Lithology Classification Using TASI Thermal Infrared Hyperspectral Data with Convolutional Neural Networks

In recent decades, lithological mapping techniques using hyperspectral remotely sensed imagery have developed rapidly. The processing chains using visible-near infrared (VNIR) and shortwave infrared (SWIR) hyperspectral data are proven to be available in practice. The thermal infrared (TIR) portion of the electromagnetic spectrum has considerable potential for mineral and lithology mapping. In particular, the abovementioned rocks at wavelengths of 8–12 μm were found to be discriminative, which can be seen as a characteristic to apply to lithology classification. Moreover, it was found that most of the lithology mapping and classification for hyperspectral thermal infrared data are still carried out by traditional spectral matching methods, which are not very reliable due to the complex diversity of geological lithology. In recent years, deep learning has made great achievements in hyperspectral imagery classification feature extraction. It usually captures abstract features through a multilayer network, especially convolutional neural networks (CNNs), which have received more attention due to their unique advantages. Hence, in this paper, lithology classification with CNNs was tested on thermal infrared hyperspectral data using a Thermal Airborne Spectrographic Imager (TASI) at three small sites in Liuyuan, Gansu Province, China. Three different CNN algorithms, including one-dimensional CNN (1-D CNN), two-dimensional CNN (2-D CNN) and three-dimensional CNN (3-D CNN), were implemented and compared to the six relevant state-of-the-art methods. At the three sites, the maximum overall accuracy (OA) based on CNNs was 94.70%, 96.47% and 98.56%, representing improvements of 22.58%, 25.93% and 16.88% over the worst OA. Meanwhile, the average accuracy of all classes (AA) and kappa coefficient (kappa) value were consistent with the OA, which confirmed that the focal method effectively improved accuracy and outperformed other methods.

Download Full-text

Wildfire detection and monitoring by using PRISMA hyperspectral data and convolutional neural networks

10.5194/egusphere-egu21-12330 ◽

2021 ◽

Author(s):

Dario Spiller ◽

Luigi Ansalone ◽

Nicolas Longépé ◽

James Wheeler ◽

Pierre Philippe Mathieu

Keyword(s):

Remote Sensing ◽

Neural Networks ◽

Deep Learning ◽

Convolutional Neural Networks ◽

Near Infrared ◽

Fire Detection ◽

Input Image ◽

Hyperspectral Data ◽

Observation Data ◽

Hyperspectral Camera

Over the last few years, wildfires have become more severe and destructive, having extreme consequences on local and global ecosystems. Fire detection and accurate monitoring of risk areas is becoming increasingly important. Satellite remote sensing offers unique opportunities for mapping, monitoring, and analysing the evolution of wildfires, providing helpful contributions to counteract dangerous situations.Among the different remote sensing technologies, hyper-spectral (HS) imagery presents nonpareil features in support to fire detection. In this study, HS images from the Italian satellite PRISMA (PRecursore IperSpettrale della Missione Applicativa) will be used. The PRISMA satellite, launched on 22 March 2019, holds a hyperspectral and panchromatic&#160; payload which is able to acquire images with a worldwide coverage. The hyperspectral camera works in the spectral range of 0.4&#8211;2.5 &#181;m, with 66 and 173 channels in the VNIR (Visible and Near InfraRed) and SWIR (Short-Wave InfraRed) regions, respectively. The average spectral resolution is less than 10 nm on the entire range with an accuracy of &#177;0.1 nm, while the ground sampling distance of PRISMA images is about 5 m and 30 m for panchromatic and hyperspectral camera, respectively.This work will investigate how PRISMA HS images can be used to support fire detection and related crisis management. To this aim, deep learning methodologies will be investigated, as 1D convolutional neural networks to perform spectral analysis of the data or 3D convolutional neural networks to perform spatial and spectral analyses at the same time. Semantic segmentation of input HS data will be discussed, where an output image with metadata will be associated to each pixels of the input image. The overall goal of this work is to highlight how PRISMA hyperspectral data can contribute to remote sensing and Earth-observation data analysis with regard to natural hazard and risk studies focusing specially on wildfires, also considering the benefits with respect to standard multi-spectral imagery or previous hyperspectral sensors such as Hyperion.The contributions of this work to the state of the art are the following:<ul><li>Demonstrating the advantages of using PRISMA HS data over using multi-spectral data.</li> <li>Discussing the potentialities of deep learning methodologies based on 1D and 3D convolutional neural networks to catch spectral (and spatial for the 3D case) dependencies, which is crucial when dealing with HS images.</li> <li>Discussing the possibility and benefit to integrate HS-based approach in future monitoring systems in case of wildfire alerts and disasters.</li> <li>Discussing the opportunity to design and develop future missions for HS remote sensing specifically dedicated for fire detection with on-board analysis.</li> </ul>To conclude, this work will raise awareness in the potentialities of using PRISMA HS data for disasters monitoring with specialized focus on wildfires.

