Monitoring growth rate of lettuce using deep convolutional neural networks
Journal
2019 ASABE Annual International Meeting
Date Issued
2019-01-01
Author(s)
Abstract
© 2019 ASABE Annual International Meeting. All rights reserved. Lettuce is one majorly consumed vegetable worldwide. According to the United States Department of Agriculture, lettuce accounted for 7% of total vegetable consumption in 2013. Monitoring the growth of lettuces is crucial for ensuring the quality and quantity of lettuce. This study proposed to monitor the growth of lettuces in greenhouses using time-lapse images and deep learning. An imaging system was constructed to acquire time-lapse images of lettuces in greenhouses. A mask region-based convolutional neural network (Mask R-CNN) model was next developed to localize the lettuces in the images and segment the leaf areas simultaneously. The growth rates were then determined as the leaf areas of the lettuces versus time. Experimental results showed that the Mask R-CNN model achieved an accuracy reached 97.63% on estimating leaf area.
Subjects
Deep learning | Growth monitoring | Mask region-based convolutional neural network
SDGs
Type
conference paper