Predictive zoning of pest and disease infestations in rice field based on UAV aerial imagery

Abstract

Bacterial leaf blight (BLB), bacterial panicle blight (BPB), and stem borer (SB) are serious infestations to the rice crop. Detection is the first essential step for effective management. The objective of the study is to provide a fast and accurate tool in detecting the infestation damages through Unmanned Aerial Vehicle (UAV) aerial imagery. System of Rice Intensification (SRI) was implemented and a UAV equipped with a digital multispectral camera was used to capture image of 20 rice plots that were treated with two types of fertilizers (organic and inorganic) in two different treatment rates namely; uniform rate and variable rate. Ground truths of infestation were observed and collected. Geospatial interpolation (kriging), linear regression analysis, and Soil Plant Analysis Development (SPAD) value models were carried out to predict the zones and level of infestation damages in the rice field. Maps showing areas with high, medium, and low counts of infestation damages were prepared using spatial analysis. The results of the relationship indicate that there were a strong correlation and high R2 between SPAD values obtained through the UAV method and infestation counts during the growth stages of 60 Days After Transplanting (DAT), 80 DAT, and 100 DAT. The findings show that the high severity of infestation happened in the plot that used a high amount of fertilizer compared to the plot that supplied with variable rate fertilizer. Infestation maps produced from the UAV aerial image would be an effective tool in detecting the pest and disease in the rice field.