Self-Organizing GA for Crop Model Parameter Estimation using Multi-resolution Satellite Images

要旨

We present a methodology for estimating the parameters for crop assimilation studies from satellite images. The procedure is optimized with an evolutionary search technique. A Genetic Algorithm (GA) operates well in high-dimensional non-linear domains. However, its parameters must be set in advance. In this paper, we use a self-organizing GA, in which the initial parameters are generated and assigned automatically. Numerical experiments were conducted to analyze the performance of the methodology, and our method’s effectiveness on both synthetic and real satellite data was proven. This study shows that the self-organizing GA methodology is better than the conventional GA approach in estimating crop assimilation.