degradation. This study examines the grain size statistical parameters and mineralogical composition of 68 sand samples collected from 17 crescentic dunes and assesses the source and depositional environment of these dunes. The analyses of samples for grain sizes resulted that the sands are characteristics to fine with an average size of 2.28 Φ and classified as moderately well-sorted (0.59 Φ), mesokurtic (0.97 Φ), and fine to coarsely skewed (0.14 Φ). X-Ray Diffraction shows that the dunes are deposited mainly by quartz, calcite, and haematite. The occurrence of absorption features near 0.5, 0.9, and 2.22 μm confirm the presence of such iron and aluminosilicate minerals in the dunes. The dunes of the provinces were mapped using TIR bands of ASTER satellite data by Carbonate index (CI) and Quartz index (QI). A good agreement among the results of grain size analyses, spectral measurements, mineralogical studies, and mapping of dunes with the field observations suggests that the sand deposits in the study area have a diversity of sources in the aeolian environment.
{"title":"Reflectance spectroscopy and ASTER mapping of aeolian dunes of Shaqra and Tharmada Provinces, Saudi Arabia: Field validation and laboratory confirmation","authors":"Yousef Salem, H. Ghrefat, R. Sankaran","doi":"10.1080/19479832.2022.2069160","DOIUrl":"https://doi.org/10.1080/19479832.2022.2069160","url":null,"abstract":"ABSTRACT Spatial variability of grain sizes and mapping of aeolian dunes is important to study the sand erosion, transport, and dune movement and to understand the dune encroachment and land degradation. This study examines the grain size statistical parameters and mineralogical composition of 68 sand samples collected from 17 crescentic dunes and assesses the source and depositional environment of these dunes. The analyses of samples for grain sizes resulted that the sands are characteristics to fine with an average size of 2.28 Φ and classified as moderately well-sorted (0.59 Φ), mesokurtic (0.97 Φ), and fine to coarsely skewed (0.14 Φ). X-Ray Diffraction shows that the dunes are deposited mainly by quartz, calcite, and haematite. The occurrence of absorption features near 0.5, 0.9, and 2.22 μm confirm the presence of such iron and aluminosilicate minerals in the dunes. The dunes of the provinces were mapped using TIR bands of ASTER satellite data by Carbonate index (CI) and Quartz index (QI). A good agreement among the results of grain size analyses, spectral measurements, mineralogical studies, and mapping of dunes with the field observations suggests that the sand deposits in the study area have a diversity of sources in the aeolian environment.","PeriodicalId":46012,"journal":{"name":"International Journal of Image and Data Fusion","volume":"14 1","pages":"162 - 181"},"PeriodicalIF":2.3,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"44619918","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
ABSTRACT Ultra-wideband is a promising technology in indoor positioning due to its accurate time resolution and good penetration. Since the positioning model is non-linear, iterative methods are often considered for solving the localisation problem. However, the positioning system is prone to become ill-posed. The iterative methods cannot easily converge to a global optimal solution. In this paper, the convergence property of four non-linear iterative methods is analytically reviewed under ill-conditioned configuration. For the iteration, three types of initial values are selected. Experimental results are given to demonstrate that although the barycentre method can converge correctly, it is inefficient with too many iterations. In addition, with a good initial value, the Gauss–Newton method can converge effectively, and it sometimes converges to a false local optimisation solution when selecting a bad initial value. Moreover, both the regularised Gauss–Newton method and closed-form Newton method can converge to the global optimum effectively with fewer iterations. This study shows that the closed-form Newton method has higher efficiency of convergence than the other methods. Meanwhile, to make complete use of measurements available to improve the accuracy, the result of non-iterative method is generally used as the initial value of the iterative method.
