基于不同叶位的茄子花期光合速率预测模型研究

摘要

在研究作物不同叶位光合速率差异的基础上,以花期茄子植株为试验材料, 设计嵌套试验, 利用回归型支持向量机算法(SVR) 实现多元非线性回归拟合, 建立了一种融合不同叶位的光合速率预测模型, 采用异校验方法对该模型进行验证分析.结果表明: 考虑叶位影响下的光合速率预测值与实测值决定系数为 0. 9960,均方误差为 0. 2503,平均相对误差为 5. 47％ ,平均绝对误差为 0. 068.该试验结果为研究作物株间按需补光以及不同叶位高效智能补光奠定了良好基础.%On the basis of studying the difference in photosynthetic rate of leaves at different positions, the eggplant crops during flowering state were used as experimental materials in nested experiments,in support vector regression algorithm (SVR) in multivariate nonlinear regression analysis and a multi-leaf photosynthetic rate prediction model was established. The model was verified by different calibration methods. Results showed that affected by different leaf positions,the determination coefficient of photosynthetic rates between actual measured and predicted values reached 0. 996,predicted of 0. 2503,the mean relative error was 5. 47％ ,the mean absolute error was 0. 068. The findings lay a good foundation for practices of supplementing lighting between crops and for leaves of specific positions.