| 基于BP神经网络的红外焦平面非均匀性校正技术 |
| |
| 引用本文： | 王娴雅,陈钱,顾国华,白俊奇.基于BP神经网络的红外焦平面非均匀性校正技术[J].光子学报,2009,38(6):1504-1506. |
| |
| 作者姓名： | 王娴雅 陈钱 顾国华 白俊奇 |
| |
| 作者单位： | 南京理工大学,电子工程与光电技术学院,南京,210094 |
| |
| 基金项目： | 武器装备预先研究项目 |
| |
| 摘 要： | 在传统BP神经网络非均匀性校正理论的基础上,结合红外焦平面阵列读出通道的时变特性,提出一种新的输出期望估计方法.该方法利用当前像素邻域和读出通道估计输出期望值.实验表明,算法能有效抑制焦平面固定图案噪音,提高场景目标的分辨能力,达到较好的视觉效果.
|
| 关 键 词： | 红外焦平面阵列 非均匀性校正 BP神经网络 |
| 收稿时间： | 2008-03-28 |
| 修稿时间： | 2008-10-09 |
A Nonuniformity Correction Technique of IRFPA Based on BP Neural Network |
| |
| WANG Xian-Ya,CHEN Qian,GU Guo-Hua,BAI Jun-qi.A Nonuniformity Correction Technique of IRFPA Based on BP Neural Network[J].Acta Photonica Sinica,2009,38(6):1504-1506. |
| |
| Authors: | WANG Xian-Ya CHEN Qian GU Guo-Hua BAI Jun-qi |
| |
| Institution: | (School of Electronic Engineering and Optoelectronic Technology,NUST,Nanjing 210094,China) |
| |
| Abstract: | Traditional BP neural network non-uniformity correction was analyzed.A new expectation estimation method was presented based on time-vary feature of IRFPA read-out channels.Output expectation was estimated using current pixel neighborhood and its read-out channel.Experiments show that the presented method can eliminate Fixed Pattern Noise effectively,improve resolved ability of scene targets and achieve good visual effects. |
| |
| Keywords: | Infrared focal plane array Non-uniformity correction BP neural network |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
| 点击此处可从《光子学报》浏览原始摘要信息 |
| 点击此处可从《光子学报》下载免费的PDF全文 |
|