{"title":"Performance analysis of parameter estimator on non-linear iterative methods for ultra-wideband positioning","authors":"Chuanyang Wang, Bing He, Liangliang Shi, Weiduo Huang, Liuxu Shan","doi":"10.1080/19479832.2022.2064554","DOIUrl":"https://doi.org/10.1080/19479832.2022.2064554","url":null,"abstract":"ABSTRACT Ultra-wideband is a promising technology in indoor positioning due to its accurate time resolution and good penetration. Since the positioning model is non-linear, iterative methods are often considered for solving the localisation problem. However, the positioning system is prone to become ill-posed. The iterative methods cannot easily converge to a global optimal solution. In this paper, the convergence property of four non-linear iterative methods is analytically reviewed under ill-conditioned configuration. For the iteration, three types of initial values are selected. Experimental results are given to demonstrate that although the barycentre method can converge correctly, it is inefficient with too many iterations. In addition, with a good initial value, the Gauss–Newton method can converge effectively, and it sometimes converges to a false local optimisation solution when selecting a bad initial value. Moreover, both the regularised Gauss–Newton method and closed-form Newton method can converge to the global optimum effectively with fewer iterations. This study shows that the closed-form Newton method has higher efficiency of convergence than the other methods. Meanwhile, to make complete use of measurements available to improve the accuracy, the result of non-iterative method is generally used as the initial value of the iterative method.","PeriodicalId":46012,"journal":{"name":"International Journal of Image and Data Fusion","volume":"14 1","pages":"145 - 161"},"PeriodicalIF":2.3,"publicationDate":"2022-04-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"43498204","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-25DOI: 10.1080/19479832.2022.2055157
Gebeyehu Abebe, T. Tadesse, B. Gessesse
ABSTRACT Accurate estimation of crop parameters, such as Leaf Area Index (LAI) and biomass over large areas using remote sensing techniques, is crucial for monitoring crop growth and yield prediction. In this study, a Gaussian Process Regression (GPR) method was developed to estimate LAI and biomass values of sugarcane during growth season using optical and synthetic-Aperture Radar (SAR) data fusion. Predicting LAI on an independent test data set using the GPR and the combined optical and SAR indices provided better prediction accuracies of LAI; with the GPR based on radial basis function (Root Mean Square Error [RMSE] = 0.34, Mean Absolute Error [MAE] = 0.28 and Mean Absolute Percentage Error [MAPE] = 10.5%) and polynomial function (RMSE = 0.42, MAE = 0.31 and MAPE = 12.58%), respectively. The test results of sugarcane biomass also showed that the GPR (poly) produced the highest statistical results (RMSE = 2.45 kg/m2, MAE = 1.72 kg/m2, MAPE = 8.1%) using the combined indices. The results suggest that the crop biophysical retrieval based on optical and SAR data fusion and GPR proposed in this study could improve LAI and biomass estimation that could help for effective crop growth monitoring and mapping applications.
{"title":"Estimating Leaf Area Index and biomass of sugarcane based on Gaussian process regression using Landsat 8 and Sentinel 1A observations","authors":"Gebeyehu Abebe, T. Tadesse, B. Gessesse","doi":"10.1080/19479832.2022.2055157","DOIUrl":"https://doi.org/10.1080/19479832.2022.2055157","url":null,"abstract":"ABSTRACT Accurate estimation of crop parameters, such as Leaf Area Index (LAI) and biomass over large areas using remote sensing techniques, is crucial for monitoring crop growth and yield prediction. In this study, a Gaussian Process Regression (GPR) method was developed to estimate LAI and biomass values of sugarcane during growth season using optical and synthetic-Aperture Radar (SAR) data fusion. Predicting LAI on an independent test data set using the GPR and the combined optical and SAR indices provided better prediction accuracies of LAI; with the GPR based on radial basis function (Root Mean Square Error [RMSE] = 0.34, Mean Absolute Error [MAE] = 0.28 and Mean Absolute Percentage Error [MAPE] = 10.5%) and polynomial function (RMSE = 0.42, MAE = 0.31 and MAPE = 12.58%), respectively. The test results of sugarcane biomass also showed that the GPR (poly) produced the highest statistical results (RMSE = 2.45 kg/m2, MAE = 1.72 kg/m2, MAPE = 8.1%) using the combined indices. The results suggest that the crop biophysical retrieval based on optical and SAR data fusion and GPR proposed in this study could improve LAI and biomass estimation that could help for effective crop growth monitoring and mapping applications.","PeriodicalId":46012,"journal":{"name":"International Journal of Image and Data Fusion","volume":"14 1","pages":"58 - 88"},"PeriodicalIF":2.3,"publicationDate":"2022-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"42895966","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-06DOI: 10.1080/19479832.2022.2047801
Xiangguo Lin, W. Xie
ABSTRACT In most Mobile Laser Scanning (MLS) applications, filtering is a necessary step. In this paper, a segmentation-based filtering method is proposed for MLS point cloud, where a segment rather than an individual point is the basic processing unit. In particular, the MLS point clouds in some blocks are clustered into segments by a surface growing algorithm, and then the object segments are detected and removed. A segment-based filtering method is employed to detect the ground segments. The experiment in this paper uses two MLS point cloud datasets to evaluate the proposed method. Experiments indicate that, compared with the classic progressive TIN (Triangulated Irregular Network) densification algorithm, the proposed method is capable of reducing the omission error, the commission error and total error by 3.62%, 7.87% and 5.54% on average, respectively.
