Error compensation based on BP neural network for airborne laser ranging

被引：61

作者：

Wu, Bing ^{[1
]}

Han, Shaojun ^{[1
]}

Xiao, Jin ^{[1
]}

Hu, Xiaoguang ^{[1
]}

Fan, Jianxin ^{[1
]}

机构：

[1] Beihang Univ, Sch Automat Sci & Elect Engn, State Key Lab Virtual Real Technol & Syst, 37 Xueyuan Rd, Beijing 100191, Peoples R China

来源：

OPTIK | 2016年 / 127卷 / 08期

关键词：

Error compensation; Quad-rotor; BP neural network; Laser ranging; FINDER;

D O I：

10.1016/j.ijleo.2016.01.066

中图分类号：

O43 [光学];

学科分类号：

070207 ; 0803 ;

摘要：

This paper proposes a general realization method of laser ranging for micro UAVs, which is based on secondary development for the laser ranging module of FLUKE 411D. And the airborne dynamic laser ranging experiments are finished with the micro UAV DJI Phantom1. Through data processing with the least-square-based method and BP neural network-based method, the measurement error with BP neural network-based method is obtained for 4.35%, which is significantly less than that 6.69% with least-square based method. It means that the measurement accuracy is improved significantly with BP neural network based method. Therefore, the error compensation method based on BP neural network for airborne laser ranging is verified effectively in the paper. (C) 2016 Elsevier GmbH. All rights reserved.

引用

页码：4083 / 4088

页数：6

共 21 条

[11] Laser range finder model for autonomous navigation of a robot in a maize field using a particle filter [J].