Loading...

Media is loading
 

Paper Type

Complete

Abstract

Traditional means of on-farm weed control mostly relies on manual labor. This process is time consuming, costly and contributes to major yield losses. The conventional application of chemical weed control, however, goes against the strive for sustainability. To solve this using computer vision, precision agriculture researchers have used remote sensing weed maps, but this has been largely ineffective for early season weed control due to problems such as solar reflectance and cloud cover in satellite imagery. With the current advances in artificial intelligence, this study leverages the automatic feature extraction capabilities of deep convolutional neural networks (DCNN) to classify plant seedlings. In a comparative study, we demonstrate that DCNNs can successfully classify crops and weeds in various phenological growth stages. Our results indicate that while training DCNNs from scratch can achieve state-of-the-art performance for weed classification tasks, model performance can be improved by fine-tuning a pre-trained model.

Share

COinS
 
Aug 10th, 12:00 AM

Towards Deep Learning for Weed Detection: Deep Convolutional Neural Network Architectures for Plant Seedling Classification

Traditional means of on-farm weed control mostly relies on manual labor. This process is time consuming, costly and contributes to major yield losses. The conventional application of chemical weed control, however, goes against the strive for sustainability. To solve this using computer vision, precision agriculture researchers have used remote sensing weed maps, but this has been largely ineffective for early season weed control due to problems such as solar reflectance and cloud cover in satellite imagery. With the current advances in artificial intelligence, this study leverages the automatic feature extraction capabilities of deep convolutional neural networks (DCNN) to classify plant seedlings. In a comparative study, we demonstrate that DCNNs can successfully classify crops and weeds in various phenological growth stages. Our results indicate that while training DCNNs from scratch can achieve state-of-the-art performance for weed classification tasks, model performance can be improved by fine-tuning a pre-trained model.

When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.