光学相关器在自动目标识别中的应用

为使光学相关器更好地应用于自动目标识别领域,保证在较高识别速度下,具备良好的识别效果,介绍了应用于自动目标识别领域的联合变换相关器和匹配滤波相关器.针对匹配滤波相关器实现过程中的两项关键技术,在滤波器编码技术上提出了构造等相关峰综合鉴别函数以解决畸变不变识别问题.在相关峰识别技术上,针对阈值法存在的不足,提出利用BP神经网络对相关平面相关峰进行识别,采用划分感兴趣区域的方法减少了输入层神经元数目,简化了神经网络结构.借助搭建的光学相关器系统验证了该方法可对相关信号和噪声进行有效分类,提高了光学相关器的识别效果.     

Optical time-domain analog pattern correlator for high-speed real-time image recognition

The speed of image processing is limited by image acquisition circuitry. While optical pattern recognition techniques can reduce the computational burden on digital image processing, their image correlation rates are typically low due to the use of spatial optical elements. Here we report a method that overcomes this limitation and enables fast real-time analog image recognition at a record correlation rate of 36.7 MHz--1000 times higher rates than conventional methods. This technique seamlessly performs image acquisition, correlation, and signal integration all optically in the time domain before analog-to-digital conversion by virtue of optical space-to-time mapping.     

Joint transform correlator for an optical face recognition system

An opto-electronic hybrid face recognition system based on real-time joint transform correlation using two types of liquid crystal spatial light modulators is proposed. The novel digital pre-processing to improve the tolerance was confirmed experimentally by face recognition from 100 references.     

Improved optical character recognition by matched filtering

The effectiveness of linear matched filters for improved character discrimination in presence of random noise and poorly defined characters has been investigated. We have found that although the performance of the filter in presence of random noise is reasonably good (16 dB gain in signal-to-noise-ratio) its performance is poor when the unknown character is distorted (linear shift and rotation).     

Super-parallel holographic correlator with optical fixing

Seven thousand five hundred holograms are stored in 15 locations within a bar of Zn:Mn:Fe:LiNbO₃, using a signal beam that propagates along the long axis of the bar. At each location, angular multiplexing is performed with the reference light changing its orientation in two-dimensional space. The same angular multiplexing is repeated at different locations along the long axis of the Zn:Mn:Fe:LiNbO₃ bar. When operating as a holographic correlator, an input image is compared simultaneously with the records stored in all locations within the bar and is recalled accurately.     

Pattern recognition with a phase-shifting interferometric correlator. Discrimination-capability enhancement

 We study an experimental setup to perform pattern recognition by correlation methods. This setup implements the Pure Phase Correlation (PPC) method, let us say, it only takes into account the phase distributions of the Fourier transform of the scene and the target. An experimental study of a PPC method is presented. A phase measurement interferometric technique is used to extract the phase of the Fourier transform of the correlation function. Some aspects related to the intermodulation are examined. Experimental PPC with improved discrimination capability obtained with binary masks is also considered. The experiments are carried out with segmented scenes and with multiobject scenes. Received: 22 March 1996/Revised version: 17 July 1996     

Chinese character recognition for LPR application

Letter and number recognition in license plates is widely considered a solved problem in many practical license plate recognition (LPR) systems. However, Chinese character recognition for LPR application still faces many challenges, such as more complex structure, defective character, partial occlusion, and sensitiveness to affine distortion, noise, scaling, illumination variation, contamination, blurring, and so no. In this paper, a novel method of Chinese character recognition is proposed, based on SIFT feature points clustering and matching in which a center matching strategy is designed to improve recognition efficiency. Promising experimental results demonstrate that the proposed is robust to the previous adverse factors in natural scenes and acquires higher efficiency that may meet requirements in practical application.     

Compact planar optical correlator

A compact planar configuration for performing optical correlation is described. It comprises two pairs of identical holographic lenses and a holographic filter. Each pair of lenses is recorded on a single substrate; together the lenses perform exact Fourier transformation. The light between the lenses propagates inside the substrate by total internal reflection. The design and recording considerations for each of the elements along with experimental results for the overall planar correlator are presented.     

Photorefractive fixed optical correlator

A photorefractive optical correlator stored and fixed in lithium niobate is presented. The device shows good correlation characteristics together with very high output efficiency and insensitivity to optical erasure during read-out. Received: 10 May 1999 / Revised version: 21 June 1999 / Published online: 16 September 1999     

All-incoherent dispersion-compensated optical correlator

We report on a simple, spatially incoherent, wavelength-independent imaging system that, in contrast to the conventional case, exhibits a dispersion-compensated point-spread function. Our hybrid (diffractive-refractive) three-lens imaging configuration thus acts as an all-incoherent dispersion-compensated optical irradiance correlator. So the optical arrangement is well adapted to processing color information (both spatially and temporally incoherent) under natural illumination.     

