Search

Generative adversarial networks (or GANs) are a specific deep learning architecture often used for different usages, such as data generation or image-to-image translation. In recent years, this structure has gained increased popularity and has been used in different fields. One area of expertise currently in vogue is the use of GANs to produce synthetic aperture radar (SAR) data, and especially expand training datasets for SAR automatic target recognition (ATR). In effect, the complex SAR image formation makes these kind of data rich in information, leading to the use of deep networks in deep learning-based methods. Yet, deep networks also require sufficient data for training. However, contrary to optical images, we generally do not have a substantial number of available SAR images because of their acquisition and labelling cost; GANs are then an interesting tool. Concurrently, how to improve explainability for SAR ATR deep neural networks and how to make their reasoning more transparent have been increasingly explored as model opacity deteriorates trust of users. This paper aims at reviewing how GANs are used with SAR images, but also giving perspectives on how GANs could be used to improve interpretability and explainability of SAR classifiers. [ABSTRACT FROM AUTHOR]

While thunderstorms can pose severe risks to property and life, forecasting remains challenging, even at short lead times, as these often arise in meta-stable atmospheric conditions. In this paper, we examine the question of how well we could perform short-term (up to 180 min) forecasts using exclusively multi-spectral satellite images and past lighting events as data. We employ representation learning based on deep convolutional neural networks in an "end-to-end" fashion. Here, a crucial problem is handling the imbalance of the positive and negative classes appropriately in order to be able to obtain predictive results (which is not addressed by many previous machine-learning-based approaches). The resulting network outperforms previous methods based on physically based features and optical flow methods (similar to operational prediction models) and generalizes across different years. A closer examination of the classifier performance over time and under masking of input data indicates that the learned model actually draws most information from structures in the visible spectrum, with infrared imaging sustaining some classification performance during the night. [ABSTRACT FROM AUTHOR]

At present, convolutional neural networks (CNN) have been widely used in building extraction from remote sensing imagery (RSI), but there are still some bottlenecks. On the one hand, there are so many parameters in the previous network with complex structure, which will occupy lots of memories and consume much time during training process. On the other hand, low-level features extracted by shallow layers and abstract features extracted by deep layers of artificial neural network cannot be fully fused, which leads to an inaccurate building extraction from RSI. To alleviate these disadvantages, a dense residual neural network (DR-Net) was proposed in this paper. DR-Net uses a deeplabv3+Net encoder/decoder backbone, in combination with densely connected convolution neural network (DCNN) and residual network (ResNet) structure. Compared with deeplabv3+net (containing about 41 million parameters) and BRRNet (containing about 17 million parameters), DR-Net contains about 9 million parameters; So, the number of parameters reduced a lot. The experimental results for both the WHU Building Dataset and Massachusetts Building Dataset, DR-Net show better performance in building extraction than other two state-of-the-art methods. Experiments on WHU building data set showed that Intersection over Union (IoU) increased by 2.4% and F1 score increased by 1.4%; in terms of Massachusetts Building Dataset, IoU increased by 3.8% and F1 score increased by 2.9%. [ABSTRACT FROM AUTHOR]

The availability of freshwater is becoming a global concern. Because agricultural consumption has been increasing steadily, the mapping of irrigated areas is key for supporting the monitoring of land use and better management of available water resources. In this paper, we propose a method to automatically detect and map center pivot irrigation systems using U-Net, an image segmentation convolutional neural network architecture, applied to a constellation of PlanetScope images from the Cerrado biome of Brazil. Our objective is to provide a fast and accurate alternative to map center pivot irrigation systems with very high spatial and temporal resolution imagery. We implemented a modified U-Net architecture using the TensorFlow library and trained it on the Google cloud platform with a dataset built from more than 42,000 very high spatial resolution PlanetScope images acquired between August 2017 and November 2018. The U-Net implementation achieved a precision of 99% and a recall of 88% to detect and map center pivot irrigation systems in our study area. This method, proposed to detect and map center pivot irrigation systems, has the potential to be scaled to larger areas and improve the monitoring of freshwater use by agricultural activities. [ABSTRACT FROM AUTHOR]

