National Research Council of Canada. Aquatic and Crop Resource Development
IEEE 30th Canadian Conference on Electrical and Computer Engineering (CCECE), 30 April-3 May 2017, Windsor, ON, Canada
age prediction; plant phenotyping; machine learning; pattern recognition in plant imaging
This paper presents the analysis of various regression based machine learning algorithms for image-based plant phenotyping application and proposes a technique for plant phenotyping. Capability to predict age/development stage of a plant is one of the important factors for plant phenotyping and for analysis of in-situ crops. With the developed technique, these algorithms can predict age of an Arabidopsis plant based on the given images and mutant types. Publicly available dataset containing 165 images at different development stages and with various mutant types was used for this experiment. Results show that with this technique and different regression algorithms, it can achieve 92% prediction accuracy. Comparatively, linear regression algorithms show greater prediction accuracy than nonlinear algorithms. This method of age prediction can help plant scientists and breeders for better analysis of crops.
2017 IEEE 30th Canadian Conference on Electrical and Computer Engineering (CCECE), 16963539.