Sign in to use this feature.

Years

Between: -

Subjects

remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline

Journals

Article Types

Countries / Regions

Search Results (43)

Search Parameters:
Keywords = polarization hyperspectral imaging

Order results
Result details
Results per page
Select all
Export citation of selected articles as:
16 pages, 4169 KiB  
Article
Asymmetric Distance in K-Means Clustering Enhances Quality of Cells Raman Imaging
by Bernadette Scopacasa and Patrizio Candeloro
Appl. Sci. 2025, 15(8), 4461; https://doi.org/10.3390/app15084461 - 17 Apr 2025
Viewed by 639
Abstract
Raman microspectroscopy is a powerful, label-free technique for the biochemical characterization of cells, but its complex spectral data require advanced computational methods for meaningful interpretation. Clustering analysis is widely used in spectroscopic imaging to extract meaningful biochemical information. Traditional methods, such as K-means [...] Read more.
Raman microspectroscopy is a powerful, label-free technique for the biochemical characterization of cells, but its complex spectral data require advanced computational methods for meaningful interpretation. Clustering analysis is widely used in spectroscopic imaging to extract meaningful biochemical information. Traditional methods, such as K-means clustering with Euclidean distance, often struggle to capture subtle spectral variations, leading to suboptimal segmentation. Alternative distance metrics, including cosine and Mahalanobis distances, have been explored to enhance cluster separability, yet challenges remain in distinguishing chemically relevant features while minimizing redundancy and noise. In this study, we introduce an asymmetric metric distance matrix with a tunable eccentricity parameter to improve clustering performance in Raman hyperspectral imaging. Our results demonstrate that suitable eccentricity values enhance the identification of subcellular structures while requiring fewer clusters than Euclidean-based approaches. Compared to polar metrics, the proposed asymmetric metric achieves better stability and reduced noise, leading to more accurate segmentation. Future research could explore its application in other clustering techniques and machine learning frameworks, as well as its application in broader spectral imaging techniques where the distance metric plays a fundamental role. Full article
(This article belongs to the Special Issue Biological Sample Analysis Techniques and Devices)
Show Figures

Graphical abstract

40 pages, 14878 KiB  
Article
Selection of Landing Sites for the Chang’E-7 Mission Using Multi-Source Remote Sensing Data
by Fei Zhao, Pingping Lu, Tingyu Meng, Yanan Dang, Yao Gao, Zihan Xu, Robert Wang and Yirong Wu
Remote Sens. 2025, 17(7), 1121; https://doi.org/10.3390/rs17071121 - 21 Mar 2025
Cited by 1 | Viewed by 1460
Abstract
The Chinese Chang’E-7 (CE-7) mission is planned to land in the lunar south polar region, and then deploy a mini-flying probe to fly into the cold trap to detect the water ice. The selection of a landing site is crucial for ensuring both [...] Read more.
The Chinese Chang’E-7 (CE-7) mission is planned to land in the lunar south polar region, and then deploy a mini-flying probe to fly into the cold trap to detect the water ice. The selection of a landing site is crucial for ensuring both a safe landing and the successful achievement of its scientific objectives. This study presents a method for landing site selection in the challenging environment of the lunar south pole, utilizing multi-source remote sensing data. First, the likelihood of water ice in all cold traps within 85°S is assessed and prioritized using neutron spectrometer and hyperspectral data, with the most promising cold traps selected for sampling by CE-7’s mini-flying probe. Slope and illumination data are then used to screen feasible landing sites in the south polar region. Feasible landing sites near cold traps are aggregated into larger landing regions. Finally, high-resolution illumination maps, along with optical and radar images, are employed to refine the selection and identify the optimal landing sites. Six potential landing sites around the de Gerlache crater, an unnamed cold trap at (167.10°E, 88.71°S), Faustini crater, and Shackleton crater are proposed. It would be beneficial for CE-7 to prioritize mapping these sites post-launch using its high-resolution optical camera and radar for further detailed landing site investigation and evaluation. Full article
(This article belongs to the Special Issue Remote Sensing and Photogrammetry Applied to Deep Space Exploration)
Show Figures