The Data Interpolating Empirical Orthogonal Functions (DINEOF) method has demonstrated usability and accuracy for filling spatial gaps in remote sensing datasets. In this study, we conducted the reconstruction of the chlorophyll-a concentration (Chl-a) data using a convolutional neural networks model called Data-Interpolating Convolutional Auto-Encoder (DINCAE), and we compared its performance with that of DINEOF. Furthermore, the cloud-free sea surface temperature (SST) was used as a phytoplankton dynamics predictor for the Chl-a reconstruction. Finally, four reconstruction schemes were implemented: DINCAE (Chl-a only), DINCAE (Chl-a and SST), DINEOF (Chl-a only), and DINEOF (Chl-a and SST), denoted rec1, rec2, rec3, and rec4 respectively. To quantitatively evaluate the accuracy of these reconstruction schemes, both the cross-validation and in situ data were used. The study domain was chosen to be the Northern South China Sea (SCS) and West Philippine Sea (WPS), bounded by 115–125°E and 16–24°N to test the model performance for the reconstruction of Chl-a under different Chl-a controlling mechanisms. The in situ validation showed that rec1 performs best among the four reconstruction schemes, and that adding SST into the Chl-a reconstruction cannot improve the reconstruction results. However, for cross validation, adding SST can slightly improve spatial distributions of the root mean square error (RMSE) between the reconstructed data and the original data, especially over the SCS continental shelf. Furthermore, the potential of DINCAE prediction is confirmed in this paper; thus, the trained DINCAE model can be re-applied to reconstruct other missing data, and more importantly, it can also be re-trained using the reconstructed data, thereby further improving reconstruction results. Another consideration is efficiency; with similar reconstruction conditions, DINCAE is 5–10 times faster than DINEOF. [ABSTRACT FROM AUTHOR]

Multisource remote sensing data acquisition has been increased in the last years due to technological improvements and decreased acquisition cost of remotely sensed data and products. This study attempts to fuse different types of prospection data acquired from dissimilar remote sensors and explores new ways of interpreting remote sensing data obtained from archaeological sites. Combination and fusion of complementary sensory data does not only increase the detection accuracy but it also increases the overall performance in respect to recall and precision. Moving beyond the discussion and concerns related to fusion and integration of multisource prospection data, this study argues their potential (re)use based on Bayesian Neural Network (BNN) fusion models. The archaeological site of Vésztő-Mágor Tell in the eastern part of Hungary was selected as a case study, since ground penetrating radar (GPR) and ground spectral signatures have been collected in the past. GPR 20 cm depth slices results were correlated with spectroradiometric datasets based on neural network models. The results showed that the BNN models provide a global correlation coefficient of up to 73%—between the GPR and the spectroradiometric data—for all depth slices. This could eventually lead to the potential re-use of archived geo-prospection datasets with optical earth observation datasets. A discussion regarding the potential limitations and challenges of this approach is also included in the paper. [ABSTRACT FROM AUTHOR]

Forest environment classification in mountain regions based on single-sensor remote sensing approaches is hindered by forest complexity and topographic effects. Temperate broadleaf forests in western Asia such as the Hyrcanian forest in northern Iran have already suffered from intense anthropogenic activities. In those regions, forests mainly extend in rough terrain and comprise different stand structures, which are difficult to discriminate. This paper explores the joint analysis of Landsat7/ETM+, L-band SAR and their derived parameters and the effect of terrain corrections to overcome the challenges of discriminating forest stand age classes in mountain regions. We also verified the performances of three machine learning methods which have recently shown promising results using multisource data; support vector machines (SVM), neural networks (NN), random forest (RF) and one traditional classifier (i.e., maximum likelihood classification (MLC)) as a benchmark. The non-topographically corrected ETM+ data failed to differentiate among different forest stand age classes (average classification accuracy (OA) = 65%). This confirms the need to reduce relief effects prior data classification in mountain regions. SAR backscattering alone cannot properly differentiate among different forest stand age classes (OA = 62%). However, textures and PolSAR features are very efficient for the separation of forest classes (OA = 82%). The highest classification accuracy was achieved by the joint usage of SAR and ETM+ (OA = 86%). However, this shows a slight improvement compared to the ETM+ classification (OA = 84%). The machine learning classifiers proved t o be more robust and accurate compared to MLC. SVM and RF statistically produced better classification results than NN in the exploitation of the considered multi-source data. [ABSTRACT FROM AUTHOR]

