Abstract
Both the simulation model and remote sensing have been extensively researched since the 1970s, and it is considered that the technologies have not yet reached the practical level despite the impression that they have almost reached it. In this article, The authors discussed the recent research trends, issues and future prospects of these technologies. Simulation models predicting the growth and yield of crops have been successful in future projections under the global warming environment, but has not yet been developed for cultivation support in Japan. This is thought to be due to the fact that much input data is required and the output remains within the range of known knowledge. Meanwhile, the price of UAV(Unmanned Air Vehicle) has been reduced in recent years, which makes farmers available for remote sensing. However, UAV is mostly used just to take photographs from the sky. The strategy to develop these technologies for farmers is necessary by clarifying the concept, data nececity and cultivation support. The authors conducted researches on remote sensing, simulation models, and their integration, mainly in farmer fields in Southeast Asia in relation to farmer's cultivation management and climate change impact assessment. Currently the authors are developing cultivation support technology for the large-sized field of agricultural corporations newly farmed in Sendai after the Great East Japan Earthquake. Although rice growth and yield has been successfully simulated with remote sensing data, further development is necessary in terms of effective cultivation support. Due to evolution of machine learning such as deep learning, technology to extract information from images and mass data is expected to be incorporated in agricultural sector. However, analogy from the results of data dependent models such as nonparametric models, it seems difficult to obtain information beyond the known by itself. The authors are convinced that analysis with directionality such as simulation model is required.
