CAJ | 학술논문

大鱼际掌纹图像在采集的过程中会引入噪声。要求寻找一种既能很好地去噪，又能增强纹线信息的滤波方法。文中给出了脉冲耦合神经网络(PCNN)滤波和中介滤波的算法原理和具体实现方法，并根据其原理分析，指出2种滤波方法可满足以上要求。从仿真实验结果看到：中介滤波方法处理的图像比其他方法更为清晰，并且能更好地保护图像细节；另外，峰值信噪比客观的度量数据也证明了2种方法的有效性。
대어제장문도상재채집적과정중회인입조성。요구심조일충기능흔호지거조，우능증강문선신식적려파방법。문중급출료맥충우합신경망락(PCNN)려파화중개려파적산법원리화구체실현방법，병근거기원리분석，지출2충려파방법가만족이상요구。종방진실험결과간도：중개려파방법처리적도상비기타방법경위청석，병차능경호지보호도상세절；령외，봉치신조비객관적도량수거야증명료2충방법적유효성。
The noise will be introduced to the thenar palmprint image acquisition process. A method which not only bears good denoising effect,but also can enhance the ridge information is required. The principles and specific implementation methods of Pulse Coupled Neural Network(PCNN) filtering and intermediary filtering are put forward. According to the principle analysis for these two filtering algorithms, it is pointed out that these two filtering methods can meet above requirements. Simulation results show that intermediary filtering can process images more clearly and protect image details better than other methods; furthermore,the Peak Signal to Noise Ratio(PSNR) metric data also proves the effectiveness of these two methods.