Naslov
Weed control using drones and robots

Autori
Grgić, Domagoj ; Ravlić, Marija

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
Book of Abstracts Digital Technologies in Agriculture / Lončarić, Zdenko ; Jović, Jurica - Osijek : Fakultet agrobiotehničkih znanosti Sveučilišta Josipa Jurja Strossmayera u Osijeku, 2022, 18-18

ISBN
978-953-8421-03-7

Skup
1st International Symposium on Digital Technologies in Agriculture & 1st Satellite Workshop – Digital Agriculture in Rural Area

Mjesto i datum
Osijek, Hrvatska, 06.12.2022. - 08.12.2022

Vrsta sudjelovanja
Poster

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
UAV, AI, robots, sustainability, site-specific weed management

Sažetak
Weeds are unwanted plants that compete with crops for space, water, light and nutrients and may cause reductions of yield and increase the cost of crop production. Modern agriculture relies on synthetic herbicides as a fast, simple and effective weed control measure. However, their excessive and improper use can cause a series of negative consequences, such as the occurrence of resistant weed populations, herbicide residues in the food chain, and environmental pollution with adverse effects on human and animal health. In order to minimize negative impact of herbicides and fit in strategies of reducing their use, adoption of precision agriculture (e.g., UAVs, remote sensing, robots, machine learning, deep learning, AI, computer vision etc.) in weed control are essential. Precision agriculture aims to establish systems that optimize resource input while maintaining high yields. Currently, uniform application of herbicides across the whole field, regardless of weed density and distribution, is typically conducted, although weeds are rarely evenly distributed and often grow in patches. The site-specific weed management (SSWM) implies application of herbicides or other direct weed control methods only to areas where weed density is above the economic weed threshold. The crucial step in SSWM is the collection of information on weed species composition, density and distribution in the field which enables application of weed control treatments at the right time, intensity and locations. UAVs equipped with different types of sensors (RGB, multispectral and hyperspectral cameras) are used for fast and accurate image acquisition which are afterwards analyzed to generate weed distribution maps. Mapping weeds at early growth stages enables for appropriate herbicide and rates selection, while pre-harvest weed mapping allows control of perennial weeds in following years or detection of herbicide resistant weeds. Weed maps are subsequently used for precise herbicide application to areas with high weed density (patch spraying) or individual plants (spot spraying). Ground sensing on the other hand enables real-time weed detection, recognition and precise spot spraying. Prototypes of weeding robots that automatically detect and identify weeds and apply small or micro-doses of herbicides are being developed. Beside selective herbicide application, robotic weed control systems that use mechanical weed removal (camera-guided inter- and intra-row weeding), flaming, laser and hot water are utilized. Their application is of great value in case of limited use of herbicides (e.g., organic agriculture), ban on active ingredients and lack of registered and effective plant protection products for specialty crops. Successful SSWM was reported for various crops, from cereals, row crops, vegetables, to permanent stands, with herbicide savings up to 90% without yield reductions compared to conventional application. Integrated Weed Management (IWM) with new tools provides environmental and economic benefits, biodiversity preservation and sustainable agriculture, although precise tools integration and implementation still poses a great challenge in commercial farming systems.

Izvorni jezik
Engleski

Znanstvena područja
Poljoprivreda (agronomija)

POVEZANOST RADA