Advanced search
Start date
Betweenand

Multi and hyperspectral imaging applied in discrimination and estimation of the arthropod pests infestation level on soybean

Grant number: 19/26099-0
Support type:Scholarships in Brazil - Post-Doctorate
Effective date (Start): June 01, 2020
Effective date (End): May 31, 2021
Field of knowledge:Agronomical Sciences - Agronomy - Plant Health
Principal Investigator:Pedro Takao Yamamoto
Grantee:Juliano de Bastos Pazini
Home Institution: Escola Superior de Agricultura Luiz de Queiroz (ESALQ). Universidade de São Paulo (USP). Piracicaba , SP, Brazil

Abstract

The Brazilian soybean production system has faced systematic and high economic losses due to the attack of pests. An effective Integrated Pest Management (IPM) program relies heavily on early detection of pest outbreaks, allowing farmers to make decisions before the infestations are well established, in order to restrict the increase in yield losses. However, traditional sampling methods make the process highly costly and challenging for large-scale agriculture, discouraging farmer adoption. Because of this, often occurs more frequent use of pesticides and timed applications. Biotic stress, such as pest herbivory, elicits responses from plant physiological defenses that lead to changes in leaf reflectance. Advanced imaging techniques can be used to detect such changes in the reflectance of soybeans and thus be used for pest monitoring. In this sense, it is intended to i) spectrally characterize healthy and pest-infested/damaged soybean plants by multi and hyperspectral imaging and; ii) to develop a system for detection and quantification of pest and related symptoms based on robust and least complex model as possible. In general, aerial multispectral sensors and terrestrial hyperspectral sensors will be used to identify reflectance profiles of arthropod-infested or damaged soybean plants. From this, univariate and multivariate statistical analyses will be employed to evaluate spectral behaviors at different levels of infestation or injury detected. This study aims to serve as a basis for the feasibility of a non-invasive monitoring technique that can combine practicality, speed, and efficiency to detect emerging outbreaks of soybean pests, aiming at full adoption of IPM in large-scale agriculture. (AU)