Download Full-text

Influence of the daylight illumination a weather conditions on Airborne Thermal Infrared Hyperspectral geological mapping

10.5194/egusphere-egu21-7494 ◽

2021 ◽

Author(s):

Stephane Boubanga Tombet ◽

Jean-Philippe Gagnon ◽

Holger Eichstaedt ◽

Joanne Ho

Keyword(s):

Near Infrared ◽

Thermal Infrared ◽

Hyperspectral Data ◽

Geological Mapping ◽

Electromagnetic Spectrum ◽

Spectral Features ◽

Absorption Bands ◽

Mineral Mapping ◽

Atmospheric Compensation ◽

Spectral Ranges

The use of airborne remote sensing techniques for geological mapping offers many benefits as it allows coverage of large areas in a very efficient way. &#160;While hyperspectral imaging from airborne/spaceborne platforms is now a well-established method applied to resolve many geological problems, it has mostly been developed only in the Visible-Near Infrared (VNIR, 0.4&#8211;1.0 mm) and Shortwave Infrared (SWIR, 1.0&#8211;2.5 mm) regions of the electromagnetic spectrum. However, the reflectance spectral features measured in the VNIR and SWIR spectral ranges are generally overtones and combination bands from fundamental absorption bands at longer wavelengths, such as in the Longwave Infrared (LWIR, 8&#8211;12 mm). The single absorption bands in the VNIR and SWIR spectral ranges are often very closely spaced so that the reflectance features measured by common spectrometers in this spectral region are typically broad and/or suffer from strong overlapping, which raises selectivity issues for mineral identification in some cases.The inherent self-emission associated with LWIR under ambient conditions allows airborne mineral mapping in various weather (cloudy, partly cloudy or clear sky) and illumination (day or night) conditions. For this reason, LWIR often refers to the thermal infrared (TIR) spectral range. Solid targets such as minerals not only emit but also reflect TIR radiation. Since the two phenomena occur simultaneously, they end-up mixed in the radiance measured at the sensor level. The spectral features observed in a TIR spectrum of the sky and the atmosphere mostly correspond to ozone, water&#160; vapor, carbon dioxide, methane and nitrous oxide with pretty sharp and narrow features compared with the infrared signature of solid materials such as minerals. The sharp spectral features of atmospheric gases are mixed up with broad minerals features in the collected geological mapping data, to unveil the spectral features associated with minerals from TIR measurements, the respective contributions of self-emission and reflection in the measurement must be &#171;unmixed&#187; and the atmospheric contributions must be compensated. This procedure refers to temperature-emissivity separation (TES). Therefore, to achieve an efficient TES and atmospheric compensation, the collection time and conditions of LWIR airborne hyperspectral data is of importance. Data of a flight mission in Southern Spain collected systematically at different times of the day (morning, mid-day and night) and in different altitudes using the Telops Hyper-Cam airborne system, a passive TIR hyperspectral sensor based on Fourier transform spectroscopy, were analyzed. TES was carried out on the hyperspectral data using two different approaches: a) Telops Reveal FLAASH IR software and b) DIMAP In-scene atmospheric compensation algorithm in order to retrieve thermodynamic temperature map and spectral emissivity data. Spectral analysis of the emissivity data with different mineral mapping methods based on commercial spectral libraries was used to compare results obtained during the different flight times and altitudes using the two post-processing methodologies. The results are discussed in the light of optimizing LWIR-based airborne operations in time and altitude to achieve best results for routine field mineral mapping applications such as in mining, soil science or archaeology, where the spatial analysis of mineral and chemical distribution is essential

Download Full-text

Convolutional Neural Networks for Semantic Segmentation as a Tool for Multiclass Face Analysis in Thermal Infrared

Journal of Nondestructive Evaluation ◽

10.1007/s10921-020-00740-y ◽

2021 ◽

Vol 40 (1) ◽

Author(s):

David Müller ◽

Andreas Ehlen ◽

Bernd Valeske

Keyword(s):

Neural Networks ◽

Convolutional Neural Networks ◽

Semantic Segmentation ◽

Thermal Infrared ◽

Relative Temperature ◽

Face Analysis ◽

Image Translation ◽

The Face ◽

Live Image ◽

Quantitative Thermography

AbstractConvolutional neural networks were used for multiclass segmentation in thermal infrared face analysis. The principle is based on existing image-to-image translation approaches, where each pixel in an image is assigned to a class label. We show that established networks architectures can be trained for the task of multiclass face analysis in thermal infrared. Created class annotations consisted of pixel-accurate locations of different face classes. Subsequently, the trained network can segment an acquired unknown infrared face image into the defined classes. Furthermore, face classification in live image acquisition is shown, in order to be able to display the relative temperature in real-time from the learned areas. This allows a pixel-accurate temperature face analysis e.g. for infection detection like Covid-19. At the same time our approach offers the advantage of concentrating on the relevant areas of the face. Areas of the face irrelevant for the relative temperature calculation or accessories such as glasses, masks and jewelry are not considered. A custom database was created to train the network. The results were quantitatively evaluated with the intersection over union (IoU) metric. The methodology shown can be transferred to similar problems for more quantitative thermography tasks like in materials characterization or quality control in production.

