Abstract: BP network learning algorithm has the shortcomings of slow convergence speed, easying to fall into local minimum. In order to solve this problem, this paper presents a remote sensing water depth inversion model based on chaotic immune optimization Radial Basis Function(RBF) network. In the model, it introduces the water depth remote sensing inversion principle, the RBF network center vector and weights are optimized by using Chaotic Immune Optimization Algorithm(CIOA), and the CIOA RBF network model are applied in the comparing experiments of remote sensing water depth inversion. Experimental results show that the average absolute error of this model between the inversion water depth and measured water depth is 0.436 7 m, the average relative error is 8.91%, the mean square root error is 0.563 5 m, this model has good nonlinear mapping ability and generalization ability.

Key words: water depth remote sensing, Chaotic Immune Optimization Algorithm(CIOA), Radial Basis Function(RBF), inversion model, generalization ability

摘要： BP网络学习算法存在收敛速度慢、易陷入局部极小的缺点。为此，提出一种基于混沌免疫优化径向基函数(RBF)网络的遥感水深反演模型。介绍水深遥感反演原理，利用混沌免疫优化算法对RBF网络的中心向量及连接权值进行优化，并将优化的RBF网络应用于遥感水深反演。实验结果表明，该模型反演水深和实测水深之间的平均绝对误差为0.436 7 m，平均相对误差为8.91%，均方根误差为0.563 5 m，具有较好的非线性映射能力和泛化能力。

关键词: 水深遥感, 混沌免疫优化算法, 径向基函数, 反演模型, 泛化能力

CLC Number: 

• TP391