Development of an Intelligent Quality Control Model Based on Speaking Plant Approach and Kansei Information for Moss Greening Product

摘要

In this study, sub-systems of intelligent quality control based on speaking plant approach and kansei information were proposed. It consists of quality and quantity (growth) model. It utilizes Artificial Neural Network (ANN), plant response, kansei index and texture analysis. The first ANN model for quality is proposed to define the relationship between textural features and kansei index. Kansei index is measured using visual appearances as the representation of plant factory owner. The target pointof the model is customer of moss product. The second ANN model for growth is proposed to define the relationship among plant response, textural features and temperature. Plant response is measured by using wet weight. The target point of the model is plant factory parameter. Four cycles of re-watering treatment were done based on two different local environments inside the same optimum environmental set point (global). Temperature of 1CPC and RH of 75% was considered as the optimum environment for the moss. The textural features have shown the various pattern compared with the changes of wet weight. It shows the difference pattern with our previous research (Ushada etal., 2006a) due to occurrence of growth. The research result shows that texture analysis is possible to be used as pattern recognition tool not only for quality but also for growth model. The first ANN model with satisfied inspection error can be used to predict the customer preferences while the second ANN model with satisfied inspectionerror can be used to predict the optimum local temperature.