基于改进型非线性最小二乘法的作物模型隐含参数估计

为了使作物生长模型中的参数最优化，提出了一种改进型非线性最小二乘法在作物生长模型中隐含参数估计的应用技术，针对多目标项的情况提出了权矩阵的自动计算方法。通过良好的人机交互界面，使得该方法更精确、更有效、更方便。实验表明：用该方法对作物模型进行参数优化后，再对番茄生长的动态过程进行仿真，对模拟值和实测值进行比较，两者拟合程度很好。

Application of Nonlinear Least Squares Method to Crop Growth Structure-function Simulation Model

In order to optimize the parameters in a crop growth simulation model, a new m ethod based on nonlinear least squares method was developed for estimation of hi dden parameters of crop growth. In allusion to multi-target items we put forwar d a new way to compute matrix. The nonlinear least squares method was adopted fo r the estimation of hidden parameters, in which a better convergence and less co mputing time could be obtained and users might adjust parameters expediently. Th e methods proved to be effective when applied in a tomato harvesting process. Th e result of simulation coincides with that from the experiments.