Figure 1

27 pages, 16278 KiB  
Article
Development of a Full-Field Integrated Radiation Observation System for Lunar Hyperspectral Irradiance Measurement
by Ye Jiang, Xin Ye, Yuwei Wang, Yuchen Lin, Dongjun Yang and Wei Fang
Remote Sens. 2025, 17(4), 626; https://doi.org/10.3390/rs17040626 - 12 Feb 2025
Viewed by 935
Abstract
The Moon serves as an ideal reference radiation source for on-orbit calibration of starborne optical remote sensing instruments. To enhance the characterization capability for lunar spectral radiation, the full-field integrated radiation observation system (FIROS) for lunar hyperspectral irradiance measurement has been developed. FIROS [...] Read more.
The Moon serves as an ideal reference radiation source for on-orbit calibration of starborne optical remote sensing instruments. To enhance the characterization capability for lunar spectral radiation, the full-field integrated radiation observation system (FIROS) for lunar hyperspectral irradiance measurement has been developed. FIROS accomplished lunar hyperspectral irradiance measurements in the 400–1000 nm range by integrating and spectrally analyzing the radiation across the entire lunar disc, reducing the angular sensitivity and polarization sensitivity to lunar radiation. Performance tests and preliminary lunar observational experiments conducted on FIROS indicate that the system possesses excellent response linearity and environmental adaptability, with a reduction in lunar tracking accuracy requirements by approximately an order of magnitude compared to push-broom imaging observations. The performance and lunar observation capabilities of the system have been well validated. FIROS provides a lunar observation method that simultaneously achieves full-disk light collection and hyperspectral measurement, demonstrating strong environmental adaptability and laying a solid foundation for enhancing long-term stable lunar observation data and establishing lunar radiation benchmarks. Full article
Show Figures

Figure 1

22 pages, 24211 KiB  
Article
Design of Polarization Spectroscopy Integrated Imaging System
by Jianan Liu, Jing Cui, Mingce Chen, Shuo Yang, Hongyu Sun, Qi Wang, Juntong Zhan, Yingchao Li, Qiang Fu and Chao Wang
Photonics 2024, 11(12), 1183; https://doi.org/10.3390/photonics11121183 - 17 Dec 2024
Cited by 1 | Viewed by 1026
Abstract
To simultaneously acquire the spectral and polarization information of the target and achieve the monitoring and identification of the target object, a polarization spectral integrated imaging system is proposed in this paper. Firstly, the structural principle of the polarization spectral integrated imaging system [...] Read more.
To simultaneously acquire the spectral and polarization information of the target and achieve the monitoring and identification of the target object, a polarization spectral integrated imaging system is proposed in this paper. Firstly, the structural principle of the polarization spectral integrated imaging system is introduced. The relationship between the spatial resolution, spectral resolution, and the system’s structural parameters is analyzed. The design of the optical part of the polarization spectral integrated imaging system is completed, along with the tolerance analysis. Secondly, the mechanical structure of the polarization spectral integrated imaging system is designed. Finally, by using a drone to carry the polarization spectral integrated imaging system, a simulation experiment for sea surface oil spill monitoring is conducted, and the hyperspectral and polarization information of the ocean, crude oil, fuel oil, palm oil, diesel, and gasoline are obtained. The polarization and spectral information were integrated. The integration of hyperspectral and polarization data yields remarkable enhancement outcomes, allowing for the clear delineation of previously challenging-to-identify crude oil contamination areas against the marine background in the fused images, characterized by sharper boundaries and improved discriminability. This accomplishment underscores the feasibility of our system for the rapid identification of large-scale oil spill events. Full article
Show Figures

