Abstract
The automatic segmentation of ears in wheat canopy images is an important step to measure ear density or extract relevant plant traits separately for the different organs. Recent deep learning algorithms appear as promising tools to accurately detect ears in a wide diversity of conditions. However, they remain complicated to implement and necessitate a huge training database. This paper is aimed at proposing an easy and quick to train and robust alternative to segment wheat ears from heading to maturity growth stage. The tested method was based on superpixel classification exploiting features from RGB and multispectral cameras. Three classifiers were trained with wheat images acquired from heading to maturity on two cultivars at different levels of fertilizer. The best classifier, the support vector machine (SVM), yielded satisfactory segmentation and reached 94% accuracy. However, the segmentation at the pixel level could not be assessed only by the superpixel classification accuracy. For this reason, a second assessment method was proposed to consider the entire process. A simple graphical tool was developed to annotate pixels. The strategy was to annotate a few pixels per image to be able to quickly annotate the entire image set, and thus account for very diverse conditions. Results showed a lesser segmentation score (F1-score) for the heading and flowering stages and for the zero nitrogen input object. The methodology appeared appropriate for further work on the growth dynamics of the different wheat organs and in the frame of other segmentation challenges.
Highlights
Grain yield is the most valuable trait for the wheat breeders, as it clearly translates the economic value of the selection process
An automatic segmentation of wheat ears based on superpixel classification was proposed
Features from RGB and multispectral cameras were fed to three classifiers
Summary
Grain yield is the most valuable trait for the wheat breeders, as it clearly translates the economic value of the selection process. Breeders are interested in improving the selection pipelines with other crop performance criteria such as the radiation use efficiency [1]. Recent advances in sensing technologies and the increase of computing power nowadays allow to extract lots of information from a crop canopy in a nondestructive way and throughout all the season [2]. [3] used five sensing modules in a soybean plant breeding program. Multisensor systems allow to use different sources of data to enrich each data point. They reported strong correlation among sensor-based plant traits providing a powerful tool for phenotype characterization
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have