October 5, 2020

How Earth observation can use in the field of Agriculture

The September Webinar organised by the S·3 Partnership on Traceability and Big Data was dedicated to the topic of Earth observation through Technology, and how Artificial Intelligence (AI) can contribute to this purpose. The two speakers were Joao Ribao and Bruno Ferreira, represented ISQ, a Portuguese Organisation that provides Scientific-Technological Support to different entities, and member of the Traceability and Big Data Partnership.

The session started with the definition of Earth observation as a remote sensing technique to collect information on the planet using different technologies like satellites, drones, aircrafts, sensors, etc. Since the launch of Explorer 7 in 1959, considered as the first step of Earth Observation (EO), a lot progresses have been made. Today EO includes Machine learning methods using Artificial Intelligence.

The speakers continued by explaining the differences among the different Earth observation satellite resolutions (spectral, spatial and temporal), and then illustrated their use in the field of Agriculture. The Copernicus Sentinel missions, for example, are being very helpful in monitoring the evolution of crops, water availability and quality, water stress prediction, vegetal species identification or weather and disaster predictions, among others.

Regarding Artificial Intelligence (AI), the first steps were developed in the 1940`s, in parallel with the invention of computers. Today, the combination between Earth Observation (EO) and AI has the potential, among other agricultural functionalities, of monitoring soil conditions or crop diseases.

Due to the broad range of uses and applications, AI has been divided in several subdomains, one of which is Machine Learning (ML). ML aims to provide to machines the ability to learn from data without being explicitly programmed for that purpose. It originates as well other subdomains like Deep Learning (DL), which is based on integrated algorithms inspired by the architecture of the biological neural networks in the human brain. In these moments, Deep Learning is gaining increasing attention from the research and the academia, and actually it has replaced the concept of Artificial Intelligence in the first positions of publications since 2009.

As a general conclusion, the speakers highlighted the main benefits of incorporating AI in the analysis of data, and in the search of hidden patterns in large datasets contributing to a smart monitoring of the agricultural sector.

More information