Figure 1

14 pages, 7636 KiB  
Article
Test Method for Mineral Spatial Distribution of BIF Ore by Imaging Spectrometer
by Wenhua Yi, Shanjun Liu, Ruibo Ding, Heng Yue, Haoran Wang and Jingli Wang
Minerals 2024, 14(9), 959; https://doi.org/10.3390/min14090959 - 23 Sep 2024
Cited by 1 | Viewed by 1275
Abstract
The spatial distribution characteristics of iron ore components are important when measuring the difficulty of their beneficiation. Polarized light microscopy and scanning electron microscopy are traditional methods with some shortcomings, including complicated operation and low efficiency. Most of the laboratory hyperspectral imaging techniques [...] Read more.
The spatial distribution characteristics of iron ore components are important when measuring the difficulty of their beneficiation. Polarized light microscopy and scanning electron microscopy are traditional methods with some shortcomings, including complicated operation and low efficiency. Most of the laboratory hyperspectral imaging techniques that have emerged in recent years have been focused on the field of mineral resource exploration. In contrast, the mineral distribution and tectonic characteristics of iron ores have been relatively poorly studied in the field of beneficiation. To address the issue, 11 experimental samples of banded iron formation (BIF)-hosted iron ores were selected and tested using an imaging spectrometer. Then, based on the differences in spectral characteristic of the three main components (quartz, hematite, and magnetite) in the samples, the identification model of the spatial distribution of the iron ore components was established using the normalized spectral amplitude index (NSAI) and spectral angle mapper (SAM). The NSAI and SAM identify minerals based on spectral amplitude features and spectral morphological features of the sample, respectively. The spatial distribution of different minerals in the samples was tested using the model, and the test results demonstrated that the spatial distribution of the three components is consistent with the banded tectonic character of the sample. Upon comparison with the chemical test results, the mean absolute errors (MAE) of the model for quartz, hematite, and magnetite in the samples were 2.03%, 1.34%, and 1.55%, respectively, and the root mean square errors (RMSE) were 2.72%, 2.08%, and 1.85%, respectively, with the exception of one martite sample that reached an MAE of 10.17%. Therefore, the model demonstrates a high degree of accuracy. The research provides a new method to test the spatial distribution of iron ore components. Full article
Show Figures

Figure 1

23 pages, 9214 KiB  
Article
DCFF-Net: Deep Context Feature Fusion Network for High-Precision Classification of Hyperspectral Image
by Zhijie Chen, Yu Chen, Yuan Wang, Xiaoyan Wang, Xinsheng Wang and Zhouru Xiang
Remote Sens. 2024, 16(16), 3002; https://doi.org/10.3390/rs16163002 - 15 Aug 2024
Cited by 3 | Viewed by 1736
Abstract
Hyperspectral images (HSI) contain abundant spectral information. Efficient extraction and utilization of this information for image classification remain prominent research topics. Previously, hyperspectral classification techniques primarily relied on statistical attributes and mathematical models of spectral data. Deep learning classification techniques have recently been [...] Read more.
Hyperspectral images (HSI) contain abundant spectral information. Efficient extraction and utilization of this information for image classification remain prominent research topics. Previously, hyperspectral classification techniques primarily relied on statistical attributes and mathematical models of spectral data. Deep learning classification techniques have recently been extensively utilized for hyperspectral data classification, yielding promising outcomes. This study proposes a deep learning approach that uses polarization feature maps for classification. Initially, the polar co-ordinate transformation method was employed to convert the spectral information of all pixels in the image into spectral feature maps. Subsequently, the proposed Deep Context Feature Fusion Network (DCFF-NET) was utilized to classify these feature maps. The model was validated using three open-source hyperspectral datasets: Indian Pines, Pavia University, and Salinas. The experimental results indicated that DCFF-NET achieved excellent classification performance. Experimental results on three public HSI datasets demonstrated that the proposed method accurately recognized different objects with an overall accuracy (OA) of 86.68%, 94.73%, and 95.14% based on the pixel method, and 98.15%, 99.86%, and 99.98% based on the pixel-patch method. Full article
(This article belongs to the Special Issue GeoAI and EO Big Data Driven Advances in Earth Environmental Science)
Show Figures

