分数相关畸变不变模式识别特性研究

本文从理论上分析了分数相关器的分数级次对相关输出面能量分布的影响,利用数值模拟方法,对基于匹配滤波结构的分数相关器畸变不变模式识别特性进行了深入的讨论,并与传统相关器进行了比较.结果表明在畸变不变模式识别中,分数相关器比传统相关器有着更强的畸变不变相关识别能力.     

基于联合广义分数相关器的相关峰特性分析

分析了联合广义分数傅里叶变换相关器相关峰的特性,得到通过改变广义分数傅里叶变换的系统参量可以提高广义分数相关峰性能的结论.进行了数值模拟和光学实验,并根据两者的结果对四个相关峰的性能指标相关峰强度最大值、峰能比、识别能力、信噪比进行了比较分析,说明只要适当控制系统参量,联合广义分数傅里叶变换相关器比联合分数傅里叶变换相关器具有更好的相关性能,有助于提高光学相关器识别的准确率.     

空变菲涅耳联合变换相关器

提出了菲涅耳联合变换相关器(Fresnel Joint Transform Correlator,FnJTC)结构,给出了数学推导过程,并通过计算机模拟验证了可行性及其与传统相关器所不同的相关特性。在相关过程第一阶段中,菲涅耳联合变换相关器对输入的联合图像作菲涅耳变换,代替了传统联合变换相关器(JTC)的傅里叶变换。与传统相关器相比,菲涅耳联合变换相关器是空变系统,相关峰值依赖于输入参考图像与目标图像的相似性、两图像之间的相对位置及参考图像(或目标图像)的线性相位调制。     

应用联合分数傅里叶变换相关器识别多个物体

通过理论分析和计算机仿真研究了联合分数傅里叶变换相关器对多个物体的识别.结果表明联合分数傅里叶变换相关器与基于傅里叶变换的联合变换相关器相比有了较大的改善,提高了目标探测的灵敏性.     

基于输入面形态学预处理的条纹调制滤波联合变换相关器

为改善联合变换相关器的相关输出结果,提高图像识别能力,提出了一种将输入面形态学处理与功率谱面FAF处理相结合的方法。首先,介绍了经典联合变换相关器的基本原理;其次,针对联合变换相关器的缺点和不足,提出将输入面形态学预处理和功率谱FAF处理相结合,以改进经典联合变换相关器的相关输出性能;最后,详细分析了上述改进型联合变换相关器的参数选取问题。实验结果表明：当参数选择为 、 时,相关峰很尖锐,峰值为 ,比传统的联合变换相关器互相关峰值提高了 。将输入面形态学处理与功率谱面FAF处理相结合效果好于单独使用其中一种方法,极大地改善了传统联合变换相关器的目标识别能力。     

基于双相位编码联合变换相关器的图像位移探测算法

提出了基于双相位编码联合变换相关器(JTC)的图像位移探测算法.随机产生一相位码对参考图像编码,与目标图像叠加作为输入图像,通过傅里叶变换得到联合功率谱,用同样的相位码对其编码并进行滤波,傅里叶逆变换后得到涉及位移矢量的互相关峰.给出了多帧图像间位移矢量的探测技术.相比传统的相关器探测技术,此算法能更有效地利用输入平面...     

面积-极坐标变换在光学相关目标识别中的应用

应用联合变换相关器对目标进行探测和识别时,由于实际探测到的目标相对于参考图像存在比例、旋转、平移畸变,因此在传统的光学相关器中很难实现目标的探测和识别。为了更好地探测到目标,提出了面积-极坐标的算法,该算法利用相似图形之间的尺寸与面积存在满映射函数关系,实现了比例不变光学探测,并结合极坐标变换实现旋转不变光学探测,最后,对变换后的联合图像进行小波边缘提取,既能提取到清晰的图像边缘,又能有效地抑制噪声,不但提高了相关峰,并使得相关峰更为尖锐,成功实现了大尺度混和畸变目标的探测和识别。作为实例,对混合畸变目标进行了计算机模拟和光学实验,验证了该算法在大尺度混合畸变光学相关探测中的可行性。     

光学相关法测量空间相机像移的性能研究

为了测量由卫星姿态不稳定或振动等原因引起的空间相机的亚像元像移,使用了光学联合变换相关器(JTC)对安装在相机焦面上的辅助面阵CCD采集到的相邻两帧图像进行相关运算.给出了使用该方法测量像移的原理,搭建了实现JTC的光学实验平台,对JTC测量相机像移的性能进行了实验研究.结果表明,使用JTC完全可以测量空间相机的亚像元...     

多目标识别的联合变换相关器的研究

提出了一种可用于多目标识别的联合变换相关器。为改善相关信号的性能,对功率谱作了优化处理。为消除相关面上的零级项和目标间的相关项,可用联合功率谱减去纯目标输入的功率谱和参考图像的功率谱;为增强和锐化相关峰,将相减的功率谱作指数函数滤波处理。分析了指数滤波参数对相关结果的影响。计算机模拟结果表明,这种相关器所输出的相关信号比经典联合变换相关器和二元联合变换相关器输出的相关信号更好,互相关得到了抑制,自相关得到了增强,具有很好的抗噪能力。     

用于指纹识别的基于离散余弦变换的二值化联合变换相关器

提出了一种利用离散斜弦变换对联合功率谱进行处理的二值化联合变换相关算法,利用这种二值联合变换相关器,进行了一个待识别指纹同四个参考指纹的比较,这种相关器同采用局部中值阈值技术的二值化联合变换相关器相比,能提高输出峰边比,提高相关峰的强度。光电混合处理实验的结果表明,这种方法能大大提高多指纹识别系统的识别性能。     

Color pattern recognition based on the joint fractional Fourier transform correlator

A new system of multi-channel single-output joint fractional Fourier transform correlator (JFRTC) for color pattern recognition is proposed based on the conventional system of multi-channel single-output joint transform correlator (JTC). The theoretical analysis and optical experiments are performed. With this method, one can obtain three correlation peaks at the output plane which show a pair of desired cross-correlation peaks and one auto-correlation peak. In comparison, the conventional system leads to more correlation peaks playing a noise role in color pattern recognition.     