We describe the application of a Neural Network (NN) previously developed by us, to the detection and tracking, of Karenia brevis Harmful Algal Blooms (KB HABs) that plague the coasts of the West Florida Shelf (WFS) using Visible Infrared Imaging Radiometer Suite (VIIRS) satellite observations. Previous approaches for the detection of KB HABs in the WFS primarily used observations from the Moderate Resolution Imaging Spectroradiometer Aqua (MODIS-A) satellite. They depended on the remote sensing reflectance signal at the 678 nm chlorophyll fluorescence band (Rrs678) needed for both the normalized fluorescence height (nFLH) and Red Band Difference algorithms (RBD) currently used. VIIRS which has replaced MODIS-A, unfortunately does not have a 678 nm fluorescence channel so we customized the NN approach to retrieve phytoplankton absorption at 443 nm (aph443) using only Rrs measurements from existing VIIRS channels at 486, 551 and 671 nm. The aph443 values in these retrieved VIIRS images, can in turn be correlated to chlorophyll-a concentrations [Chla] and KB cell counts. To retrieve KB values, the VIIRS NN retrieved aph443 images are filtered by applying limiting constraints, defined by (i) low backscatter at Rrs 551 nm and (ii) a minimum aph443 value known to be associated with KB HABs in the WFS. The resulting filtered residual images, are then used to delineate and quantify the existing KB HABs. Comparisons with KB HABs satellite retrievals obtained using other techniques, including nFLH, as well as with in situ measurements reported over a four year period, confirm the viability of the NN technique, when combined with the filtering constraints devised, for effective detection of KB HABs. [ABSTRACT FROM AUTHOR]

The use of Landsat data to answer ecological questions is greatly increased by the effective removal of cloud and cloud shadow from satellite images. We develop a novel algorithm to identify and classify clouds and cloud shadow, SPARCS: Spatial Procedures for Automated Removal of Cloud and Shadow. The method uses a neural network approach to determine cloud, cloud shadow, water, snow/ice and clear sky classification memberships of each pixel in a Landsat scene. It then applies a series of spatial procedures to resolve pixels with ambiguous membership by using information, such as the membership values of neighboring pixels and an estimate of cloud shadow locations from cloud and solar geometry. In a comparison with FMask, a high-quality cloud and cloud shadow classification algorithm currently available, SPARCS performs favorably, with substantially lower omission errors for cloud shadow (8.0% and 3.2%), only slightly higher omission errors for clouds (0.9% and 1.3%, respectively) and fewer errors of commission (2.6% and 0.3%). Additionally, SPARCS provides a measure of uncertainty in its classification that can be exploited by other algorithms that require clear sky pixels. To illustrate this, we present an application that constructs obstruction-free composites of images acquired on different dates in support of a method for vegetation change detection. [ABSTRACT FROM AUTHOR]

An extensive record of current velocities at all levels in the water column is an indispensable requirement for a tidal resource assessment and is fully necessary for accurate determination of available energy throughout the water column as well as estimating likely energy capture for any particular device. Traditional tidal prediction using the least squares method requires a large number of harmonic parameters calculated from lengthy acoustic Doppler current profiler (ADCP) measurements, while long-term in situ ADCPs have the advantage of measuring the real current but are logistically expensive. This study aims to show how these issues can be overcome with the use of a neural network to predict current velocities throughout the water column, using surface currents measured by a high-frequency radar. Various structured neural networks were trained with the aim of finding the network which could best simulate unseen subsurface current velocities, compared to ADCP data. This study shows that a recurrent neural network, trained by the Bayesian regularisation algorithm, produces current velocities highly correlated with measured values: r2 (0.98), mean absolute error (0.05 ms−1), and the Nash–Sutcliffe efficiency (0.98). The method demonstrates its high prediction ability using only 2 weeks of training data to predict subsurface currents up to 6 months in the future, whilst a constant surface current input is available. The resulting current predictions can be used to calculate flow power, with only a 0.4% mean error. The method is shown to be as accurate as harmonic analysis whilst requiring comparatively few input data and outperforms harmonics by identifying non-celestial influences; however, the model remains site specific. [ABSTRACT FROM AUTHOR]

