摘要
自动拼图问题可应用到文物修复、碎片拼接等领域,有重要的研究意义。针对目前研究较少的纹理特征在自动拼图中的应用进行了研究,利用二维傅里叶变换的频谱能够很好地描述图像纹理的粗糙度和方向性的特点,采用楔环统计方法提取纹理的粗糙性和方向性特征,然后采用欧式距离分类器的方法,进行相邻拼块的查找,取得了很好的效果,为后续利用形状特征实现精确匹配提供参考,提高了拼接的速度。
The puzzle assembly problem is important in that it can be applied to many areas such as restoration archeological findings,repairing of fragments,etc.In the paper,the authors investigate the texture features which,study less in current in this field,be applied to the problem,2D FFT spectrum is good at describing the roughness and direction of texture.So using the wedge and ring statistic method to gain the roughness and direction feature then calculating the euclidean distance of the rough- ness and direction to find the matching puzzles have acquired better effect.The method can improve the speed of automatic puzzles.
出处
《计算机工程与应用》
CSCD
北大核心
2008年第30期199-201,共3页
Computer Engineering and Applications
关键词
纹理分析
自动拼图
傅里叶变换
楔环特征
欧式距离
texture
puzzle assembly
Fourier transform
wedge and ring feature
Euclidean distance
