基于最优梯度估计及小波变换的指纹图像处理技术

Fingerprint Image Processing Technology Based on BEG and Wavelet Transform

提出了最优梯度估计算法 ,用于提取指纹图像局部窗口中的纹理梯度特征参数 ,并在此局部窗口内计算出相应的频率参数 .小波滤波器再根据这些特征参数对指纹图像进行自适应滤波 ,并用滤波后的合成梯度强度对滤波性能进行了衡量 .实验结果表明 ,该算法较好地去除了图像中的噪声 ,并较大程度地提高了图像的质量 .

Best estimate of gradient (BEG) algorithm was proposed, which is used for extracting the feature parameters of texture gradient of fingerprint image in local window. And it calculated the frequency parameters in the local window at the same time. The wavelet filter did the adaptive filtering using these parameters. At last the composite gradient intensity was used to test performance. The experiment results show that the method can well remove the image noise and enhance the fingerprint image quality.