Non-destructive measurement system for plant growth information based on machine vision

A non-destructive measurement system for plural plants morphological characteristics of was developed and constructed. A CCD camera (Sony, XC-77) with an infrared filter and an 850 nm infrared LED lighting source was installed on a stepping motor (Oriental motor, DG130R-ASAA). The all plants images was captured and stored by rotated the stepping motor 20° according to the time interval that can be preset to any interval 10 min was used in this study. The images that captured by infrared irradiation take the same level of intensity value during day and night. The reflective marker was attached at the node position of a plant; stem elongation up to the node position could be measured by tracking the reflective marker. The image analysis and measurements were performed using an image processing of area-based pattern matching technique. Plant stem node coordinates were computed by the image processing algorithms―the sub-pixel estimation on area-based matching technique. One of the problems in the pattern matching technique is determination of an appropriate reference image size. The result of the research shows that the measurement accuracy became raise if the reference window size is more than 15×15 pixel in this research. The system was used to measure the stem elongation of Chrysanthemum (Dendranthema×grandiflorum “Reangan”). In the research, the daily pattern of chrysanthemum stem elongation was constructed, and it was observed that the daily stem elongation was greater than that during the night.