论文部分内容阅读
提出一种基于振幅调制纯相位滤波器(AMPOF)和片状正交非线性广义(SONG)相关的畸变不变光学相关目标识别算法。利用振幅调制纯相位滤波器区分能力强的优点,结合综合鉴别函数(SDF)方法,提出基于振幅调制纯相位滤波器的畸变不变识别算法,达到明显优于传统综合鉴别函数方法相关输出效果的目的。在此基础上引入SONG相关技术,利用其突出的区分能力和抗噪性能,结合这两种非线性滤波技术,得到能够很好地兼顾多项性能指标的畸变不变识别算法。仿真实验结果表明,算法既提高了相关峰输出质量和多目标区分能力,又具有很强的抵御噪声能力,在输入图像被噪声严重破坏的情况下,识别度仍然能够达到97%以上。和其他结合SONG相关的畸变不变识别算法相比,该算法目标识别性能更为优良,且更实用、广义。
A distortion-invariant optical correlation-related target recognition algorithm based on amplitude-modulated pure phase filter (AMPOF) and flaky orthogonal nonlinear generalized (SONG) correlation is proposed. By using the advantage of the ability of the amplitude-modulated pure phase filter to differentiate, this paper proposes a distortion-invariant identification algorithm based on the amplitude-modulated pure phase filter combined with the integrated discriminant function (SDF) method to achieve a significantly better output than the conventional synthetic discriminant function purpose. On this basis, SONG related technology is introduced. By utilizing its outstanding distinguishing ability and anti-noise performance, combined with these two nonlinear filtering techniques, a distortion invariant identification algorithm capable of well taking into account a number of performance indexes is obtained. The simulation results show that the proposed algorithm not only improves the output quality of the correlation peak and the ability of multi-object discrimination, but also has a strong ability to resist noise. In the case of serious damage to the input image, the recognition rate can still reach more than 97%. Compared with other distortion invariant recognition algorithms related to SONG, this algorithm has better target recognition performance and is more practical and generalized.