Search

Neural architecture search (NAS) is a subset of automated machine learning that tries to find the best neural network to perform a given task. In this article, a network search space is defined and applied to perform the semantic segmentation of satellite imagery. Due to the spatial nature of the data, the search space uses cells that group parallel operations with kernels of different sizes, providing options to accommodate the neighborhood information required to perform a better classification. The architecture search space follows a UNet-like network. The proposed approach uses scaled sigmoid gates, a strategy for network pruning that was adapted to search for the best operations on the cell search space. The architecture achieved by the proposed approach uses wider kernels on lower resolution feature maps, which leads to the interpretation that some pixels required information from pixels farther away than expected. The resulting network was compared to a very similar UNet-like network that only used 3×3 convolutions. The resulting network shows slightly better results on the test set.

This paper reports the results of a study that aims to develop semi-automatic methods for assessing the degree of corrosion in industrial plant. We evaluated two fully convolutional networks (U-Net and DeepLab v3 +) to segment corroded areas in panoramic images of offshore platforms. The experimental analysis was based on two datasets built for this study. The datasets comprise 9,112 2D images and 3,732 panoramic images. Both FCNs trained on 2D images were tested on 2D images and cubic projections of panoramic images. In addition to pointing out encouraging results, the experiments indicated that most prediction errors concentrated in corrosion defects with a small pixel area.

Recent works have studied crop recognition in regions with highly complex spatio-temporal dynamics typical of a tropical climate. However, most proposals have only been evaluated in a single agricultural year, and their capabilities to generalize to dates outside the temporal sequence have not been properly addressed thus far. This work assesses the generalization capabilities of a recent convolutional recurrent architecture, testing it in a temporal sequence two years ahead of the sequence with which it was trained. Furthermore, a N-to-1 variant of such network is proposed, which is able to produce classification outcomes for every month in the agricultural year, and it is compared with two baselines designed in a more traditional approach, in which a separate specific network is trained for each month of the year. The approaches are evaluated on two public datasets from a tropical region. The first dataset comprehends the period from June 2017 to May 2018, while the second goes from October 2019 to September 2020. Results show a decrease of up to 24.6% in per-date average F1 score when training the network with data of an agricultural year different from the one it is tested on, which indicates a domain shift that demands further research. Additionally, the proposed approach presented only a slight decrease in performance compared to its baseline when trained on the same dataset, with a 2.7% drop in average F1 score. This performance drop is a small cost in exchange for its operational advantages, such as reduced training time and a more straightforward pipeline.

Knowing the spatial distribution of endangered tree species in a forest ecosystem or forest remnants is a valuable information to support environmental conservation practices. The use of Unmanned Aerial Vehicles (UAVs) offers a suitable alternative for this task, providing very high-resolution images at low costs. In parallel, recent advances in the computer vision field have led to the development of effective deep learning techniques for end-to-end semantic image segmentation. In this scenario, the DeepLabv3+ is well established as the state-of-the-art deep learning method for semantic segmentation tasks. The present paper proposes and assesses the use of DeepLabv3+ for mapping the threatened Dipteryx alata Vogel tree, popularly also known as cumbaru. We also compare two backbone networks for feature extraction in the DeepLabv3+ architecture: the Xception and MobileNetv2. Experiments carried out on a dataset consisting of 225 UAV/RGB images of an urban area in Midwest Brazil demonstrated that DeepLabv3+ was able to achieve in mean overall accuracy and F1-score above 90%, and IoU above 80%. The experimental analysis also pointed out that the MobileNetv2 backbone overcame its counterpart by a wide margin due to its comparatively simpler architecture in view of the available training data.

Applying remote sensing technology to map and monitor agriculture and its impacts can greatly contribute for the proper development of this activity, promoting efficient food, fiber and energy production. For that, not only remote sensing images are needed, but also ground truth information, which is a key factor for the development and improvement of methodologies using remote sensing data. While a variety of images are current available, inclusive cost-free images, field reference data is scarcer. For agricultural applications, especially in tropical regions such as Brazil, where the agriculture is very dynamic and diverse (recent agricultural frontiers, crop rotations, multiple cropping systems, several management practices, etc.), and cultivated over a vast territory, this task is not trivial. One way of boosting the researches in agricultural remote sensing is to stimulate people to share their data, and to foster different groups to use the same dataset, so distinct methods can be properly compared. In this context, our group created the LEM Benchmark Database (a project funded by the ISPRS Scientific Initiative project - 2017) from the Luiz Eduardo Magalhães (LEM) municipality, Bahia State, Brazil. The database contains a set of pre-processed multitemporal satellite images (Landsat-8/OLI, Sentinel-2/MSI and SAR band-C Sentinel-1) and shapefiles of agricultural fields with their correspondent monthly land use classes, covering the period of one Brazilian crop year (2017–2018). In this paper we present the first results obtained with this database.

