Aplikasi Mobile untuk Rekonstruksi dan Identifikasi Area Kerusakan Daun Menggunakan Convex Hull

Abstract

This research proposes a method to calculate the area of damaged leaf attacked by pests, implemented as a mobile application on the Android operating system. The mobile application works on mobile device (client) for image acquisition and preprocess, while the calculation are performed on the server. This research uses Otsu method for image segmentation process and Convex Hull to reconstruct the area of damaged leaf. Reconstruction is done because the leaves were damaged so that their original shape was unrecognizable. After the reconstruction, the original shape of the leaves can be identified so the level of leaves damage can be calculated. The average error rate on the accuracy of the reconstruction of damaged salam (Syzygium polyanthum) leaves, kemuning (Murraya paniculata) leaves and sirih (Piper betle L) leaves is 19.19%, 16.81% and 22.31% respectively. Evaluation results show that the level of damage that occurs in the leaves and the reconstruction accuracy rate are inversely related. Convex Hull algorithm can reconstruct damaged leaves that have pinnate bone, but it is not suitable for leaves that have finger bone.