Among the systems that aim to help users to perform map reading tasks, augmented reality (AR) is one of the most promising. However, the impact of this new technology in terms of acceptance, motivation and improvement of the learning process has still not been sufficiently explored. This study aims to assess the contribution of an AR system for map reading and for improving users’ understanding of geographic data. Landscape and census data from New Zealand were used to evaluate the performance of an AR system in an experiment involving 60 participants. Differences in the participants’ backgrounds were reflected in the way they completed the task for both printed maps and the AR system and 90% of the participants preferred to work with the AR system rather than with traditional printed maps. Users that had previous experience of the geographical dataset provided performed the task better using only the printed maps, while those without that experience performed better with the AR system.
14 August 2024 in IEEE Geoscience and Remote Sensing Letters
Active fire segmentation in satellite imagery is a critical remote sensing task, providing essential support for planning, decision-making, and policy development. Several techniques have been proposed for this problem over the years, generally based on specific equations and thresholds, which are sometimes empirically chosen. Some satellites, such as MODIS and Landsat-8 have consolidated algorithms for this task. However, for other important satellites such as Sentinel-2, this is still an open problem. In this paper, we explore the possibility of using transfer learning to train convolutional and transformer-based deep architectures (U-Net, DeepLabV3+ and SegFormer) for active fire segmentation. We pre-train these architectures based on Landsat-8 images and automatically labeled samples, and fine-tune them to Sentinel-2 images. Experiments show that the proposed method achieves F 1-scores of up to 88.4% for Sentinel-2 images, outperforming three threshold-based algorithms by at least 19%, while maintaining a low demand for manually labeled samples. We also address detection over seam-line regions that present a particular challenge for existing methods. The source code and trained models are available on GitHub.
2 August 2024 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Active fire segmentation through satellite imagery is fundamental to prevention and damage prediction. Algorithms to address this problem usually rely on sensor-specific thresholds, empirically chosen based on a few image samples, and thus, are susceptible to many errors. Deep learning algorithms automatically extract information through various levels of abstraction, avoiding human intervention for feature engineering. However, training a deep network from scratch requires massive labeled data, and efforts in this direction have already been made for the Landsat-8 satellite. For Sentinel-2, another important satellite that has band wavelength similarities with Landsat-8, there is a limited number of such initiatives, with few studies even concerning hand-crafted (traditional) algorithms. In this context, we explored in this paper the transfer of knowledge for active fire segmentation from Landsat-8 to Sentinel-2, avoiding the need of a vast amount of labeled data from Sentinel-2, and also reducing the computational resources for training. We also compiled a benchmark containing 12,584 image patches extracted from 26 Sentinel-2 images from around the globe, along with manually annotated fire pixels, to assess the algorithm's response compared to a human specialist. In a series of transfer learning experiments by using the U-Net architecture, we showed that even a single labeled image from Sentinel-2 for training was sufficient to allow achieving an $F$ -score metric of 84.9% when Landsat-8 knowledge is transferred and further fine-tuned in the machine learning process, while the best hand-crafted algorithm designed for Sentinel-2 achieved an $F$ -score of 75.8% in the segmentation task. This project is available on GitHub.
2 January 2019 in Canadian Journal of Remote Sensing
The objective of this research was to distinguish and estimate cultivated areas of soybean and corn in Paraná State, Brazil, in the 2014/2015 crop season. The main obstacle in mapping summer crops using vegetation index images is to separate the cultivated areas with soybean and corn. These crops planted in a similar period present similar spectral signatures. Thus, with the use of Data Mining techniques (DM) and Artificial Neural Network (ANN) it was possible to carry out the crop mapping, even for those that present similarities in spectral-temporal profile of vegetation indexes. The accuracy obtained in the mappings resulted in a KI (Kappa Index) of 0.78 and 89% of OA (overall accuracy) indicating a high accuracy in the separation of soybean and corn summer crops.
Recent technological advancements in many areas have changed the way that individuals interact with the world. Some daily tasks require visualization skills, especially when in a map-reading context. Augmented Reality systems could provide substantial improvement to geovisualization once it enhances a real scene with virtual information. However, relatively little research has worked on assessing the effective contribution of such systems during map reading. So, this research aims to provide a first look into the usability of an Augmented Reality system prototype for interaction with geoinformation. For this purpose, we have designed an activity with volunteers in order to assess the system prototype usability. We have interviewed 14 users (three experts and 11 non-experts), where experts were subjects with the following characteristics: a professor; with a PhD degree in Cartography, GIS, Geography, or Environmental Sciences/Water Resources; and with experience treating spatial information related to water resources. The activity aimed to detect where the system really helps the user to interpret a hydrographic map and how the users were helped by the Augmented Reality system prototype. We may conclude that the Augmented Reality system was helpful to the users during the map reading, as well as allowing the construction of spatial knowledge within the proposed scenario.
28 August 2019 in Raega - O Espaço Geográfico em Análise