Y. Hendrawan, H. Murase
An automated irrigation system in plant factory would be desirable since it could reduce the work load on human, increase the speed of production, improve the quality of the products and optimize the growth of the plant products. Sunagoke moss Rachomitrium japonicum is one of the plant products which are produced by plant factory. A novel irrigation system using intelligent approaches which was able to minimize the water use in plant factory as well as to optimize the growth of moss by maintaining the optimum water content was introduced. The overall goal of this research was to develop a coupled computer vision and neural network system for detection of water content of cultured Sunagoke moss. Machine vision was used as non-destructive sensing to extract image features. Back-propagation Neural Network (BPNN) had been tested successfully to describe the relationship between the water content of moss and image features. However, the prediction performance of BPNN could be improved through feature selection techniques.
Department of Agricultural Engineering, Faculty of Agricultural Technology, Brawijaya University, Malang, Jl. Veteran, Indonesia; Bio-production Engineering Laboratory, Department of Mechanical Engineering, Graduate School of Engineering Osaka Prefecture University, Sakai, Osaka, 1-1 Gakuen-cho, Japan
