基于小波边缘提取的灰度图象联合相关识别预处理

本文将小波变换方法用于灰度图象的联合变换相关识别中,采用不同的尺度因子对输入图象进行边缘提取预处理,使相关识别结果得到不同程度的改善.通过计算机模拟对比了一阶、二阶微商的边缘提取方法和小波变换边缘提取方法的预处理结果和对识别的影响,在同时衡量相关识别能力及其对噪音的敏感性前提下,小波变换边缘提取预处理明显优于各种微商边缘提取方法.调节小波变换尺度因子还能使识别能力与噪音敏感性这两方面得到更好地均衡,使小波变换边缘提取预处理能够适应不同的图象输入条件和相关输出要求.结果表明,在联合变换相关识别中采用小波变换对输入图象进行预处理是一种更理想的方法。

THE GREY IMAGE PREPROCESS BASED ON WAVELET TRANSFORM FOR EDGE DETECTION TO IMPROVE JOINT-TRANSFORM CORRELATION RECOGNITION

In this paper,the wavelet transform is used to preprocess the grey image in the joint transform correlator.Computer simulation results of multiscale edges are presented.Compared with other differential operation edge detection,wavelet transform based on multiscale edge detection is the best method for preprocess image to improve the joint transform correlation recognition.By using of the wavelet transform,the correlation recognition ability can be raised and the noise can be smoothed simultaneously The recognition ability and its sensitivity of noise can be equilibrated to suit to different input condition and output requirement by adjusting the scale.The wavelet multiscale edge detection method is an ideal preprocess method for joint transform correlation recognition.