Figure 1

27 pages, 2083 KiB  
Article
A Wide-Angle Hyperspectral Top-of-Atmosphere Reflectance Model for the Libyan Desert
by Fuxiang Guo, Xiaobing Zheng, Yanna Zhang, Wei Wei, Zejie Zhang, Quan Zhang and Xin Li
Remote Sens. 2024, 16(8), 1406; https://doi.org/10.3390/rs16081406 - 16 Apr 2024
Cited by 3 | Viewed by 1343
Abstract
Reference targets with stability, uniformity, and known reflectance on the Earth’s surface, such as deserts, can be used for the absolute radiometric calibration of satellite sensors. A wide-angle hyperspectral reflectance model at the top of atmosphere (TOA) over such a reference target will [...] Read more.
Reference targets with stability, uniformity, and known reflectance on the Earth’s surface, such as deserts, can be used for the absolute radiometric calibration of satellite sensors. A wide-angle hyperspectral reflectance model at the top of atmosphere (TOA) over such a reference target will expand the applicability of on-orbit calibration to different spectral bands and angles. To achieve the long-term, continuous, and high-precision absolute radiometric calibration of remote sensors, a wide-angle hyperspectral TOA reflectance model of the Libyan Desert was constructed based on spectral reflectance data, satellite overpass parameters, and atmospheric parameters from the Terra/Aqua and Earth Observation-1 (EO-1) satellites between 2003 and 2012. By means of angle fitting, viewing angle grouping, and spectral extension, the model is applicable for absolute radiometric calibration of the visible to short-wave infrared (SWIR) bands for sensors within viewing zenith angles of 65 degrees. To validate the accuracy and precision of the model, a total of 3120 long-term validations of model accuracy and 949 cross-validations with the Landsat 8 Operational Land Imager (OLI) and Suomi National Polar-Orbiting Partnership (NPP) Visible Infrared Imaging Radiometer Suite (VIIRS) satellite sensors between 2013 and 2020 were conducted. The results show that the TOA reflectance calculated by the model had a standard deviation (SD) of relative differences below 1.9% and a root-mean-square error (RMSE) below 0.8% when compared with observations from the Moderate Resolution Imaging Spectroradiometer (MODIS) and Landsat 8 OLI. The SD of the relative differences and the RMSE were within 2.7% when predicting VIIRS data. Full article
(This article belongs to the Special Issue Remote Sensing Satellites Calibration and Validation)
Show Figures

Graphical abstract

20 pages, 6947 KiB  
Article
Fusion of Multimodal Imaging and 3D Digitization Using Photogrammetry
by Roland Ramm, Pedro de Dios Cruz, Stefan Heist, Peter Kühmstedt and Gunther Notni
Sensors 2024, 24(7), 2290; https://doi.org/10.3390/s24072290 - 3 Apr 2024
Cited by 3 | Viewed by 2401
Abstract
Multimodal sensors capture and integrate diverse characteristics of a scene to maximize information gain. In optics, this may involve capturing intensity in specific spectra or polarization states to determine factors such as material properties or an individual’s health conditions. Combining multimodal camera data [...] Read more.
Multimodal sensors capture and integrate diverse characteristics of a scene to maximize information gain. In optics, this may involve capturing intensity in specific spectra or polarization states to determine factors such as material properties or an individual’s health conditions. Combining multimodal camera data with shape data from 3D sensors is a challenging issue. Multimodal cameras, e.g., hyperspectral cameras, or cameras outside the visible light spectrum, e.g., thermal cameras, lack strongly in terms of resolution and image quality compared with state-of-the-art photo cameras. In this article, a new method is demonstrated to superimpose multimodal image data onto a 3D model created by multi-view photogrammetry. While a high-resolution photo camera captures a set of images from varying view angles to reconstruct a detailed 3D model of the scene, low-resolution multimodal camera(s) simultaneously record the scene. All cameras are pre-calibrated and rigidly mounted on a rig, i.e., their imaging properties and relative positions are known. The method was realized in a laboratory setup consisting of a professional photo camera, a thermal camera, and a 12-channel multispectral camera. In our experiments, an accuracy better than one pixel was achieved for the data fusion using multimodal superimposition. Finally, application examples of multimodal 3D digitization are demonstrated, and further steps to system realization are discussed. Full article
(This article belongs to the Special Issue Multi-Modal Image Processing Methods, Systems, and Applications)
Show Figures

