Research
Statistical and physically based hyperspectral and multispectral reflectance modelling for agricultural monitoring: a case study in Vilankulo, Mozambique
Research Overview
Agriculture faces challenges related to infestation by weeds, diseases and pests; soil treatment; drainage and irrigation. As a result, we are witnessing high crop losses, but also environmental disasters due to the excessive use of agrochemicals and water. Artificial Intelligence (AI), due to its ability to learn rigorously, has become a crucial technique for solving the problems of agriculture (Bannerjee, Sarkar, Das, & Ghosh, 2018). The application of AI in agriculture dates back to 1983 (Baker, Lambert, & McKinion, 1983), however, its use in Mozambique is incipient. In view of the implementation of strategic initiatives for the structural transformation of agriculture in Mozambique, there has been an increase in cultivated areas and in the level of use of agricultural inputs. For example, in the 2020/2021 agricultural campaign, the National Program for the Integration of Family Agriculture into Value Chains, SUSTENTA, made available around 300,000 technological packages to small farmers, supporting them in increasing the areas of cultivation, the level of use of agricultural inputs and, consequently, to increase their production in different cultures (MADER, 2021). If these results can be considered the beginning of the desired structural transformation in the agricultural sector, there are concerns regarding the sustainability of this intensification, above all in terms of efficient use of production inputs. In this sense, it is urgent to develop innovative tools conducive to the establishment of robust, economically competitive, socially responsible and environmentally efficient production systems. Said tools should be able to monitor the cultivation areas and provide timely information that gives the farmer greater decision-making capacity; localized application and, in the appropriate amount and time of inputs and/or interventions necessary to sustain high productivity, thus contributing to the preservation of the environment, namely carbon sequestration. In this context, this research aims to develop AI algorithms for optimizing the control of weeds, diseases, pests and crop irrigation in Mozambique.