Quantification Method for In-Vitro Tissue Culture Plants Morphology using Object Tracking and Digital Image Analysis

Manual measurement of morphology variables on in-vitro stored plants usually cause either physical damage or microorganism infection such that further monitoring of their in-vitro performance is precluded. This study adapted computer vision technology by which it is possible to conduct such measurement without physical contact or destructive test. Moreover, by applying object tracking and pattern recognition technique in the algorithm, the system could provide automatic and real time analysis. It was shown that this quantification method reach 80.2% and 87.9% in the measurement of leaf area and chlorophyll intensity. Intensity histogram and Fourier spectrum found to be the best feature for leaf recognition and interpolation usage to adjust pixel amount over the camera distance provide better estimation on leaf area.