系统工程与电子技术 ›› 2020, Vol. 42 ›› Issue (8): 1679-1684.doi: 10.3969/j.issn.1001-506X.2020.08.06

• 电子技术 • 上一篇    下一篇

基于改进滤波器和图像多尺度变换的背景抑制算法

花思齐(), 赵伟(), 刘建业()   

  1. 南京航空航天大学自动化学院, 江苏 南京 211100
  • 收稿日期:2019-11-25 出版日期:2020-08-01 发布日期:2020-07-27
  • 作者简介:花思齐(1996-),女,硕士研究生,主要研究方向为图像处理、目标检测。E-mail:h47@nuaa.edu.cn|赵伟(1971-),男,副教授,博士,主要研究方向为数据融合、组合导航。E-mail:zhwac@nuaa.edu.cn|刘建业(1957-),男,教授,博士,主要研究方向为检测技术、自动化装置。E-mail:ljyac@nuaa.edu.cn
  • 基金资助:
    国家自然科学基金(61603181);中央高校基本科研业务费专项资金(NS2018021)

Background suppression algorithms based on improved filter and image multi-scale transformation

Siqi HUA(), Wei ZHAO(), Jianye LIU()   

  1. College of Automation Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 211100, China
  • Received:2019-11-25 Online:2020-08-01 Published:2020-07-27
  • Supported by:
    国家自然科学基金(61603181);中央高校基本科研业务费专项资金(NS2018021)

摘要:

红外目标的探测背景具有复杂性和非平稳性,为提高后续检测性能,常常通过抑制背景来增强目标能量。针对传统背景抑制方法检测率低、虚警率高的问题,提出一种基于改进滤波器和图像多尺度变换的复杂背景抑制算法。首先,对红外图像进行改进的滤波处理获得预处理图像。其次,通过高斯金字塔多分辨技术,平滑图像背景。然后,采用Cubic插值算法提高图像的分辨率,得到背景估计图像。最后,将预处理图像和背景估计图像差分,获得背景抑制的结果。经验证,该算法实时性相对传统方法提高了19%,对于多种复杂的背景情况具有良好的适应性。同时,算法计算复杂度较低,有利于实现实时性工程应用。

关键词: 红外图像, 背景抑制, 图像多尺度变换, 插值算法

Abstract:

The detection background of the infrared target is always complex and non-stationary. In order to improve the performance of target detection, the target signal is often enhanced by suppressing the background. Aiming at the problems of low detection rate and high false alarm rate of traditional background suppression methods, a c omplex background suppression algorithm based on improved filter and image multi-scale transformation is proposed. Firstly, the infrared image is filtered to obtain the preprocessed image. Secondly, the image background is smoothed by Gaussian pyramid multi-resolution technology. Thirdly, the Cubic interpolation algorithm is used to improve the resolution of the image and get the background estimation image. Finally, the difference between the preprocessed image and the background estimated image is estimated to obtain the result of background suppression. Experiments show that the real-time performance of the proposed algorithm is 19% better than that of the traditional method, and it has good adaptability to a variety of complex background situations. Meanwhile, the algorithm has low computational complexity, which is conducive to the realization of real-time engineering applications.

Key words: infrared image, background suppression, image multi-scale transformation, interpolation algorithm

中图分类号: