基于小波变换的拓片文字边缘检测

Edge detection of rubbing text images based on wavelet transform

针对拓片得到的文字图像具有模糊细节多、效果差等特征,以及传统算法对其边缘检测的精度不高,根据拓片文字边缘独立于尺度传播的特性,提出了一种基于二进小波变换的拓片文字图像边缘提取和增强算法。首先用二进小波对拓片文字图像进行多尺度分解,再结合小波变换模值跨尺度传递的不同特性,进行多尺度下的图像边缘提取、增强和细化。实验表明,该算法克服了传统算法的不足,弱化了单尺度下噪声抑制与边缘细节提取精度之间的矛盾,从而具有更好的实用性。

The text images obtained through rubbing were featured by many fuzzy details, bad effect and so on, so it might lose more details in the traditional handling process. Proposed a new algorithm of the rubbing text image edge detection and enhancement based on dyadic wavelet transform. Firstly, transformed the rubbing text image using dyadic wavelet. Then combined with the property of cross-scale transmission for wavelet transform modulus value to extract enhance and refine the multiscale edge. Experiments show this algorithm overcomes the shortcomings of traditional method and weakens the contradiction between the noise suppression and the accuracy of detecting edge details, so it has better practicality.