Detecting the severity of maize streak virus infestations in maize crop using in situ hyperspectral data

Abstract

Maize streak geminivirus (MSV) causes maize streak disease, a major disease limiting maize production over widespread areas of Africa. There has always been an urgency to develop quick and efficient methods of detecting such a disease for control purposes as well as increased food production and security. The use of remote sensing techniques for detecting the MSV infected maize was evaluated in this study based on experiments in Ofcolaco, Tzaneen, South Africa. Specifically, the potential of hyperspectral data for detecting different levels of MSV infection in maize was tested based on the guided regularized random forest (GRRF) algorithm. The findings of this study illustrate the strength of hyperspectral data in detecting different levels of MSV infection. Specifically, the optimal bands for detecting different levels of maize streak disease in maize were 552 nm, 603 nm, 683 nm, 881 nm and 2338 nm based on the GRRF algorithm. This study underscores the potential of remotely sensed data in the accurate detection of food crop diseases such as MSV and their severity, which is critical in crop monitoring to foster food security especially in the resource-limited sub-Saharan Africa.

Keywords:

• field spectroscopy
• random forest
• remote sensing
• maize streak virus (MSV)