Figure 1

18 pages, 9610 KiB  
Article
Dual-Channel Switchable Metasurface Filters for Compact Spectral Imaging with Deep Compressive Reconstruction
by Chang Wang, Xinyu Liu, Yang Zhang, Yan Sun, Zeqing Yu and Zhenrong Zheng
Nanomaterials 2023, 13(21), 2854; https://doi.org/10.3390/nano13212854 - 27 Oct 2023
Cited by 2 | Viewed by 2698
Abstract
Spectral imaging technology, which aims to capture images across multiple spectral channels and create a spectral data cube, has been widely utilized in various fields. However, conventional spectral imaging systems face challenges, such as slow acquisition speed and large size. The rapid development [...] Read more.
Spectral imaging technology, which aims to capture images across multiple spectral channels and create a spectral data cube, has been widely utilized in various fields. However, conventional spectral imaging systems face challenges, such as slow acquisition speed and large size. The rapid development of optical metasurfaces, capable of manipulating light fields versatilely and miniaturizing optical components into ultrathin planar devices, offers a promising solution for compact hyperspectral imaging (HSI). This study proposes a compact snapshot compressive spectral imaging (SCSI) system by leveraging the spectral modulations of metasurfaces with dual-channel switchable metasurface filters and employing a deep-learning-based reconstruction algorithm. To achieve compactness, the proposed system integrates dual-channel switchable metasurface filters using twisted nematic liquid crystals (TNLCs) and anisotropic titanium dioxide (TiO2) nanostructures. These thin metasurface filters are closely attached to the image sensor, resulting in a compact system. The TNLCs possess a broadband linear polarization conversion ability, enabling the rapid switching of the incidence polarization state between x-polarization and y-polarization by applying different voltages. This polarization conversion facilitates the generation of two groups of transmittance spectra for wavelength-encoding, providing richer information for spectral data cube reconstruction compared to that of other snapshot compressive spectral imaging techniques. In addition, instead of employing classic iterative compressive sensing (CS) algorithms, an end-to-end residual neural network (ResNet) is utilized to reconstruct the spectral data cube. This neural network leverages the 2-frame snapshot measurements of orthogonal polarization channels. The proposed hyperspectral imaging technology demonstrates superior reconstruction quality and speed compared to those of the traditional compressive hyperspectral image recovery methods. As a result, it is expected that this technology will have substantial implications in various domains, including but not limited to object detection, face recognition, food safety, biomedical imaging, agriculture surveillance, and so on. Full article
(This article belongs to the Special Issue Photofunctional Nanomaterials and Nanostructures)
Show Figures

Figure 1

18 pages, 3320 KiB  
Article
Vigor Detection for Naturally Aged Soybean Seeds Based on Polarized Hyperspectral Imaging Combined with Ensemble Learning Algorithm
by Qingying Hu, Wei Lu, Yuxin Guo, Wei He, Hui Luo and Yiming Deng
Agriculture 2023, 13(8), 1499; https://doi.org/10.3390/agriculture13081499 - 27 Jul 2023
Cited by 6 | Viewed by 2072
Abstract
To satisfy the increasing demand for soybeans, identifying and sorting high-vigor seeds before sowing is an effective way to improve the yield. Polarized hyperspectral imaging (PHI) technology is here proposed as a rapid, non-destructive method for detecting the vigor of naturally aged soybean [...] Read more.
To satisfy the increasing demand for soybeans, identifying and sorting high-vigor seeds before sowing is an effective way to improve the yield. Polarized hyperspectral imaging (PHI) technology is here proposed as a rapid, non-destructive method for detecting the vigor of naturally aged soybean seeds. First, the spectrum of 396.1–1044.1 nm was collected to automatically extract the region of interest (ROI). Then, first derivative (FD), Savitzky–Golay (SG), multiplicative scatter correction (MSC), and standard normal variate (SNV) preprocessed hyperspectral and polarized hyperspectral data (0°, 45°, 90°, and 135°) for the soybean seeds was obtained. Finally, the seed vigor prediction model based on polarized hyperspectral components such as I, Q, and U was constructed, and partial least squares regression (PLSR), back-propagation neural network (BPNN), generalized regression neural network (GRNN), support vector regression (SVR), random forest (RF), and blending ensemble learning were applied for modeling analysis. The results showed that the prediction accuracy when using PHI was improved to 93.36%, higher than that for the hyperspectral technique, with a prediction accuracy up to 97.17%, 98.25%, and 97.55% when using the polarization component of I, Q, and U, respectively. Full article
Show Figures

