Regresar

Use of neural networks to the detection of pyroclastic flows in Ecuador volcanoes

Abstract:

Neural networks are oriented to provide solutions that require a complex computational solution and this answer must be precise, from they belong to a group of techniques grouped under deep learning, a branch that has proven to be effective in many tasks as image recognition and voice recordings in the last years. For this reason, a neural network was created for volcanic images recognition that allow presence of volcanic fluids detection, as well as their trajectory, to generate timely alerts when a volcanic phenomenon is present. For the algorithm implementation will be used open source computer vision libraries such as Open Computer Vision (OpenCV).