基于粒子群神经网络模型反演玉米、小麦叶面积指数

Inversion of maize and wheat leaf area index based on particle swarm optimization neural network model

基于高分1号遥感影像,分别采用粒子群神经网络模型、神经网络模型和植被指数回归模型3种方法,反演廊坊市玉米、小麦叶面积指数(LAI)。结果表明,粒子群神经网络模型反演玉米、小麦LAI的精度要高于其他方法,其模型的决定系数R2均高于0.9,均方根误差均低于0.196,可满足反演精度的要求。本研究提出的基于高分1号影像的粒子群神经网络模型反演玉米和小麦LAI的方法具有一定的普适性。

In this paper, based on GF-1 remote sensing image, three methods, namely particle swarm optimization neural network model, artificial neural network model, vegetation index regression model ,were adopted to invert leaf area index (LAI) of maize and wheat in Langfang City. It was shown that the accuracy of maize and wheat LAI inversion by particle swarm optimization neural network model was the highest. The calculated determination coefficient R 2 of this method was higher than 0.9, and its root mean square error was lower than 0.196, which could satisfy the requirement of inversion precision. To sum up, maize and wheat LAI inversion based on the proposed particle swarm optimization neural network model was feasible on GF-1 images, and possessed universality.