Detection of grassy weeds in bermudagrass with deep convolutional neural networks

Citation

Yu, J., Schumann, A.W., Sharpe, S.M., Li, X., Boyd, N.S. (2020). Detection of grassy weeds in bermudagrass with deep convolutional neural networks. Weed Science, [online] 68(5), 545-552. http://dx.doi.org/10.1017/wsc.2020.46

Plain language summary

Due to the patchiness of weeds, precision spraying has the potential to significantly reduce herbicide inputs in various settings. Turfgrass is subjected to many herbicides so the opportunities to reduce inputs are many, depending on the herbicide and weed combinations. Image classification networks were trained and one (VGGNet) with near perfect accuracy to detect 4 different grass weeds. Future direction will be testing derived networks in the field when linked to developed precision spraying prototypes.

Abstract

Spot spraying POST herbicides is an effective approach to reduce herbicide input and weed control cost. Machine vision detection of grass or grass-like weeds in turfgrass systems is a challenging task due to the similarity in plant morphology. In this work, we explored the feasibility of using image classification with deep convolutional neural networks (DCNN), including AlexNet, GoogLeNet, and VGGNet, for detection of crabgrass species (Digitaria spp.), doveweed [Murdannia nudiflora (L.) Brenan], dallisgrass (Paspalum dilatatum Poir.), and tropical signalgrass [Urochloa distachya (L.) T.Q. Nguyen] in bermudagrass [Cynodon dactylon (L.) Pers.]. VGGNet generally outperformed AlexNet and GoogLeNet in detecting selected grassy weeds. For detection of P. dilatatum, VGGNet achieved high F1 scores (≥0.97) and recall values (≥0.99). A single VGGNet model exhibited high F1 scores (≥0.93) and recall values (1.00) that reliably detected Digitaria spp., M. nudiflora, P. dilatatum, and U. distachya. Low weed density reduced the recall values of AlexNet at detecting all weed species and GoogLeNet at detecting Digitaria spp. In comparison, VGGNet achieved excellent performances (overall accuracy = 1.00) at detecting all weed species in both high and low weed-density scenarios. These results demonstrate the feasibility of using DCNN for detection of grass or grass-like weeds in turfgrass systems.

Publication date

2020-09-01

Author profiles