×

People

Roberto Arias

Roberto Arias

Cohort III, NERTO, Masters

M.S, Digital Signal Processing, Computer Engineering, Graduate

Cohort Level: Cohort - III

Career Goal: To pursue a PhD either in Computer Engineering Machine Learning or Quantum Physics. Or to pursue a career in Remote Sensing at NOAA or the private sector.

Expected Graduation Date: December 20, 2021

Degree: M.S Digital Signal Processing, Computer Engineering

Research Title: Multisource Satellite Data Fusion for monitoring sediments in coastal waters using Deep-Learning.

Research Synopsis: Advances in remote sensing technologies have resulted in generation of varied large-scale data that requires efficient processing algorithms for information extraction, fusion and decision making. Furthermore, having such efficient tools provide the ability of providing information of an event under circumstances of sensor failure by using previously learned behavior and relying on functioning sensors.

Multi-level sensor fusion has the advantages of increased robustness, confidence and reliability of sensor applications, reduced ambiguity and uncertainty based on redundant sensor measurements, extended spatial and temporal coverage, low-cost. We propose to study new sensor fusion mechanisms for data (pixel scale), feature, and classification tools using a large data analytics framework for remote sensing systems, focusing on the use of VIIRS, MODIS and LANDSAT data for the monitoring and identification of sedimentation in coastal environments.

A master student Mr. Roberto Arias from the program of Computer Engineering program at the Department of Electrical and Computer Engineering will conduct his thesis in this project. The project will involve the study of deep learning algorithms for more efficient data fusion models:

To study algorithms for fusion of multiresolution (spectral and spatial images) images

Propose a model to improve data fusion for remote sensing applications

Advances in remote sensing technologies have resulted in generation of varied large-scale data that requires efficient processing algorithms for information extraction, fusion and decision making. Furthermore, having such efficient tools provide the ability of providing information of an event under circumstances of sensor failure by using previously learned behavior and relying on functioning sensors.

Multi-level sensor fusion has the advantages of increased robustness, confidence and reliability of sensor applications, reduced ambiguity and uncertainty based on redundant sensor measurements, extended spatial and temporal coverage, low-cost. We propose to study new sensor fusion mechanisms for data (pixel scale), feature, and classification tools using a large data analytics framework for remote sensing systems, focusing on the use of VIIRS, MODIS and LANDSAT data for the monitoring and identification of sedimentation in coastal environments.

The project will involve the study of deep learning algorithms for more efficient data fusion models:

To study algorithms for fusion of multiresolution (spectral and spatial images) images

Propose a model to improve data fusion for remote sensing applications.

CESSRST Consortium

CESSRST is led by The City University of New York and brings together Hampton University, VA; University of Puerto Rico at Mayaguez, PR; San Diego State University, CA; University of Maryland Baltimore County, MD; University of Texas at El Paso, TX.