摘要
根据高斯噪声三阶以上累积量为零,从而可以将其有效抑制的特点,提出了一种基于高阶累计量的文本图像去噪方法.与平滑滤波和中值滤波2种经典算法相比,高阶累积量的使用不仅提高了文本图像去噪效果,而且能够较好地保留图像的边缘和细节信息,保证了原有文字的完整性.实验表明,提出的去噪算法对于去除文本图像的加性高斯噪声较为理想.
Since the third order or higher of Gaussian noise cumulants are null, they can be effectively suppressed. In the paper, a denoising method for text images by use of high order cumulants was proposed. Compared with the two classical algorithms-smooth filtering and medium value filtering, the use of high order cumulants can not only improve the denoising effect, but better preserve the image information of edges and details, so that the integrality of the original characters is guaranteed. Experimental results show that the proposed algorithm is ideal for removal of additive Gaussian noise in text images.
出处
《应用科技》
CAS
2007年第10期1-4,共4页
Applied Science and Technology
关键词
图像去噪
高阶累积量
平滑滤波
中值滤波
image denoising
higher-order cumulant
smooth filtering
medium value filtering