Figure 1

16 pages, 7880 KiB  
Article
High Light Efficiency Spectral Polarization Imaging Method Based on Mach–Zehnder Structured Liquid Crystal Tunable Filters and Variable Retarders
by Lixin Chen, Shiyuan Zhang, Wenbin Zheng and Lishuang Yao
Photonics 2023, 10(7), 765; https://doi.org/10.3390/photonics10070765 - 3 Jul 2023
Cited by 1 | Viewed by 2290
Abstract
Liquid crystal tunable filters (LCTFs) are extensively used in hyperspectral imaging systems to obtain spectral information of target scenes. However, a typical LCTF can only filter linearly polarized light, greatly reducing the transmittance of the system and limiting its application in spectral and [...] Read more.
Liquid crystal tunable filters (LCTFs) are extensively used in hyperspectral imaging systems to obtain spectral information of target scenes. However, a typical LCTF can only filter linearly polarized light, greatly reducing the transmittance of the system and limiting its application in spectral and polarization imaging. In this paper, a spectropolarimeter using Mach–Zehnder structured LCTFs (MZ-LCTFs) combined with liquid crystal variable retarders (LCVRs) is proposed. The polarized beam splitter (PBS) can make full use of the two polarization components of the incident light to improve the transmittance of the system. Specifically, the results show that the mean pixel intensity (MPI) of spectral images is improved by 93.48% compared to a typical LCTF. Subsequently, the average signal to noise ratio (SNR) of filtered and unfiltered images when simultaneously using polarization S and P channels is increased by 2.59 dB compared to a single channel. In addition, the average Standard Deviations (STDs) of DoLP and DoCP are 0.016 and 0.018, respectively. The proposed method has the potential to be applied to obtain polarization information with high optical efficiency and a full spectrum in a wide band. Full article
(This article belongs to the Special Issue Optical Design in Night Vision Imaging)
Show Figures

Figure 1

20 pages, 6683 KiB  
Article
Cropland Mapping in Tropical Smallholder Systems with Seasonally Stratified Sentinel-1 and Sentinel-2 Spectral and Textural Features
by Manushi B. Trivedi, Michael Marshall, Lyndon Estes, C.A.J.M. de Bie, Ling Chang and Andrew Nelson
Remote Sens. 2023, 15(12), 3014; https://doi.org/10.3390/rs15123014 - 9 Jun 2023
Cited by 7 | Viewed by 2614
Abstract
Mapping arable field areas is crucial for assessing agricultural productivity but poses challenges in sub-Saharan agroecosystems because of diverse crop calendars, small and irregularly shaped fields, persistent cloud cover, and lack of high-quality model training data. This study proposes several methodological improvements to [...] Read more.
Mapping arable field areas is crucial for assessing agricultural productivity but poses challenges in sub-Saharan agroecosystems because of diverse crop calendars, small and irregularly shaped fields, persistent cloud cover, and lack of high-quality model training data. This study proposes several methodological improvements to overcome these challenges. Specifically, it utilizes long-term MODIS data to stratify finer Sentinel-2 reflectance and Sentinel-1 backscatter image features on a per-pixel basis. It also incorporates texture features and employs a machine learning approach with over 300,000 samples. The eastern region of Ghana was stratified into seven seasonal strata exhibiting distinct vegetation seasonality, capturing diversity in crop calendars, using long-term MODIS (2001–2009) normalized difference vegetation index phenology. Three years (2017–2019) of Sentinel-1 and Sentinel-2 original bands at 20 m were composited into dry and wet seasonal features according to the strata, from which spectral, polarimetric, and texture features were extracted. The field boundaries were digitized using PlanetScope images (2018–2019). Random Forest classifier with 10-fold cross-validation and recursive feature elimination was used for feature selection and model building. Including topographic variables, out of 137 image features, only 11 features were found important. Sentinel-2 SWIR-based spectral features were most important, followed by Sentinel-1 polarimetric (VV) and elevation features. Half of the 11 features were variance texture features, followed by spectral features. The Random Forest classifier produced a 0.78 AUC score with overall precision, recall, and F1-score of 0.96, 0.78, and 0.85, respectively. While the precision for both classes was >0.90, the recall rate for arable areas was half that of non-arable areas. Future studies could improve the technical workflow with reliable balanced sampling, narrowband hyperspectral images, and fully polarized SAR images. Full article
Show Figures

