Performance Evaluation of Deep Learning based approach for Paddy Head Detection in Images
Abstract
Prediction of crop yield using a computer vision-based approach is one of the active research areas in precision agriculture. Paddy head detection is a key role in yield prediction. The performance analysis of deep learning-based models such as Faster RCNN with EfficientNet, ResNet, and YOLO to detect paddy heads available in images has been investigated in this research since rice is one of the staple food of Asian countries. The performance of these methods is evaluated on the paddy dataset collected by the authors from the paddy lands in Sri Lanka since the lack of a publically available paddy dataset. Out of these deep learning-based approaches, YOLO V5 was able to achieve 88.10% accuracy.
Building: SickKids Hospital / University of Toronto
Room: Science Hall
Date: July 2, 2023 - 02:05 PM – 02:20 PM