MEX02 - Smart Segmentation of the Swell as an Energy Resource in Mexican Coasts

This project will seek, based on a series of satellite images and the implementation of techniques related to the field of artificial intelligence, to carry out an evaluation of the energy potential by waves for the geographical region corresponding to the coasts of Ensenada Baja California. The proposed computational code will have the fundamental task of recognizing the surface patterns of the waves and establishing a relationship with respect to variables that are relative to the estimation of the energy potential, for example, the significant height of the waves, the average period and the energy density. The general perspective of resolution will start from a pre-treatment of the images where issues such as scaling or dimensional reduction will be addressed. Regarding the computational model, one of the value proposals for the challenge will be the search for an integration of complementary information through variables such as wind speed or temperature; this once the corresponding correlation analysis is carried out. The accuracy of the model will be determined statistically by comparing the predictions made against the actual data measured at the site. Convolutional neural networks are probably recognized as the most suitable model to address the problem, however, they present a high computational cost. In this context, its implementation is not ruled out, however, the exploration of new models and algorithms that allow facing the present challenge is encouraged. Finally, presenting the data in the form of time series will allow the identification of certain trends and temporal patterns in favor of a better understanding of the disposition of the energy resource, and even the realization of forecasts in the short term if a recurrent neural network is attached.

A computational code that, making use of artificial intelligence, has the ability to receive satellite images of the water surface, and from them, determine the energy use potential of the corresponding site.

In addition, the data obtained may be re-entered into an auxiliary model focused on the analysis of the time series, this in order to identify time patterns that allow a more appropriate evaluation of the energy resource.

Python programmers
Satellite imagery specialists
Data science specialists
Specialists in oceanography and meteorology
Climate change specialists.