基于双隐含层BP算法的激光主动成像识别系统

王灿进; 孙涛; 石宁宁; 王锐; 王挺峰; 王卫兵; 郭劲; 陈娟

doi:10.3788/OPE.20142206.1639

您当前的位置：

首页 >

文章列表页 >

基于双隐含层BP算法的激光主动成像识别系统

信息科学 | 更新时间：2020-08-13

- 基于双隐含层BP算法的激光主动成像识别系统
- Laser active imaging and recognition system based on double hidden layer BP algorithm
- 光学精密工程 2014年22卷第6期页码：1639-1647
- 作者机构：
  
  1. 中国科学院大学北京,中国,100049
  2. 北京航天自动控制研究所北京,100039
  3. 中国科学院长春光学精密机械与物理研究所激光与物质相互作用国家重点实验室,吉林长春,130033
- 作者简介：
- 基金信息：
  
  吉林省重大科技攻关专项资助项目（No.20120615）()
- DOI：10.3788/OPE.20142206.1639
  中图分类号： TN249;TP391
- 收稿日期：2013-04-10，
  
  修回日期：2013-05-10，
  
  纸质出版日期：2014-06-25
- 稿件说明：
移动端阅览
王灿进, 孙涛, 石宁宁等. 基于双隐含层BP算法的激光主动成像识别系统[J]. 光学精密工程, 2014,22(6): 1639-1647

WANG Can-jin, SUN Tao, SHI Ning-ning etc. Laser active imaging and recognition system based on double hidden layer BP algorithm[J]. Editorial Office of Optics and Precision Engineering, 2014,22(6): 1639-1647
王灿进, 孙涛, 石宁宁等. 基于双隐含层BP算法的激光主动成像识别系统[J]. 光学精密工程, 2014,22(6): 1639-1647 DOI： 10.3788/OPE.20142206.1639.

WANG Can-jin, SUN Tao, SHI Ning-ning etc. Laser active imaging and recognition system based on double hidden layer BP algorithm[J]. Editorial Office of Optics and Precision Engineering, 2014,22(6): 1639-1647 DOI： 10.3788/OPE.20142206.1639.

摘要

在传统激光主动成像系统的基础上，结合目标识别技术搭建了一个激光主动成像识别系统实验平台，用于研究激光主动成像后的目标识别。介绍了实验平台的工作原理，基于Hu矩特征的双隐含层BP神经网络算法以及实验处理流程和实验结果。特征量由7个不变Hu矩构成，通过240张原始目标样本库对由136个权值系数构成的双隐含层BP神经网络算法进行了训练。利用训练好的双隐含层BP算法对黑夜条件下远处的运动目标43式冲锋模具枪进行了实验研究，成功获得了清晰的红外激光主动成像效果。实验显示对450 m处2 740帧和550 m处2 420帧激光主动成像图像的统计识别率达到了68.87%和72.11%，其中旋转变换下的统计识别率可达80.05%和84%，好于仿射变换的识别效果。

Abstract

An experiment platform for laser active imaging and recognition was established based on the traditional laser active imaging system to investigate the target recognition after laser active imaging. The working mechanism of the platform was introduced and the Hu moment feature based BP neural network algorithm with double hidden layers and an experimental process were given. The target feature vector was consisted of seven invariant Hu moments. The BP neural network algorithm with double hidden layers including 136 weight coefficients was trained by 240 original sample libraries. The trained BP neural network algorithm was used to research a distance moving target in the dark condition

a model of 43 submachine gun

and a clear infrared laser active image was obtained. Experiment results show that statistical recognition probability is 68.87% for 2 740 frames of images at 450 m and 72.11% for 2 420 frames of images at 550 m. The corresponding recognition probabilities from rotation transformation are 80.05% and 84%

respectively

which is better than the results by affine transformation.

关键词

Keywords

references

DAYTON D, BROWNE S, GONGLEWSKI J. Long-range laser illuminated imaging: analysis and experimental demonstration[J]. Opt. Eng.,2001, 40(6):1001-1009.

聂瑞杰,徐智勇,张启衡,等. 长距离红外主动成像系统目标获取性能分析[J]. 红外与激光工程, 2012, 41(9):2276-2282. NIE R J, XU ZH Y, ZHANG Q H, et al.. Analysis on target acquisition performance of long-range IR active imaging system [J]. Infrared and Laser Engineering, 2012,41(9): 2276-2282. (in Chinese)

庞春颖,张涛. 激光主动成像系统信噪比模型的研究[J]. 光学精密工程, 2008, 16(2):319-324. PANG CH Y, ZHANG T. Signal-to-noise ratio model of laser active imaging system [J]. Opt. Precision Eng., 2008, 16(2):319-324. (in Chinese)

王智,贾书洪,张晓辉,等. 激光主动成像的多帧后处理算法研究[J]. 光学精密工程, 2007, 15(4):615-621. WANG ZH, JIA SH H, ZHANG X H,et al.. Multiframe post processing algorithm of laser active imaging[J]. Opt. Precision Eng., 2007, 15(4):615-621. (in Chinese)

杨红丽. 目标识别与跟踪在激光主动侦查中的研究 [D]. 成都:电子科技大学, 2009. YANG H L. Research on target recognition and tracking used in laser active reconnaissance [D]. Chengdu: University of Electronic Science and Technology of China, 2009. (in Chinese)

吕砚山,赵正琦. BP神经网络的优化及应用研究[J]. 北京化工大学学报, 2001, 28(1):67-69. L Y SH, ZHAO ZH Q. Optimization and application research of BP neural network [J]. Journal of Beijing University of Chemical Technology,2001,28(1):67-69. (in Chinese)

孙红辉,王红霞,田涛. 一种基于不变矩和BP网络的目标识别方法[J]. 微电子学与计算机, 2011, 28(3):63-69. SUN H H, WANG H X, TIAN T. The recognition method of objects based on moment invariant and BP neural network[J]. Microelectronics & Computer, 2011, 28(3): 63-69. (in Chinese)

田华,石圣羽,宗晓萍. 基于不变矩特征及BP神经网络的图像模式识别[J]. 河北大学学报, 2008, 28(2): 214-217. TIAN H, SHI SH Y, ZONG X P. Pattern recognition based on moment invariant feature and BP neural network for image [J]. Journal of Hebei University, 2008, 28(2):214-217. (in Chinese)

刘天舒. BP神经网络的改进研究及应用 [D]. 哈尔滨:东北农业大学, 2011. LIU T SH. The research and application on BP neural network improvement [J]. Harbin: Northeast Agricultural University, 2011. (in Chinese)

郭婉露. 红外图像目标识别及跟踪技术研究 [D]. 哈尔滨:哈尔滨工程大学, 2011. GUO W L. Researches for infrared image target identification and tracking technology [D]. Harbin: Harbin Engineering University, 2011. (in Chinese)

浏览量

475

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

凝视型激光主动成像系统性能验证

直接测距型无扫描激光主动成像验证系统

结合光斑与目标特征的激光干扰效果评估

激光主动成像图像的多帧后处理算法研究