Download Full-text

Generation and Annotation of Simulation-Real Ship Images for Convolutional Neural Networks Training and Testing

Applied Sciences ◽

10.3390/app11135931 ◽

2021 ◽

Vol 11 (13) ◽

pp. 5931

Author(s):

Ji’an You ◽

Zhaozheng Hu ◽

Chao Peng ◽

Zhiqiang Wang

Keyword(s):

Neural Networks ◽

Convolutional Neural Networks ◽

Image Annotation ◽

Three Dimensional ◽

Image Data ◽

Detection Algorithm ◽

Simulation Software ◽

High Quality ◽

Annotation Method ◽

Detection Of Objects

Large amounts of high-quality image data are the basis and premise of the high accuracy detection of objects in the field of convolutional neural networks (CNN). It is challenging to collect various high-quality ship image data based on the marine environment. A novel method based on CNN is proposed to generate a large number of high-quality ship images to address this. We obtained ship images with different perspectives and different sizes by adjusting the ships’ postures and sizes in three-dimensional (3D) simulation software, then 3D ship data were transformed into 2D ship image according to the principle of pinhole imaging. We selected specific experimental scenes as background images, and the target ships of the 2D ship images were superimposed onto the background images to generate “Simulation–Real” ship images (named SRS images hereafter). Additionally, an image annotation method based on SRS images was designed. Finally, the target detection algorithm based on CNN was used to train and test the generated SRS images. The proposed method is suitable for generating a large number of high-quality ship image samples and annotation data of corresponding ship images quickly to significantly improve the accuracy of ship detection. The annotation method proposed is superior to the annotation methods that label images with the image annotation software of Label-me and Label-img in terms of labeling the SRS images.

Download Full-text

Atmospheric Compensation of PRISMA Data by Means of a Learning Based Approach

Remote Sensing ◽

10.3390/rs13152967 ◽

2021 ◽

Vol 13 (15) ◽

pp. 2967

Author(s):

Nicola Acito ◽

Marco Diani ◽

Gregorio Procissi ◽

Giovanni Corsini

Keyword(s):

Spectral Range ◽

Near Infrared ◽

Random Noise ◽

Synthetic Data ◽

Regression Function ◽

Short Wave ◽

Hyperspectral Data ◽

Electromagnetic Spectrum ◽

Great Opportunity ◽

Atmospheric Compensation

Atmospheric compensation (AC) allows the retrieval of the reflectance from the measured at-sensor radiance and is a fundamental and critical task for the quantitative exploitation of hyperspectral data. Recently, a learning-based (LB) approach, named LBAC, has been proposed for the AC of airborne hyperspectral data in the visible and near-infrared (VNIR) spectral range. LBAC makes use of a parametric regression function whose parameters are learned by a strategy based on synthetic data that accounts for (1) a physics-based model for the radiative transfer, (2) the variability of the surface reflectance spectra, and (3) the effects of random noise and spectral miscalibration errors. In this work we extend LBAC with respect to two different aspects: (1) the platform for data acquisition and (2) the spectral range covered by the sensor. Particularly, we propose the extension of LBAC to spaceborne hyperspectral sensors operating in the VNIR and short-wave infrared (SWIR) portion of the electromagnetic spectrum. We specifically refer to the sensor of the PRISMA (PRecursore IperSpettrale della Missione Applicativa) mission, and the recent Earth Observation mission of the Italian Space Agency that offers a great opportunity to improve the knowledge on the scientific and commercial applications of spaceborne hyperspectral data. In addition, we introduce a curve fitting-based procedure for the estimation of column water vapor content of the atmosphere that directly exploits the reflectance data provided by LBAC. Results obtained on four different PRISMA hyperspectral images are presented and discussed.

Download Full-text

Identification of Mode Shapes of a Composite Cylinder Using Convolutional Neural Networks

Materials ◽

10.3390/ma14112801 ◽

2021 ◽

Vol 14 (11) ◽

pp. 2801

Author(s):

Bartosz Miller ◽

Leonard Ziemiański

Keyword(s):

Neural Networks ◽

Convolutional Neural Networks ◽

Composite Structures ◽

Three Dimensional ◽

Mode Shapes ◽

Identification Algorithm ◽

Material Degradation ◽

Local Material ◽

Noisy Input ◽

Shape Vector

The aim of the following paper is to discuss a newly developed approach for the identification of vibration mode shapes of multilayer composite structures. To overcome the limitations of the approaches based on image analysis (two-dimensional structures, high spatial resolution of mode shapes description), convolutional neural networks (CNNs) are applied to create a three-dimensional mode shapes identification algorithm with a significantly reduced number of mode shape vector coordinates. The CNN-based procedure is accurate, effective, and robust to noisy input data. The appearance of local damage is not an obstacle. The change of the material and the occurrence of local material degradation do not affect the accuracy of the method. Moreover, the application of the proposed identification method allows identifying the material degradation occurrence.

Download Full-text