Growth estimation of leaf vegetable plants using camera image analysis

Abstract

Abstract Fresh weight and leaf canopy area of vegetable plants are the main parameters of plant growth in determining productivity and yield quality, affecting consumer acceptance. Visual observations and manual measurements are time-consuming, labor-intensive, and subject to high errors. Plant phenotyping technology can be applied to measure physiological and growth conditions and predict plant productivity non-destructively. The objective of the study was to estimate the canopy area and fresh weight of leaf vegetable plants non-destructively using camera image analysis. Pakcoy mustard and romaine lettuce are grown hydroponically in a greenhouse. Plant images were taken using a digital web camera installed on a photo box set with dimensions of 90cm×60cm×90cm (l×w×h). One hundred images of each pakcoy and romaine lettuce plant were analyzed using image processing to see the profile of the relationship between the canopy area and the fresh weight of vegetable plants. Camera image analysis can describe leaf vegetable plants’ growth profile in greenhouses. Canopy area and plant fresh weight correlated well to form a quadratic equation with R2 0.96 for pakcoy and R2 0.98 for romaine lettuce. The plant growth estimation model built using the results of camera image analysis can estimate the parameters of the canopy area and fresh weight of leaf vegetable plants with high accuracy.