J4 ›› 2009, Vol. 36 ›› Issue (6): 1070-1074.

• 研究论文 • 上一篇    下一篇

一种自适应红外图像增强技术

寇小明1,2;刘上乾1;洪鸣1;汪大宝1
  

  1. (1. 西安电子科技大学 技术物理学院,陕西 西安  710071
    2. 中船重工第705研究所,陕西 西安  710075)
  • 收稿日期:2009-05-10 出版日期:2009-12-20 发布日期:2010-01-20
  • 通讯作者: 寇小明
  • 作者简介:寇小明(1967-),男,中船重工第705所研究员,西安电子科技大学博士研究生,E-mail: casscamm@126.com.
  • 基金资助:

    教育部科学技术重点研究项目资助(108114)

Approach for adaptive infrared thermal image enhancement

KOU Xiao-ming1,2;LIU Shang-qian1;HONG Ming1;WANG Da-bao1
  

  1. (1. School of Technical Physics, Xidian Univ., Xi'an  710071, China
    2. No. 705 Research Institute CSIC, Xi'an  710075, China)
  • Received:2009-05-10 Online:2009-12-20 Published:2010-01-20
  • Contact: KOU Xiao-ming

摘要:

针对红外热成像对比度不足的问题,提出了一种S曲线映射函数,该函数具有定义域与值域区间一致、拐点非对称的特点,在目标图像增强的同时拟制了杂波背景.为了满足动态场景自适应增强的需要,建立了基于场景灰度特征调整S曲线参数的动态自适应模型,该模型用灰度直方图的阈值和调制度作为输入,用图像增强算法的参数作为输出,并基于典型样本数据通过人工神经网络的BP算法确定出了模型中权系数的值,工程应用实例表明这种方法是切实有效的.

关键词: 红外成像, 背景抑制, 图像增强, 自适应, BP算法

Abstract:

As for the inadequate contrast of infrared thermal imaging, as S curve mapping function is proposed to enhance target objectives and in the meantime suppress background clutters, which has the same range of domain with a variable interval and a non-symmetrical inflection point. With the need of adaptive enhancement for the changing scene, a dynamic self-adaptive model is proposed to adjust the parameters of the S curve based on the scene. By using the threshold and modulation of histogram as the input parameters and the parameters of the enhancement algorithm as the outputs, the coefficients of the model is determined based on typical samples of data through the artificial neural network BP algorithm, with the application showing that this approach is effective.

Key words: infrared imaging, background suppression, image enhancement, adaptive, BP algorithm

中图分类号: 

  • TN911.73