Accuracy assessment of automation oil palm tree counting using ecognition based of UAV imagery / Fairuz Maizan Sazeli

A good management of oil palm plantation become a growing concern as the demand
on oil palm is increasing. The farmers need a high technological system that can make
it easier to identify the number of oil palm in the plantation for the fertilizers and
prediction of yield. There is already manual technique and automation technique in
determining the number of oil palm trees exist but the manual technique is time
consuming and more expensive. The existence of automation method has not been
widely known by the farmers and the accuracy is not yet been verified. Therefore, the
accuracy of automation technique should be verified as it can be beneficial to the
agriculture sector. This research aim is to identify the accuracy of automation method
in oil palm tree counting by using the result from manual digitizing as a verification
subject to the automation technique. The data used in this research which is UAV
imagery data has been obtained from Braintree company. This data is suitable to be
used in this project as it provides high spatial resolution where the trees can be identified
easily. The result analyzes whether the accuracy of automation method by manipulating
five different threshold values approximately same as the accuracy of manual method.
It is suggested that the result of one of the threshold value in the automation method
has a high accuracy and can be used for oil palm tree counting.