An investigation into the detection of sugarcane African stalk borer (Eldana saccharina Walker (Lepidoptera : Pyralidae)) using hyperspectral data (spectroradiometry).
Loading...
Date
2009
Authors
Mokhele, Tholang Alfred.
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The South African Sugarcane production is one of the world’s leading sugarcane (Saccharum spp. Hybrid) producers. However, in recent years Eldana saccharina Walker has been the most destructive pest in South African sugarcane production, causing great crop loses per annum and is the most important factor limiting sugarcane productivity. The pest has been monitored using a traditional visual approach whereby a representative sample of stalks is taken from a field and split longitudinally to assess damage and count the number of E. saccharina larvae and pupae. However, this approach is time-consuming, labour intensive and sometimes biased as only easily accessible areas are often surveyed. In order to investigate a more economical but equally effective survey methodology, this study aimed to determine the potential of using hyperspectral remote sensing (spectroradiometry) for identifying sugarcane attacked by E. saccharina. A hand-held spectroradiometer ASD Field Spec® 3 was used to collect leaf spectral measurements of sugarcane plants from a potted-plant trial taking place under shade house conditions at the South African Sugarcane Research Institute (SASRI). In this trial, nitrogen (N) and silicon (Si) fertilizers were applied at known levels to sugarcane varieties. Varieties were either resistant or intermediate resistant or susceptible to E. saccharina attack. In addition, watering regimes and artificial infestation of E. saccharina were carefully controlled. Results illustrated that severe E. saccharina infestation increased spectral reflectance throughout the whole spectrum range (400 – 2500 nm) and caused a red-edge shift to the shorter wavelength. Eldana saccharina stalk damage was also linearly related to modified normalized difference vegetation index (mNDVI) using R2025 and R2200 (R2 = 0.69). It was concluded that hyperspectral data has a potential for use in monitoring E. saccharina in sugarcane rapidly and non-destructively under controlled conditions. A followup study is recommended in field conditions and using airborne and/or spaceborne hyperspectral sensors.
Description
Thesis (M.Sc.)-University of KwaZulu-Natal, Durban, 2009.
Keywords
Sugarcane borer., Stem borers., Theses--Environmental management.