Deep learning classification models require large amounts of labeled training data to perform properly, but the production of reference data for most Earth observation applications is a labor intensive, costly process. In that sense, transfer learning is an option to mitigate the demand for labeled data. In many remote sensing applications, however, the accuracy of a deep learning-based classification model trained with a specific dataset drops significantly when it is tested on a different dataset, even after fine-tuning. In general, this behavior can be credited to the domain shift phenomenon. In remote sensing applications, domain shift can be associated with changes in the environmental conditions during the acquisition of new data, variations of objects’ appearances, geographical variability and different sensor properties, among other aspects. In recent years, deep learning-based domain adaptation techniques have been used to alleviate the domain shift problem. Recent improvements in domain adaptation technology rely on techniques based on Generative Adversarial Networks (GANs), such as the Cycle-Consistent Generative Adversarial Network (CycleGAN), which adapts images across different domains by learning nonlinear mapping functions between the domains. In this work, we exploit the CycleGAN approach for domain adaptation in a particular change detection application, namely, deforestation detection in the Amazon forest. Experimental results indicate that the proposed approach is capable of alleviating the effects associated with domain shift in the context of the target application.

Climate change and its effects are taking more importance nowadays; and glaciers are one of the most affected ecosystems by that, considering that the energy of Earth’s surface and its temperature may be directly related to glacier temporal changes. Then, the comprehension of glaciers behaviour, by its retreating or melting critical conditions, can be achieved by the analysis of Remote Sensing data, but considering the unprecedented volumes of information currently provided by satellites sensors, we can refer to this analysis as a big data problem. Machine learning techniques have the potential to improve the analysis of this type of data; however, most current machine learning algorithms are unable to properly process such huge volumes of data. In the attempt to overcome the computational limitations related to Remote Sensing Big Data analysis, we implemented the K-Means and Expectation Maximization algorithms, as distributed clustering solutions, exploiting the capabilities of cloud computing infrastructure for processing very large datasets. The solution was developed over the InterCloud Data Mining Package, which is a suite of distributed classification methods, previously employed in hyperspectral image analysis. In this work we extended the functionalities of that package, by making it able to process multispectral images using the aforementioned clustering algorithms. To validate our proposal, we analysed the Ausangate glacier, located on the Andes Mountains, in Peru, by mapping the changes in such environment through a multi-temporal Remote Sensing analysis. Our results and conclusions are focused on the thematic accuracy and the computational performance achieved by our proposed solution. Thematic accuracy was assessed by comparing the automatically detected glacier areas by the clustering approaches against the manually selected ground truth data. We compared the computational load involved in executing the clustering processes sequentially and in a distributed fashion, using a local mode and cluster configuration over a cloud computing infrastructure.

This work aims at investigating unsupervised and semi-supervised representation learning methods based on generative adversarial networks for remote sensing scene classification. The work introduces a novel approach, which consists in a semi-supervised extension of a prior unsupervised method, known as MARTA-GAN. The proposed approach was compared experimentally with two baselines upon two public datasets, UC-MERCED and NWPU-RESISC45. The experiments assessed the performance of each approach under different amounts of labeled data. The impact of fine-tuning was also investigated. The proposed method delivered in our analysis the best overall accuracy under scarce labeled samples, both in terms of absolute value and in terms of variability across multiple runs.

Recently, recurrent neural networks have been proposed for crop mapping from multitemporal remote sensing data. Most of these proposals have been designed and tested in temperate regions, where a single harvest per season is the rule. In tropical regions, the favorable climate and local agricultural practices, such as crop rotation, result in more complex spatio-temporal dynamics, where the single harvest per season assumption does not hold. In this context, a demand arises for methods capable of recognizing agricultural crops at multiple dates along the multitemporal sequence. In the present work, we propose to adapt two recurrent neural networks, originally conceived for single harvest per season, for multidate crop recognition. In addition, we propose a novel multidate approach based on bidirectional fully convolutional recurrent neural networks. These three architectures were evaluated on public Sentinel-1 data sets from two tropical regions in Brazil. In our experiments, all methods achieved state-of-the-art accuracies with a clear superiority of the proposed architecture. It outperformed its counterparts in up to 3.8% and 7.4%, in terms of per-month overall accuracy, and it was the best performing method in terms of F1-score for most crops and dates on both regions.

