Proyectos vigentes

Development and validation of new technologies of remote sensing and machine learning applied to smart weed control

Equipo

Acrónimo

DECIMAL

IP

José Dorado (ICA), José Manuel Peña (ICA), Ana I. de Castro (INIA)

Resumen

This project aims at researching, developing and validating technological innovations to improve diverse management strategies of the agricultural sector in an environment of digital transformation. In a context of Precision Agriculture, the implementation of intelligent systems for identifying weed species and characterizing crop structure has proven to be critical to reduce the use of agrochemicals through selective and site-specific applications. Advances in artificial intelligence, together with new remote sensing platforms and sensors, make it possible to develop more efficient and accurate systems for data acquisition, processing and analysis, ultimately facilitating the design of weed management strategies adapted to the problem identified (type and severity), its spatial extension and temporal evolution, as well as its impact on crop yield. This project envisages the use and validation of unmanned aerial vehicle- (UAV-) based technologies and powerful analysis procedures (machine/deep learning) for the early and accurate identification of the main weed species affecting three of Spain’s main crops (maize, vine and tomato). The results will provide information of high technological and agronomic value to feed a decision support system (DSS) and to apply more timely, cost-efficient and sustainable crop protection strategies in various agroecosystems in Spain.

Actividades del proyecto

1) Identifying and mapping major weed species in maize and tomato by combining UAV images and machine/deep learning procedures; 2) Seasonal monitoring of invasive weeds (e.g., Amaranthus palmeri) in maize fields by both UAV and satellite remote sensing and evaluation of its spatio-temporal spread; 3) Design and validation of a DSS for SSWM in maize based on species-specific weed maps, crop vigour maps in early season and crop yield at harvest; 4) Assessing spatio-temporal dynamics of major weed species in vineyards under different weed management systems using UAV images and geo-referenced field data; 5) Detecting and evaluating weeds-vineyard competition problems by monitoring vines vigour in weed infested vineyards using UAV-based 3D crop models; and 6) High-throughput vineyard phenotyping by assessing vine development and phenology metrics using UAV technology.

Fecha inicio

09/01/2021

Fecha finalización

31/08/2024

Organismo financiador

AEI (PID2020-113229RB-C41)

Cuantía

180,290.00€