Drone shaped bird will arrive in Latin America to protect crops

A team from the Analog Microelectronics and Mixed Signal Group of the Institute of Microelectronics of Seville (IMSE), a mixed center of the Higher Council of Scientific Research (CSIC) and the University of Seville (US), have developed several devices for capturing images with intelligence embedded that have been licensed by Fobos Solutions.

This engineering company that manufactures and markets an Unmanned Aerial Vehicle developed by the research team of Jordi Figuerola of the Biological Station of Doñana of the CSIC, which emulates the behavior of raptors and has application in the protection of crops and fish farms.

Farmers with strawberry plantations in Huelva, apricot in the Sierra Norte de Sevilla or quinoa in Malaga have already tried this drone in the shape of a raptor, called to become the 'scarecrow of the future' by being able to mimic the behavior of these predators to scare flocks of smaller birds that each year involve thousands of crop losses.

Artificial intelligence

The Galician company now seeks to improve the services it offers by incorporating into the licensed avian mimetic drone systems capable of processing images and making decisions from the vehicle itself in real time, without the need to transmit them to a server or store them in a memory on the drone itself. developed from the IMSE.

"These systems are very light at the same time powerful. Our patents allow us to have a very specific hardware embedded in the sensor layer of the chip installed in the drone, which accelerates the processing necessary for decision making from the beginning, when the computational load is larger ", explains Ángel Rodríguez Vázquez, professor at the Department of Microelectronics and Electromagnetism of the US and one of the researchers responsible for the licensed patents.

This solves one of the main restrictions when it comes to improving or incorporating artificial vision systems in these vehicles, and it is recommended that the deployments installed in the drones be light, since the weight directly affects the design and autonomy of the vehicles. appliances.

Computational load

The inventions developed by the IMSE researchers also aim to accelerate the postprocessing of images with millions of pixels: "The biggest computational load when dealing with images occurs at the beginning, when you have to discard that huge amount of pixels and keep the information that interests you to perform a certain processing", adds Jorge Fernández, another of the researchers involved.

In your opinion, "Thanks to these innovations, it will also be possible to have information already processed when the drone lands, saving time and effort in image post-production processes, reducing the computational load that artificial intelligence tasks imply".

Source: UCC + i US

Previous article

next article


Representatives of the Chilean Blueberry Fruit Committee analyzed the...
Climate change is affecting the vital microbes of blueberries
Demand drives growing blueberry production in China