Graphical abstract

28 pages, 7900 KiB  
Article
Two-Branch Convolutional Neural Network with Polarized Full Attention for Hyperspectral Image Classification
by Haimiao Ge, Liguo Wang, Moqi Liu, Yuexia Zhu, Xiaoyu Zhao, Haizhu Pan and Yanzhong Liu
Remote Sens. 2023, 15(3), 848; https://doi.org/10.3390/rs15030848 - 2 Feb 2023
Cited by 25 | Viewed by 3704
Abstract
In recent years, convolutional neural networks (CNNs) have been introduced for pixel-wise hyperspectral image (HSI) classification tasks. However, some problems of the CNNs are still insufficiently addressed, such as the receptive field problem, small sample problem, and feature fusion problem. To tackle the [...] Read more.
In recent years, convolutional neural networks (CNNs) have been introduced for pixel-wise hyperspectral image (HSI) classification tasks. However, some problems of the CNNs are still insufficiently addressed, such as the receptive field problem, small sample problem, and feature fusion problem. To tackle the above problems, we proposed a two-branch convolutional neural network with a polarized full attention mechanism for HSI classification. In the proposed network, two-branch CNNs are implemented to efficiently extract the spectral and spatial features, respectively. The kernel sizes of the convolutional layers are simplified to reduce the complexity of the network. This approach can make the network easier to be trained and fit the network to small sample size conditions. The one-shot connection technique is applied to improve the efficiency of feature extraction. An improved full attention block, named polarized full attention, is exploited to fuse the feature maps and provide global contextual information. Experimental results on several public HSI datasets confirm the effectiveness of the proposed network. Full article
Show Figures

Figure 1

25 pages, 13257 KiB  
Article
Combination of Hyperspectral and Quad-Polarization SAR Images to Classify Marsh Vegetation Using Stacking Ensemble Learning Algorithm
by Hang Yao, Bolin Fu, Ya Zhang, Sunzhe Li, Shuyu Xie, Jiaoling Qin, Donglin Fan and Ertao Gao
Remote Sens. 2022, 14(21), 5478; https://doi.org/10.3390/rs14215478 - 31 Oct 2022
Cited by 17 | Viewed by 3180
Abstract
Combinations of multi-sensor remote sensing images and machine learning have attracted much attention in recent years due to the spectral similarity of wetland plant canopy. However, the integration of hyperspectral and quad-polarization synthetic aperture radar (SAR) images for classifying marsh vegetation has still [...] Read more.
Combinations of multi-sensor remote sensing images and machine learning have attracted much attention in recent years due to the spectral similarity of wetland plant canopy. However, the integration of hyperspectral and quad-polarization synthetic aperture radar (SAR) images for classifying marsh vegetation has still been faced with the challenges of using machine learning algorithms. To resolve this issue, this study proposed an approach to classifying marsh vegetation in the Honghe National Nature Reserve, northeast China, by combining backscattering coefficient and polarimetric decomposition parameters of C-band and L-band quad-polarization SAR data with hyperspectral images. We further developed an ensemble learning model by stacking Random Forest (RF), CatBoost and XGBoost algorithms for marsh vegetation mapping and evaluated its classification performance of marsh vegetation between combinations of hyperspectral and full-polarization SAR data and any of the lone sensor images. Finally, this paper explored the effect of different polarimetric decomposition methods and wavelengths of radar on classifying wetland vegetation. We found that a combination of ZH-1 hyperspectral images, C-band GF-3, and L-band ALOS-2 quad-polarization SAR images achieved the highest overall classification accuracy (93.13%), which was 5.58–9.01% higher than that only using C-band or L-band quad-polarization SAR images. This study confirmed that stacking ensemble learning provided better performance than a single machine learning model using multi-source images in most of the classification schemes, with the overall accuracy ranging from 77.02% to 92.27%. The CatBoost algorithm was capable of identifying forests and deep-water marsh vegetation. We further found that L-band ALOS-2 SAR images achieved higher classification accuracy when compared to C-band GF-3 polarimetric SAR data. ALOS-2 was more sensitive to deep-water marsh vegetation classification, while GF-3 was more sensitive to shallow-water marsh vegetation mapping. Finally, scattering model-based decomposition provided important polarimetric parameters from ALOS-2 SAR images for marsh vegetation classification, while eigenvector/eigenvalue-based and two-component decompositions produced a great contribution when using GF-3 SAR images. Full article
(This article belongs to the Special Issue Remote Sensing of Wetlands and Biodiversity)
Show Figures

