Weed Detection using Lightweight DL models with Transfer Learning & Hyperparameter Optimization
Outlet Title
AMCIS 2024 Proceedings
Document Type
Conference Proceeding
Publication Date
2024
Abstract
Weeds significantly reduce agricultural yields. This study focuses on using artificial intelligence, specifically deep learning, for early-stage weed detection on edge devices, which have limited computational power. We propose a three-part strategy to address this challenge: employing lightweight architectures to reduce the model size and computational demand; using transfer learning to overcome the limitations of small data sets; and applying Bayesian optimization to fine-tune model parameters. Our results show that MobileNetV2 and EfficientNetV2B0 models achieve high accuracy (95.09% and 95.79% respectively) with MobileNetV2 being nearly as accurate but much smaller in size (13.96 MB compared to 23 MB). This demonstrates MobileNetV2's suitability for computationally constrained edge devices.
Recommended Citation
Shaheen, Ali and El-Gayar, Omar F., "Weed Detection using Lightweight DL models with Transfer Learning & Hyperparameter Optimization" (2024). Research & Publications. 402.
https://scholar.dsu.edu/bispapers/402