To precisely forecast downstream water levels in catchment areas during typhoons, the deep learning artificial neural networks were employed to establish two water level forecasting models using sequential neural networks (SNNs) and multiple-input functional neural networks (MIFNNs). SNNs, which have a typical neural network structure, are network models constructed using sequential methods. To develop a network model capable of flexibly consolidating data, MIFNNs are employed for processing data from multiple sources or with multiple dimensions. Specifically, when images (e.g., radar reflectivity images) are used as input attributes, feature extraction is required to provide effective feature maps for model training. Therefore, convolutional layers and pooling layers were adopted to extract features. Long short-term memory (LSTM) layers adopted during model training enabled memory cell units to automatically determine the memory length, providing more useful information. The Hsintien River basin in northern Taiwan was selected as the research area and collected relevant data from 2011 to 2019. The input attributes comprised one-dimensional data (e.g., water levels at river stations, rain rates at rain gauges, and reservoir release) and two-dimensional data (i.e., radar reflectivity mosaics). Typhoons Saola, Soudelor, Dujuan, and Megi were selected, and the water levels 1 to 6 h after the typhoons struck were forecasted. The results indicated that compared with linear regressions (REG), SNN using dense layers (SNN-Dense), and SNN using LSTM layers (SNN-LSTM) models, superior forecasting results were achieved for the MIFNN model. Thus, the MIFNN model, as the optimal model for water level forecasting, was identified. [ABSTRACT FROM AUTHOR]

Surface skin temperature (Tskin) derived from infrared remote sensors mounted on board satellites provides a continuous observation of Earth's surface and allows the monitoring of global temperature change relevant to climate trends. In this study, we present a fast retrieval method for retrieving Tskin based on an artificial neural network (ANN) from a set of spectral channels selected from the Infrared Atmospheric Sounding Interferometer (IASI) using the information theory/entropy reduction technique. Our IASI Tskin product (i.e., TANN) is evaluated against Tskin from EUMETSAT Level 2 product, ECMWF Reanalysis (ERA5), SEVIRI observations, and ground in situ measurements. Good correlations between IASI TANN and the Tskin from other datasets are shown by their statistic data, such as a mean bias and standard deviation (i.e., [bias, STDE]) of [0.55, 1.86 °C], [0.19, 2.10 °C], [−1.5, 3.56 °C], from EUMETSAT IASI L-2 product, ERA5, and SEVIRI. When compared to ground station data, we found that all datasets did not achieve the needed accuracy at several months of the year, and better results were achieved at nighttime. Therefore, comparison with ground-based measurements should be done with care to achieve the ±2 °C accuracy needed, by choosing, for example, a validation site near the station location. On average, this accuracy is achieved, in particular at night, leading to the ability to construct a robust Tskin dataset suitable for Tskin long-term spatio-temporal variability and trend analysis. [ABSTRACT FROM AUTHOR]

A large variety of sound sources in the ocean, including biological, geophysical, and man-made, can be simultaneously monitored over instantaneous continental-shelf scale regions via the passive ocean acoustic waveguide remote sensing (POAWRS) technique by employing a large-aperture densely-populated coherent hydrophone array system. Millions of acoustic signals received on the POAWRS system per day can make it challenging to identify individual sound sources. An automated classification system is necessary to enable sound sources to be recognized. Here, the objectives are to (i) gather a large training and test data set of fin whale vocalization and other acoustic signal detections; (ii) build multiple fin whale vocalization classifiers, including a logistic regression, support vector machine (SVM), decision tree, convolutional neural network (CNN), and long short-term memory (LSTM) network; (iii) evaluate and compare performance of these classifiers using multiple metrics including accuracy, precision, recall and F1-score; and (iv) integrate one of the classifiers into the existing POAWRS array and signal processing software. The findings presented here will (1) provide an automatic classifier for near real-time fin whale vocalization detection and recognition, useful in marine mammal monitoring applications; and (2) lay the foundation for building an automatic classifier applied for near real-time detection and recognition of a wide variety of biological, geophysical, and man-made sound sources typically detected by the POAWRS system in the ocean. [ABSTRACT FROM AUTHOR]