Figure 1

13 pages, 3451 KiB  
Communication
Assessing FY-3E HIRAS-II Radiance Accuracy Using AHI and MERSI-LL
by Hongtao Chen and Li Guan
Remote Sens. 2022, 14(17), 4309; https://doi.org/10.3390/rs14174309 - 1 Sep 2022
Cited by 5 | Viewed by 2251
Abstract
The FY-3E/HIRAS-II (Hyperspectral Infrared Atmospheric Sounder-II), as an infrared hyperspectral instrument onboard the world’s first early morning polar-orbiting satellite, plays a major role in improving the accuracy and timeliness of global numerical weather predictions. In order to assess its observation quality, the geometrically, [...] Read more.
The FY-3E/HIRAS-II (Hyperspectral Infrared Atmospheric Sounder-II), as an infrared hyperspectral instrument onboard the world’s first early morning polar-orbiting satellite, plays a major role in improving the accuracy and timeliness of global numerical weather predictions. In order to assess its observation quality, the geometrically, temporally, and spatially matched scene homogeneous HIRAS-II hyperspectral observations were convolved to the channels corresponding to the Himawari-8/AHI (Advanced Himawari Imager) and FY-3E/MERSI-LL (Medium-Resolution Spectral Imager) imagers from 15 March to 21 April 2022, and their brightness temperature deviation characteristics were statistically calculated in this paper. The results show that the HIRAS-II in-orbit observed brightness temperatures are slightly warmer than the AHI observations in all the matched AHI channels (long wave infrared channel 8 to channel 16) with a mean brightness temperature bias less than 0.65 K. The bias of the atmospheric absorption channel is slightly larger than that of the window channel. A standard deviation less than 0.31 K and a correlation coefficient higher than 0.98 in all channels means that the quality of the observation is satisfactory. The thresholds chosen for the colocation approximation factors (e.g., observation geometry angle, scene uniformity, observation azimuth, and observation time) for matching the HIRAS-II with AHI contribute little and negligible uncertainty to the bias assessment, so the difference between the two observed radiations is considered to be mainly from the systematic bias of the two-instrument measurement. Compared with MERSI-LL window channel 5, the observations of both instruments are very close, with a mean bias of 0.002 K and a standard deviation of 0.31 K. The mean brightness temperature bias (HIRAS-II minus MERSI-LL) of the MERSI-LL water vapor channel 4 is 0.66 K with a standard deviation of 0.22 K. The mean brightness temperature bias of channel 6 and channel 7 is 0.63 K (the standard deviation is 0.36 K) and 0.5 K (the standard deviation is 0.3 K), respectively. The biases of channel 4 are significantly and positively correlated with the target scene temperature, and the biases of channel 6 and 7 show a U-shaped change with the increase in the scene temperature, and the biases are smallest (close to 0 K) when the scene temperature is between 250 K and 280 K. The statistical characteristics of the HIRAS-II–MERSI-LL difference vary minimally and almost constantly over a period of time, indicating that the performance of the HIRAS-II instrument is stable. Full article
Show Figures

Figure 1

Back to TopTop