Weed Detection in Perennial Ryegrass With Deep Learning Convolutional Neural Network

Jialin Yu; Arnold W Schumann; Zhe Cao; Shaun M Sharpe; Nathan S Boyd

doi:10.3389/fpls.2019.01422

Weed Detection in Perennial Ryegrass With Deep Learning Convolutional Neural Network

Front Plant Sci. 2019 Oct 31:10:1422. doi: 10.3389/fpls.2019.01422. eCollection 2019.

Authors

Jialin Yu¹, Arnold W Schumann², Zhe Cao³, Shaun M Sharpe⁴, Nathan S Boyd⁴

Affiliations

¹ Co-Innovation Center for Sustainable Forestry in Southern China, Nanjing Forestry University, Nanjing, China.
² Citrus Research and Education Center, University of Florida, Lake Alfred, FL, United States.
³ Crop Development Centre/Department of Plant Sciences, University of Saskatchewan, Saskatoon, SK, Canada.
⁴ Gulf Coast Research and Education Center, University of Florida, Wimauma, FL, United States.

Abstract

Precision herbicide application can substantially reduce herbicide input and weed control cost in turfgrass management systems. Intelligent spot-spraying system predominantly relies on machine vision-based detectors for autonomous weed control. In this work, several deep convolutional neural networks (DCNN) were constructed for detection of dandelion (Taraxacum officinale Web.), ground ivy (Glechoma hederacea L.), and spotted spurge (Euphorbia maculata L.) growing in perennial ryegrass. When the networks were trained using a dataset containing a total of 15,486 negative (images contained perennial ryegrass with no target weeds) and 17,600 positive images (images contained target weeds), VGGNet achieved high F₁ scores (≥0.9278), with high recall values (≥0.9952) for detection of E. maculata, G. hederacea, and T. officinale growing in perennial ryegrass. The F₁ scores of AlexNet ranged from 0.8437 to 0.9418 and were generally lower than VGGNet at detecting E. maculata, G. hederacea, and T. officinale. GoogleNet is not an effective DCNN at detecting these weed species mainly due to the low precision values. DetectNet is an effective DCNN and achieved high F₁ scores (≥0.9843) in the testing datasets for detection of T. officinale growing in perennial ryegrass. Moreover, VGGNet had the highest Matthews correlation coefficient (MCC) values, while GoogleNet had the lowest MCC values. Overall, the approach of training DCNN, particularly VGGNet and DetectNet, presents a clear path toward developing a machine vision-based decision system in smart sprayers for precision weed control in perennial ryegrass.

Keywords: artificial intelligence; machine learning; machine vision; precision herbicide application; weed control.