Work

An efficientnet-based model for classification of oil palm, coconut and banana trees in drone images

Year: 2024

Type: article

Abstract: Oil palm tree detection and classification from coconut and banana trees is vital for increasing productions of oil palm tree business in the globe particularly in Malaysia. Since oil palm, coconut and banana trees share common characteristics such as shape and structure of trees, classification is challenging. Further, this work considers images captured by drones, which makes the classification problem more complex. Unlike most existing methods aim to detect oil palm trees, the proposed work aims to detect trees and classify them. Inspired by the success of the Segment Anything Model (SAM), which is generalized model work for segmenting objects and persons etc, we adapt the same SAM for detecting and localizing oil palm, coconut and banana trees in the drone images. Similarly, motivated by the efficiency and effective feature extraction of the EfficientNetB3, we adapt the EfficientNetB3 for classification of oil palm, coconut and banana trees. Therefore, the proposed work fuses the above two models as end-to-end in novel for classification. For experimentation, we collected dataset from a Malaysian drone services company, which provides frames captured across diverse locations. The results of the proposed model and state-of-the art models on our dataset show the proposed model is the best.

Source: Smart Agricultural Technology

DOI: 10.1016/j.atech.2024.100748

Authors: Kwee Kim Teo, Nurul Fazmidar Binti Mohd Noor, Shivakumara Palaiahnakote, Mohamad Nizam Bin Ayub

Institutions: University of Malaya, University of Salford

Language: English

Topic: Advanced Neural Network Applications

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

SDG: Zero hunger

Open Access status: gold

Estimated APC paid: $1,800