Ant nests in planted forests can cause significant damage. Ants can disturb the soil, affecting tree growth and nutrient absorption. These interactions can result in damage to tree health and reduced productivity of forest plantations. Therefore, ant nest control is often necessary to protect growth and wood quality.

In response to this demand, Radaz trained its artificial intelligence to locate and measure the size of anthills. This initiative became viable thanks to the use of P Band, which penetrates the ground to depths of up to 60 meters (tested to date), and helical flight, which provides higher resolution.

Artificial Intelligence Training

Comparison of optical image (Google) with radar image:

Optical image

SAR image (Synthetic Aperture Radar)

The points marked in red were used as training examples to allow the AI to learn the characteristics of an anthill, while the blue point (B) represented the target that should be identified.

Mapped anthills: data for AI algorithm

The result of the training was a margin of error of just 12 meters. Currently reaching 8 meters.

Result of the AI algorithm

Radaz's technology for detecting anthills was validated by Klabin

Press release - click here.

Post LinkedIn - click here.Post LinkedIn - click here.

Detailed information about this application is described in the article:

G. Oré, A. Santos, D. Ukan, R. Zanetti, M. Camargo, L. P. Oliveira, H. E. Hernandez-Figueroa, "Ant Nests Detection in Industrial Forests by SAR P-band Tomography," in IGARSS, Kuala Lumpur, Malaysia , Jul. 17–22, 2022.