GENDUS
Challenge
The researchers realized that the aquaculture industry was not automated and that many steps still required manual labor to perform. Therefore, only salmon and trout aquaculture companies could afford the costs of manual methods of fish genus classification, due to the high economic value of these species.
Solution
Development of an automatic fish genus classification system based on image processing using Artificial Intelligence (AI).
Work completed
Development and construction of the prototype system based on convolutional neural networks and image processing algorithms.
OV-CARE
Challenge
Company members were able to see the growing global trend of problems related to the spread of pests and diseases in vineyards and olive trees, as well as the problems related to the spread of crop-related diseases and the impact this problem could have in the near future.
Solution
Development of a tool for the global control of pests and diseases in vineyards and olive trees based on spectroscopy, neural networks, image processing and the use of unmanned aerial vehicles.
Work completed
Development of the software and follow-up of the project up to the testing phase, obtaining promising results with high accuracy rate.
CESSAR
Challenge
The researchers realized the need to monitor the specific strength of concrete in civil engineering works in an automatic and simple way, reducing construction times and bringing greater safety to designs.
Solution
Development of a tool that allows finding the maximum power point of photovoltaic panels in real time through the combined use of Artificial Intelligence (AI), Machine Learning algorithms and different sensors with the main objective of improving their efficiency.
Work completed
Development of the algorithms and implementation in software that can be easily incorporated into photovoltaic systems.
MASTEEC
Challenge
The researchers realized that the non-linearity characteristic of the I-V (current-voltage) curve in PV panels allowed the existence of a single value of Voltage (V.opt) for which the electrical power supplied by the PV panel was maximum (P.max). This task was performed by the so-called maximum power point tracking (MPPT) systems. However, MPPTs had performance limitations that restricted the overall efficiency of the solar plant, as the I-V curve was constantly changing under variations of insolation and temperature, as well as under partial shading conditions.
Solution
Development of a tool that combines the use of sensors, artificial intelligence (AI) and cloud computing technology to accurately determine the specific strength of concrete in real time.
Work completed
Development of sensors and hardware system with real-time monitoring capacity and communication with the cloud that stores the information.