This title appears in the Scientific Report :
2018
Plant Screen Mobile –a new smartphone app for plant trait analysis
Plant Screen Mobile –a new smartphone app for plant trait analysis
Leaf area is one of the most important parameters to quantify plant growth and physiological function. Therefore it is widely used to characterize genotypes and their interaction with the environment. Analysis of leaf area often requires destructive measurements or elaborate imaging-based methods. C...
Saved in:
Personal Name(s): | Müller-Linow, Mark (Corresponding author) |
---|---|
Wilhelm, Jens / Briese, Christoph / Wojciechowski, Tobias / Fiorani, Fabio | |
Contributing Institute: |
Pflanzenwissenschaften; IBG-2 |
Imprint: |
2018
|
Conference: | International Plant Phenotyping Symposium, Adelaide (Australia), 2018-10-02 - 2018-10-05 |
Document Type: |
Poster |
Research Program: |
Deutsches Pflanzen Phänotypisierungsnetzwerk Plant Science |
Publikationsportal JuSER |
Leaf area is one of the most important parameters to quantify plant growth and physiological function. Therefore it is widely used to characterize genotypes and their interaction with the environment. Analysis of leaf area often requires destructive measurements or elaborate imaging-based methods. Consequently, there is a clear trend towards simple and affordable sensor solutions and methodologies. A major focus is on applications developed for smartphones that provide access to analysis tools to a wide user basis. With the development of the Plant Screen Mobile app we want to contribute to this development by providing a suitable solution for Android-based smartphones to estimate proxies for leaf area and biomass in various imaging scenarios from lab to field. To distinguish between plant and background the core of the app comprises different classification approaches that deliver results on-the-fly or in batch mode. The resulting clusters of pixels reflect projected leaf, shoot or root area, respectively, and can also be used to count plant organs like flowers, fruits or seeds. We included a simple camera calibration that allows to convert pixels into metric measures as well as a genetic algorithm for parameter optimization. We tested our approach on different plant species with contrasting shoot architecture and found high correlations between our proxies and ground truth measures for leaf area and biomass. Additionally, we compared our results to a more sophisticated imaging and processing method using support vector machines and found only marginal differences. |