摘要
针对织物图像周期性较强,且特征不易提取的特点,采用元胞自动机的方法,将元胞空间取作二维,将图像的像素和元胞一一对应起来。通过建立元胞自动机模型和元胞转移方程,完成特征值提取及边缘检测。根据对摄像机的事先标定,确定出2幅图像重叠的部分,然后在相互重叠的部分寻找2幅图像之间像素值的最大相似性,以互相关系数最大为原则进行迭代搜索,求出最佳拼接点。再对拼接处产生的拼接缝隙迹进行消除,从而完成图像拼接,得到1幅完整的图像。通过具体的织物图像进行验证计算,结果表明,采用元胞自动机可有效地进行织物图像拼接。
Owing to the fact that fabric images exhibit strong periodic and their features are difficuh to extract,it is suggested to use cellular automaton to take the cellular space as a two-dlmensional one, and the mosaic of image pixels and cellular is performed one-to-one correspondingly. It completes the feature extraction and edge detection through the establishment of cellular automata model and cellular transfer equation. According to the pre-calibration of the camera, it determines overlapping parts of two images, and then searches the maximum similarity in pixel values of the overlapping parts of the two images and conducts iterative search on the principle of maximum correlated factor to get the optimal stitching point. By eliminating apertures generated from mosaicking,image mosaicking is finished and a complete image is gotten. It is shown by verification that using cellular automata can achieve fabric image mosaicking effectively.
出处
《纺织学报》
EI
CAS
CSCD
北大核心
2011年第1期29-33,共5页
Journal of Textile Research
基金
湖北省教育厅重点项目(D20091705)
关键词
元胞自动机
织物图像
特征提取
拼接
融合
cellular automata
fabric image
feature extraction
mosaic
fusion
