Oil palm tree detection and counting for precision farming using deep learning CNN

• Main Authors: Kipli, Kuryati, Lee, Paul Jaw Bin, Sam, Huai En, Joseph, Annie, Zen, Hushairi, Gan, Brandon Yong Kien, A. Jalil, M., Ray, Kanad, Kaiser, M. Shamim, Mahmud, Mufti
• Format: Book Section
• Published: Springer Science and Business Media Deutschland GmbH 2022
• Subjects: Q Science (General)
• Online Access: http://eprints.utm.my/id/eprint/101093/; http://dx.doi.org/10.1007/978-981-16-7597-3_45

Oil palm tree is a very important crop in Malaysia and other tropical areas. The number of oil palm trees in a plantation area is crucial as it could help to estimate the potential yield of palm oil, monitoring the growing situation of palm trees after plantation such as the age and the survival rate and also the amount of fertilizer and pesticides needed. In this paper, a deep learning-based oil palm tree detection and counting method is proposed and designed into a functioning app. Images of oil palm plantation are collected by using drones then they are pre-processed. The pre-processed images are used to train and optimize the convolutional neural network (CNN). After the CNN model is trained, it is used to predict the label for all the samples in an image dataset collected through the sliding window technique. Its performance is tested. The performance of the classifier is tested on three different tree conditions, from small number of properly separated trees to big number of crowded trees. Based on the result, accuracy ranging from 83.5% to 100% is obtained. Finally, the method is built into an application for a better user experience.