Abstract
Coconut palm tree plantation, monitoring and management in tropical countries is vital to improving exports, domestic food use and the economy. Remote sensing and GIS technologies are widely used to maintain and monitor the coconut palm’s health and production using very high-resolution satellite data. Deep learning algorithms will help identify and monitor the growth and health of coconut palms automatically using satellite data. In this study, a deep learning model has been developed to identify the coconut trees using Worldview-2 satellite data. ArcGIS Pro software is used to develop the deep learning model. Single Shot Detector (SSD) algorithm is adopted for developing the model to detect the coconut trees. The model is based on the object detection technique. The generated model accurately identified the individual trees for the study area using the Worldview-2 satellite data. Produced maps will be used to calculate the population of coconut trees.
