Unmanned Aircraft System (UAS) Based Large Area Identification of Basal Stem Rot (BSR) Infected Oil Palm Tree Using Convolutional Neural Network (CNN)

Abstract: Malaysia is the second largest palm oil producer in the world but ironically, the overall production of the palm oil is declining versus planned growth trajectory. One of the factors affecting the exponential growth of the palm oil estate is fungal or bacterial infection. One of the major diseases that infected palm oil tree is Ganoderma Boninense that which known as Basal Stem Rot (BSR). A fast, accurate, wide-coverage and non-destructive method for detection of infected oil palm tree is required. Therefore, low-cost unmanned aircraft for detection of the health status of palm tree plantation was implemented and image processing of the data collected from the low-cost unmanned aircraft system was processed for oil palm tree health status. MATLAB based detection was implemented. Convolutional Neural Network (CNN) and few image processing techniques were introduced in this project.

Keywords: Basal Stem Rot (BSR), Unmanned Aircraft System (UAS), Convolutional Neural Network (CNN), detection