Repository logo

Browsing by Issue Date, starting with "2022-04-22"

Filter results by year or month
Now showing 1 - 2 of 2
  • Assessing the impact of the loss function and encoder architecture for fire aerial images segmentation using deeplabv3+
    Publication . Harkat, Houda; Nascimento, Jose; Bernardino, Alexandre; Ahmed, Hasmath Farhana Thariq
    Wildfire early detection and prevention had become a priority. Detection using Internet of Things (IoT) sensors, however, is expensive in practical situations. The majority of present wildfire detection research focuses on segmentation and detection. The developed machine learning models deploy appropriate image processing techniques to enhance the detection outputs. As a result, the time necessary for data processing is drastically reduced, as the time required rises exponentially with the size of the captured pictures. In a real-time fire emergency, it is critical to notice the fire pixels and warn the firemen as soon as possible to handle the problem more quickly. The present study addresses the challenge mentioned above by implementing an on-site detection system that detects fire pixels in real-time in the given scenario. The proposed approach is accomplished using Deeplabv3+, a deep learning architecture that is an enhanced version of an existing model. However, present work fine-tuned the Deeplabv3 model through various experimental trials that have resulted in improved performance. Two public aerial datasets, the Corsican dataset and FLAME, and one private dataset, Firefront Gestosa, were used for experimental trials in this work with different backbones. To conclude, the selected model trained with ResNet-50 and Dice loss attains a global accuracy of 98.70%, a mean accuracy of 89.54%, a mean IoU 86.38%, a weighted IoU of 97.51%, and a mean BF score of 93.86%.
    2022-04-22Journal article Open access Show more
  • Imaging the crust and uppermost mantle structure of Portugal (West Iberia) with seismic ambient noise
    Publication . Silveira, Graça; Dias, Nuno; Kiselev, Sergey; Stutzmann, Eleonore; Custodio, Susana; Schimmel, Martin
    We present a new high-resolution three-dimensional (3D) shear wave velocity (Vs) model of the crust and uppermost mantle beneath Portugal, inferred from ambient seismic noise tomography. We use broadband seismic data from a dense temporary deployment covering the entire Portuguese mainland between 2010 and 2012 in the scope of the WILAS project. Vertical component data are processed using phase correlation and phase weighted stack to obtain Empirical Green functions (EGF) for 3900 station pairs. Further, we use a random sampling and subset stacking strategy to measure robust Rayleigh wave group velocities in the period range 7-30 s and associated uncertainties. The tomographic inversion is performed in 2 steps: First, we determine group velocity lateral variations for each period. Next, we invert them at each grid point using a new trans-dimensional inversion scheme to obtain the 3D shear wave velocity model. The final 3D model extends from the upper crust (5 km) down to the uppermost mantle (60 km) and has a lateral resolution of similar to 50 km. In the upper and middle crust, the Vs anomaly pattern matches the tectonic units of the variscan massif and alpine basins. The transition between the Lusitanian Basin and the Ossa Morena Zone is marked by a contrast between moderate and high velocity anomalies, in addition to two arched earthquake lineations. Some faults, namely the Manteigas-Vilarica-Braganca fault and the Porto-Tomar-Ferreira do Alentejo fault, have a clear signature from the upper crust down to the uppermost mantle (60 km). Our 3D shear wave velocity model offers new insights into the continuation of the main tectonic units at depth and contributes to better understanding the seismicity of Portugal.
    2022-04-22Journal article Restricted access Show more