Non-destructive measurement of leaf area and leaf number of hydroponic pak-choy plants (Brassica rapa)

Abstract

Leaf area and leaf number is an important factor in many agronomic and plant physiological studies. The measurements are commonly obtained by destructive and time consuming via laboratory. Therefore, there is a trend in utilizing fast and non-destructive methods to estimate leaf area and leaf number. The aim of this study was to estimate leaf area and leaf number of pak choy (Brassica rapa) under field condition. For this purpose, 180 pots of pak choy plants were captured. Afterward, the images were analysed via image processing. Distance transform-watershed segmentation was used to extract and to separate every leaf from its background. The area of each leaves were estimated according to the pixel values obtained. The area of leaves also compared to the ImageJ measurement. The number of extracted leaves was evaluated according to precision, recall, and f-measure calculation. In conclusion, our system achieves a good performance for minimal occlusion leaves and provides quick and non-destructive method. This method could potentially be applied to design phenotyping system in the greenhouse even for small scale companies.