Crop quality control system: a tool to control the visual quality of pot plants

Key words: quality, growth, model, leaf unfolding rate, internode, plant height, plant width, leaf area, temperature, plant spacing, season, light, development, image processing, grading, neural network, pot plant, Ficus benjamina 'Exotica'.The market is increasingly dictating the specifications for products. A well-defined marketable product must be delivered at a defined moment in time. A system was developed for growers to control development and growth of pot plants to achieve a defined quality on a specified delivery date.A theoretical framework of quality modelling used in the food industry was successfully implemented within the domain of pot plants. The visual quality of Ficus benjamina 'Exotica' was explained for 88.7% by four variables (in addition to the fixed plant height of 1.20 m.): the density of the leaf mass in front-view, 40-60 cm above pot rim; the width in side view, 40-60 cm above pot rim; the width in front-view, 0-20 and 20-40 cm above pot rim. Effect of temperature, plant density (control factors) and season (not controlled) on the growth and development of Ficus in relation to these quality features was quantified. Development of Ficus benjamina 'Exotica' showed a regular pattern. Plant height increased with temperature up to 29.5°C. The effect of plant density on this variable was small. Leaf unfolding rate was influenced by light and temperature, plant density was of minor influence. Its response to temperature showed an optimum at 30°C. The final internode length was mainly affected by light intensity, temperature had no effect. The width of the plant and the density of leaf mass increased at lower plant densities. The effect of temperature on visual quality of Ficus was small. A simple model based on only two major quality features, plant height and plant width, was developed and could successfully control visual quality of Ficus . Plant height as a function of plastochron age and internode length in relation to temperature and light was adequately predicted. Plant width was successfully controlled by spacing operations, using image processing to monitor this crop feature.By integration of four functional modules, a model for visual quality, a crop growth control model, image processing to monitor crop development and a neural network to grade plants, a crop quality control system (CQCS) was obtained and successfully implemented. This system was compared to other systems and its limitations and transferability to other crops was discussed. Furthermore, the system was evaluated for commercial practice.