AUTHOR=Narisetti Narendra , Neumann Kerstin , Röder Marion S. , Gladilin Evgeny 

TITLE=Automated Spike Detection in Diverse European Wheat Plants Using Textural Features and the Frangi Filter in 2D Greenhouse Images

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 11 - 2020

YEAR=2020

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2020.00666

DOI=10.3389/fpls.2020.00666

ISSN=1664-462X

ABSTRACT=Spike is one of the crop yield organs in wheat plants. The determination of the spike phenological stages including heading time point (HTP) and the area provides the necessary information for the interpretation of growth-related traits acquired from non-invasive phenotyping images. The algorithm previously developed by Qiongyan et al for the spike detection in 2-D images turns out to be less accurate when applied to the European cultivars that produce many more leaves. So, here we present an improved and extended method where (i) wavelet amplitude is used as an input to the Law’s texture energy-based neural network instead of original grayscale images, and (ii) non-spike structures (e.g., leaves) are subsequently suppressed by combining the result of the neural network prediction with the Frangi filtered image. Using this two-step approach, the overall 98.6% accuracy of neural network segmentation by direct comparison with ground-truth data could be achieved. Moreover, the comparative error rate in the spike HTP detection and growth correlation among the ground truth, Qiongyan et al, and the proposed algorithm are discussed in this paper. The proposed algorithm was also capable to significantly reduce the error rate of the HTP detection by 75% and to improve the accuracy of spike area estimation compared by 50% in comparison to the Qionagyan et al. With these algorithmic improvements, the HTP detection on diverse set of 369 plants were performed in a high-throughput manner. This analysis demonstrated that the HTP of 104 plants (comprises of 57 genotypes) with lower biomass and tillering range (e.g., earlier heading types) were correctly determined. However, the fine-tuning or extension of the developed method is required for high biomass plants where spike emerges within the green bushes. In conclusion, our proposed method allows to achieve significantly more reliable results of the HTP detection and spike growth analysis by application to the European cultivars with earlier heading types.