This work follows a line of research dedicated to the parallelization of image segmentation algorithms on distributed computing environments, which is motivated by the increasing resolutions and availability of Remote Sensing (RS) images. Here we focus on region-growing segmentation, which is regarded as a time consuming and demanding approach in terms of computational resources. Its parallelization is a complex problem since it usually affects the final outcome in comparison to what would be delivered by a sequential solution. This is due to the fact that subdividing an image to perform segmentation of its tiles concurrently usually introduces undesirable artifacts near to the borders of the image tiles. Additional processing steps are then required to properly stitch together the segments alongside tiles borders in order to eliminate such artifacts. In this work we evaluated alternative implementations of a previously proposed region-growing distributed segmentation approach, which was originally built on top of the Hadoop distributed computing framework. We developed a new implementation of the approach, which was built with the Spark framework, and compared its performance with that of the original implementation. In this investigation RS images of various sizes were processed using different configurations of a physical computer cluster. We evaluated computational performances and accessed the differences among the segmentation outcomes generated by the alternative implementations. We also assessed the stability of the implementations by comparing the segmentations produced with different cluster configurations. Although the approach is, in principle, suitable to any region growing algorithm, the experiments were performed with a particular segmentation method, and the results showed that the Spark implementation consistently outperformed the Hadoop counterpart, bringing in most cases a significant improvement in terms of processing time. The experiment results also attested the stability of the distributed segmentation approach, as very similar results were produced with the alternative implementations, running on different cluster configurations.

Deforestation is one of the main causes of biodiversity reduction, climate change among other destructive phenomena. Thus, early detection of deforestation processes is of paramount importance. Motivated by this scenario, this work presents an evaluation of methods for automatic deforestation detection, specifically Early Fusion (EF) Convolutional Network, Siamese Convolutional Network (S-CNN) and the well-known Support Vector Machine (SVM), taken as the baseline. These methods were evaluated in a region of the Brazilian Legal Amazon (BLA). Two Landsat 8 images acquired in 2016 and 2017 were used in our experiments. The impact of training set size was also investigated. The Deep Learning-based approaches clearly outperformed the SVM baseline in our approaches, both in terms of F1-score and Overall Accuracy, with a superiority of S-CNN over EF.

Advances in the development of Earth observation data acquisition systems have led to the continuously growing production of remote sensing datasets, for which timely analysis has become a major challenge. In this context, distributed computing technology can provide support for efficiently handling large amounts of data. Moreover, the use of distributed computing techniques, once restricted by the availability of physical computer clusters, is currently widespread due to the increasing offer of cloud computing infrastructure services. In this work, we introduce a cloud computing approach for object-based image analysis and classification of arbitrarily large remote sensing datasets. The approach is an original combination of different distributed methods which enables exploiting machine learning methods in the creation of classification models, through the use of a web-based notebook system. A prototype of the proposed approach was implemented with the methods available in the InterCloud system integrated with the Apache Zeppelin notebook system, for collaborative data analysis and visualization. In this implementation, the Apache Zeppelin system provided the means for using the scikit-learn Python machine learning library in the design of a classification model. In this work we also evaluated the approach with an object-based image land-cover classification of a GeoEye-1 scene, using resources from a commercial cloud computing infrastructure service provided. The obtained results showed the effectiveness of the approach in efficiently handling a large data volume in a scalable way, in terms of the number of allocated computing resources.

Optical imagery is often affected by the presence of clouds. Aiming to reduce their effects, different reconstruction techniques have been proposed in the last years. A common alternative is to extract data from active sensors, like Synthetic Aperture Radar (SAR), because they are almost independent on the atmospheric conditions and solar illumination. On the other hand, SAR images are more complex to interpret than optical images requiring particular handling. Recently, Conditional Generative Adversarial Networks (cGANs) have been widely used in different image generation tasks presenting state-of-the-art results. One application of cGANs is learning a nonlinear mapping function from two images of different domains. In this work, we combine the fact that SAR images are hardly affected by clouds with the ability of cGANS for image translation in order to map optical images from SAR ones so as to recover regions that are covered by clouds. Experimental results indicate that the proposed solution achieves better classification accuracy than SAR based classification.