Improved approach for target detection and recognition on optoelectronic hybrid joint transform correlator

A conventional joint transform correlator (JTC) often generates large correlation sidelobes as well as a large correlation peak width, strong zero-order peak, and low diffraction efficiency in target detection and recognition, which make the detection ability of JTC lower. To conquer these difficulties, firstly, a joint power spectrum (JPS) subtraction technique was proposed in Fourier plane, where power spectrum of reference image and power spectrum of object image are subtracted from the JPS before inverse Fourier-transform operation, it is evident that the improved JPS removes the zero-order term. Secondly, a fringe-adjusted filter (FAF) was presented to restrain sidelobes and noises. The revised JPS is multiplied by a FAF before the inverse Fourier-transform operation to obtain the cross-correlation peak. Computer simulations showed the improved method can markedly eliminate zero-order diffraction and effectively control the sidelobes and noises compared to traditional JTC, and then enhance the detection ability for JTC. Experimental results presented the sharp correlation peak and also demonstrated this approach effectiveness.     

Improved target detection and recognition in complicated background with joint transform correlator

We present an effective method to enhance target detection and recognition by Canny algorithm, which can obtain the continuous and distinct image edges, also suppress back-ground noise effectively, for this reason, correlation output processed by our proposed method provides higher discrimination than the classical JTC. Computer simulations demonstrate the improved method can obviously suppress noise and sidelobes, also enhance the performance of detection and recognition for JTC. An optoelectronic hybrid JTC is constructed to verify the efficiency. Experimental results present the sharp correlation peak and also confirm the method effectiveness.     

Image displacement based on phase-encoded reference joint fractional transform correlator

In order to utilize the space of an input plane efficiently and make the optical structure more flexible, an image displacement measurement based on phase-encoded reference joint fractional transform correlator (PER-JFrTC) is proposed. We use a random phase mask to encode the reference image and overlay it with the target image forming the input image. Joint power spectrum (JPS) of the input image is obtained by Fourier transform and the resultant is encoded by the same phase mask. Then a fractional Fourier transform with an order p is applied to the phase-encoded JPS (PJPS), resulting in a correlation output with a sharp cross-correlation peak, which includes the displacement information between the reference and the target image. Contrast to displacement measurement based on traditional joint transform correlator (JTC), PER-JFrTC can use the space of the input plane efficiently and reduces the influence of the auto-correlation. Also the position of cross-correlation peak can be fixed arbitrarily according to the fractional order p as well as the optical set-up can be more flexible and easier to implement. Results based on digital computation show that PER-JFrTC could detect the displacement accurately and verify our proposal. A possible optical set-up is suggested.     

Extended fractional wavelet joint transform correlator

An extended fractional wavelet joint transform correlator is implemented for real-time target recognition applications. The real-time input scene captured using a charge-coupled device camera along with the reference image is fractional Fourier transformed. The obtained joint power spectrum is multiplied by an appropriately scaled wavelet filter and the resultant function is differentiated. The application of wavelet filter enhances the correlation outputs and differential processing of wavelet-filtered joint power spectrum improves the detection efficiency by reducing the zero-order spectra. Targets with Gaussian and speckle noise have also been used to check the correlation output. The performance metrics: correlation peak intensity, peak-to-correlation energy, peak-to-sidelobe ratio and signal to clutter ratio have been calculated. The experimental results are presented in support of the proposed idea.     

最大平均相关高度滤波算法在畸变目标识别中的应用

为了解决在联合变换相关器上无法识别复杂背景下畸变目标(目标的平面内旋转或尺寸改变)的问题,对其应用最大平均相关高度滤波算法进行畸变目标识别.首先在频域内优化最大平均相关高度滤波器的三个性能控制参量|然后将优化好的最大平均相关高度频域滤波器映射到物空间,得到最大平均相关高度参考模板.作为实例,本文对复杂背景下的军舰和汽车目标进行了光学识别实验,结果表明,当滤波器的控制参量为γ=1,α=0,β=0.1时,得到的MACH参考模板最为清晰,目标产生的相关峰最为尖锐明亮.通过这种方法,在联合变换相关器上能够得到的比例畸变容差为28%,旋转畸变容差为20°,对于超出上述畸变容差范围的目标,相关峰强度逐步减弱,但部分仍然可以识别.改进后的MACH滤波器在畸变目标的探测领域具有很强的优势和良好的应用前景.     

Displacement measurement based on joint fractional Fourier transform correlator

A displacement measurement technology based on joint fractional Fourier transform is firstly proposed. Contrast to conventional displacement measurement based on joint Fourier transform correlator, the position of cross correlation peak in the proposed technology could be fixed arbitrarily according to the order of fractional Fourier transform. The optical setup in the proposed technology is more flexible and easier to implement. Simulation and experiment results are given out to verify the analysis.