{"title":"A segment-based filtering method for mobile laser scanning point cloud","authors":"Xiangguo Lin, W. Xie","doi":"10.1080/19479832.2022.2047801","DOIUrl":"https://doi.org/10.1080/19479832.2022.2047801","url":null,"abstract":"ABSTRACT In most Mobile Laser Scanning (MLS) applications, filtering is a necessary step. In this paper, a segmentation-based filtering method is proposed for MLS point cloud, where a segment rather than an individual point is the basic processing unit. In particular, the MLS point clouds in some blocks are clustered into segments by a surface growing algorithm, and then the object segments are detected and removed. A segment-based filtering method is employed to detect the ground segments. The experiment in this paper uses two MLS point cloud datasets to evaluate the proposed method. Experiments indicate that, compared with the classic progressive TIN (Triangulated Irregular Network) densification algorithm, the proposed method is capable of reducing the omission error, the commission error and total error by 3.62%, 7.87% and 5.54% on average, respectively.","PeriodicalId":46012,"journal":{"name":"International Journal of Image and Data Fusion","volume":"13 1","pages":"136 - 154"},"PeriodicalIF":2.3,"publicationDate":"2022-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"43575013","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-02-22DOI: 10.1080/19479832.2022.2033855
Ya-Guang Tian, Yuan-Wei Chen, Wan Diming, Yuan Shaoguang, Mao Wandeng, Wang Chao, Chun-xiao Xu, Yifan Long
ABSTRACT Checking the status of the power grid is very important. However, the low occurrence of defects in an actual power grid makes it difficult to collect training samples, which affects the training of defect-detection models. In this study, we proposed a method for enhancing the defective image of a power grid based on cycle-consistent adversarial networks (CycleGAN). The defective image sample dataset was expanded by fusing artificial defective samples, converted from defect-free components of samples with the trained CycleGAN model and updating its corresponding label file. Comparing the accuracy of the object detection model trained by the augmented dataset, we found a 2%–3% Average Precision (AP) improvement over baseline, and the fusing method of histogram specification reaches the best performance. In conclusion, the generative adversarial network (GAN) and its variants have considerable potential for dataset augmentation as well as scope for further improvement.
{"title":"Augmentation Method for anti-vibration hammer on power transimission line based on CycleGAN","authors":"Ya-Guang Tian, Yuan-Wei Chen, Wan Diming, Yuan Shaoguang, Mao Wandeng, Wang Chao, Chun-xiao Xu, Yifan Long","doi":"10.1080/19479832.2022.2033855","DOIUrl":"https://doi.org/10.1080/19479832.2022.2033855","url":null,"abstract":"ABSTRACT Checking the status of the power grid is very important. However, the low occurrence of defects in an actual power grid makes it difficult to collect training samples, which affects the training of defect-detection models. In this study, we proposed a method for enhancing the defective image of a power grid based on cycle-consistent adversarial networks (CycleGAN). The defective image sample dataset was expanded by fusing artificial defective samples, converted from defect-free components of samples with the trained CycleGAN model and updating its corresponding label file. Comparing the accuracy of the object detection model trained by the augmented dataset, we found a 2%–3% Average Precision (AP) improvement over baseline, and the fusing method of histogram specification reaches the best performance. In conclusion, the generative adversarial network (GAN) and its variants have considerable potential for dataset augmentation as well as scope for further improvement.","PeriodicalId":46012,"journal":{"name":"International Journal of Image and Data Fusion","volume":"13 1","pages":"362 - 381"},"PeriodicalIF":2.3,"publicationDate":"2022-02-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49581576","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","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：