商标图像平滑滤波算法研究与应用

Study and Application of Filtering Algorithm for Trademark Image

为了更加准确的完成商标图像的自动检索,在预处理过程中,针对商标图像中存在的噪声,研究了图像平滑滤波算法,分别利用均值滤波、中值滤波和改进的中值滤波方法对商标图像进行了滤波去噪处理。实验结果表明,改进的中值滤波算法在实现去除噪声的同时,既克服了均值滤波处理后图像边缘模糊和图像色变的缺点,也克服了中值滤波法处理后图像形变的不足。

In order to retrieve trademark image accurately and eliminate noise in trademark image pre-processing,some filtering algorithms are investigated.In these methods,mean filtering,median filtering and an improved median filtering are used to achieve its smoothing filtering.Experimental results show that the improved median filtering can not only has good effect in eliminating noise,but also overcome the shortcomings of marginal vague and color-change in mean filtering algorithm,and the image deformation in median filtering algorithm.