摘要
由于探测器本身固有的特性以及军事上的极限使用要求,红外图像普遍存在目标-背景间对比度较差、目标边缘模糊和噪声较大等特点,采用常规模板匹配、Hu矩方法难以取得理想的检测效果。针对低信噪比红外目标图像,分析了伪Zernike矩的基本原理、不变性和计算方法,提出了低信噪比红外目标检测的伪Zernike矩方法,并比较了模板匹配、Hu矩、伪Zernike矩方法的目标检测效果。理论分析和实验验证了该方法的有效性。
As inherent characters in infrared detectors and utmost military application demands, low-SNR infrared images often present bad features such as low target-background contrast, blurred edges and large noises, so traditional target recogni- tion methods as template matching or Hu moments cannot detect target effectively. Aimed at low-SNR infrared images, the principle, invariant property and computational methods of Pseudo-Zemike moments are analyzed, and a method for low-SNR infrared target detection based on Pseudo-Zemike moments is proposed. Target recognition effects among template matching, Hu moments and Pseudo-Zemike moments are compared via experiments. Theoretical analysis and experiment results show that the proposed method is effective for low-SNR infrared target detection.
出处
《光学技术》
EI
CAS
CSCD
北大核心
2007年第6期938-941,945,共5页
Optical Technique