The forecasting and reconstruction of oceanic dynamics is a crucial challenge. While model driven strategies are still the state-of-the-art approaches in the reconstruction of spatio-temporal dynamics. The ever increasing availability of data collections in oceanography raised the relevance of data-driven approaches as computationally efficient representations of spatio-temporal fields reconstruction. This tools proved to outperform classical state-of-the-art interpolation techniques such as optimal interpolation and DINEOF in the retrievement of fine scale structures while still been computationally efficient comparing to model based data assimilation schemes. However, coupling this data-driven priors to classical filtering schemes limits their potential representativity. From this point of view, the recent advances in machine learning and especially neural networks and deep learning can provide a new infrastructure for dynamical modeling and interpolation within a data-driven framework. In this work we adress this challenge and develop a novel Neural-Network-based (NN-based) Kalman filter for spatio-temporal interpolation of sea surface dynamics. Based on a data-driven probabilistic representation of spatio-temporal fields, our approach can be regarded as an alternative to classical filtering schemes such as the ensemble Kalman filters (EnKF) in data assimilation. Overall, the key features of the proposed approach are two-fold: (i) we propose a novel architecture for the stochastic representation of two dimensional (2D) geophysical dynamics based on a neural networks, (ii) we derive the associated parametric Kalman-like filtering scheme for a computationally-efficient spatio-temporal interpolation of Sea Surface Temperature (SST) fields. We illustrate the relevance of our contribution for an OSSE (Observing System Simulation Experiment) in a case-study region off South Africa. Our numerical experiments report significant improvements in terms of reconstruction performance compared with operational and state-of-the-art schemes (e.g., optimal interpolation, Empirical Orthogonal Function (EOF) based interpolation and analog data assimilation). [ABSTRACT FROM AUTHOR]

This work presents quantitative prediction of severity of the disease caused by Phytophthora infestans in potato crops using machine learning algorithms such as multilayer perceptron, deep learning convolutional neural networks, support vector regression, and random forests. The machine learning algorithms are trained using datasets extracted from multispectral data captured at the canopy level with an unmanned aerial vehicle, carrying an inexpensive digital camera. The results indicate that deep learning convolutional neural networks, random forests and multilayer perceptron using band differences can predict the level of Phytophthora infestans affectation on potato crops with acceptable accuracy. [ABSTRACT FROM AUTHOR]

Knowledge of tree species composition is obligatory in forest management. Accurate tree species maps allow for detailed analysis of a forest ecosystem and its interactions with the environment. The research presented here focused on developing methods of tree species identification using aerial hyperspectral data. The research area is located in Southwestern Poland and covers the Karkonoski National Park (KNP), which was significantly damaged by acid rain and pest infestation in the 1980s. High-resolution (3.35 m) Airborne Prism Experiment (APEX) hyperspectral images (288 spectral bands in the range of 413 to 2440 nm) were used as a basis for tree species classification. Beech (Fagus sylvatica), birch (Betula pendula), alder (Alnus incana), larch (Larix decidua), pine (Pinus sylvestris), and spruce (Picea abies) were classified. The classification algorithm used was feed-forward multilayered perceptron (MLP) with a single hidden layer. To simulate such a network, we used the R programming environment and the nnet package. To provide more accurate measurement of accuracy, iterative accuracy assessment was performed. The final tree species maps cover the whole area of KNP; a median overall accuracy (OA) of 87% was achieved, with median producer accuracy (PA) for all classes exceeding 68%. The best-classified classes were spruce, beech, and birch, with median producer accuracy of 93%, 88% and 83%, respectively. The pine class achieved the lowest median producer and user accuracies (68% and 75%, respectively). The results show great potential for the use of hyperspectral data as a tool for identifying tree species locations in diverse mountainous forest. [ABSTRACT FROM AUTHOR]