A real-time programmable joint transform correlator

A real-time programmable joint transform correlator (PJTC) utilizing a magneto-optic device (MOD) with a liquid crystal light value (LCLV) is presented. Object functions to be correlated are written into a light MOD with a programmable microcomputer, as input signals. Real-time correlation is taken place by a LCLV, as a square low detector, with a coherent read out. Cross correlation between the input functions can be obtained through the inverse Fourier transform of the read-out coherent illumination. A brief discussion of the proposed PJTC and a few experimental results are provided.     

Sparse representation-based classification algorithm for optical Tibetan character recognition

A sparse representation-based two-phase classification algorithm is proposed for off-line handwritten Tibetan character recognition. The first phase realizes coarse classification with the K-NN classifier by finding the K nearest neighbours of a test sample in the dictionary constructed by K-SVD with samples of all classes, and the classes of these neighbours are regarded as the candidate classes of the test sample. The second phase performs fine classification with the sparse representation classifier by sparsely representing the test sample with all elements of the dictionary constructed by K-SVD with samples of all candidate classes, and the test sample is finally classified into the candidate class with the maximal contribution in sparse representation. Experiments on the Tibetan off-line handwritten character database show that an optimal recognition rate of 97.17% has been reached and it is 2.12% higher than that of K-NN.     

Critical adjustment of correlation filters for optical character recognition

As yet, optical correlation filters were not applied successfully in commercially available equipment since 1964. Since Vander Lugt¹ prepared the first correlation filter using a Mach Zender interferometer a considerable amount of effort both experimental and theoretical was devoted to them.

In this paper an attempt will be made to evaluate their usefulness in optical character recognition on the basis of our experimental work. No attempt will be made to survey other work in this field.

For the sake of completeness a brief outline of the theory will be given.     

Real time optical correlator with a Bil2SiO20 crystal

In this paper we propose a new design of a joint transform correlator in which two transform lenses are used to minimize the influence of their aberration, and a magnification lens is used to make the optics compact. Experimental results show the correlator can be used to recognize characters and/or halL tone images in real time.     

Engineering implementation of a hybrid optical correlator using SLMs

Optical correlators have been widely used for such applications as automatic target identification, invariant pattern recognition and machine vision. In hybrid optical correlators, Spatial light modulators (SLMs) are used to dynamically update the input and the filter. However, implementation is difficult because of the interference of “ghost” images; rigorous requirements for filter registration, accurate and effective filter sizing and complex valued filtering functions. This paper proposes some practical approaches to handle these issues and provides engineering details of how to build a hybrid optical correlator, where a SLM serves as an updateable intensity filter for real-time pattern recognition.     

基于光学相关器的光学稳像器研究

为了突破传统电子稳像和机械稳像技术在处理速度、精度等方面的局限性,提出基于光学相关器的稳像技术,旨在实现对高速视频的实时高精度光学稳像。系统使用光学联合相关变换的方法,提取视频前后两帧变换后的相关峰,给出两帧的位移差值,通过改进后的补偿算法进行实时稳像。基于光学相关器的光学稳像器能够实现平移和旋转稳像,通过仿真和实验分析,平移稳像的精度可以达到0.1个像元（改进补偿算法后可达0.04个像元）,旋转稳像的精度可达0.1。     

透射式光学相关器小型化设计

 针对光学相关器的技术指标,对其光学系统进行小型化设计。傅里叶透镜组与准直扩束系统均采用摄远型式结构,在满足其长焦距的前提下缩短了系统的光学筒长;采用直角反射棱镜折叠空间光路充分利用空间,使系统布局更紧凑;利用计算机辅助设计软件ZEMAX对光学系统结构进行优化,最后研制出的小型化透射式光学相关器总长在150 mm以内,总宽在130 mm以内,实现了小型化目的。改进后的光学系统结构为双光路处理结构,具有实时性强的优点。经实验验证,该装置具有较好的相关探测性能,并且噪声小、探测精度高。     

Programmable optical hit-miss transformation using an incoherent optical correlator

Using an incoherent optical correlator, a programmable optical morphological hit-miss transform processor with large structuring element is proposed and demonstrated. The experimental results are given.     

Fringe pattern correlator for three-dimensional object recognition

We propose a novel three-dimensional (3-D) object-recognition method based on a Fourier-transform profilometry technique and a two-dimensional (2-D) correlation technique. Height information on 3-D objects is transformed to phase information on 2-D complex amplitude by use of the Fourier-transform profilometry technique. 3-D objects are recognized using correlation by use of the transformed complex amplitude.