Free and Unfree Weed Classification in Young Palm Oil Crops Using Artificial Neural Network

• Published in: Lecture Notes in Networks and Systems
• Main Author: Jopony S.T.M.; Ahmad F.; Osman M.K.; Idris M.; Yahaya S.Z.; Daud K.; Ismail A.P.; Ibrahim A.H.; Soh Z.H.C.
• Format: Conference paper
• Language: English
• Published: Springer Science and Business Media Deutschland GmbH 2023
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85174485337&doi=10.1007%2f978-3-031-43520-1_2&partnerID=40&md5=6a6f4a1bbb09b45488ac242a8b2d026b

An automatic classification of whether a circle area of a young palm oil crop is free or unfree from unwanted weed can be a crucial step towards improving the growth management of a palm oil plantation. However, there is still less study that utilizes artificial intelligence techniques to address this issue. Most previous studies focused on leaves diseases, ripeness of the fruit bunch of the palm oil and counting of the palm oil crop instead of the groundcover management of the palm circle. Hence, this study proposes the development of an automatic and intelligent technique to classify the condition of a young palm oil crop which is based on the condition of ground cover management. Images of the different young palm oil conditions where the palm circles must be visible in the image acted as the datasets for this system. Local Binary Pattern is implemented as the feature extraction method and the classification result is compared with and without feature selection technique. ReliefF was chosen as the feature selection technique for this system. The developed ANN model with the implementation of feature selection technique produced the highest accuracy of 92.9% of correct classification. Consequently, this system is suitable for the palm oil industries to monitor the health condition of the palm oil plant. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.