计算机自动拼图中颜色匹配算法的研究与应用  

ON COLOUR MATCHING ALGORITHM IN AUTOMATIC COMPUTER IMAGE STITCHING AND ITS APPLICATION

在线阅读下载全文

作  者:姬照中[1] 葛庆平[1] 刘宁[1] 

机构地区:[1]首都师范大学信息工程学院

出  处:《计算机应用与软件》2009年第7期68-71,共4页Computer Applications and Software

基  金:北京市教委项目(KM200710028018)

摘  要:计算机自动拼图是利用计算机自动完成拼图游戏,它可以作为教育软件展示使用颜色、纹理和形状进行图像匹配的过程,也可以应用于考古中的文物碎片拼接等领域。针对计算机智能拼图中拼块匹配的问题,提出应用拼块的颜色信息进行匹配的方法。颜色作为一种重要的视觉特征可以提高系统的执行效率并提高匹配精度。算法采用符合人类视觉特性的HSV颜色空间,利用人对颜色的感知特性对颜色分量进行非等间隔的量化并形成特征矢量,应用直方图相交距离进行了特征的匹配,并且针对拼块图像自身的特点采用了图像分割技术对拼块图像的背景进行剔除的同时实现了按方向进行匹配的方法。实验表明,该方法可以有效地进行拼块图像的匹配。Automatic computer image stitching is a system using computer to automatically implement image stitching work. It can be applied as education software to display the image matching process with colour, texture and shape information, and can also be applied to many other areas such as stitching culture relic fragments in archaeology, etc. In view of pieces matching issue in intelligent computer image stitching, an approach that uses colour information of the pieces to match is proposed in this paper. Colour is deemed as very important vision feature and can improve the system executing efficiency and matching accuracy. This algorithm uses HSV colour space which is more compatible with human' s visual sense feature and takes advantage of human' s perception feature on colour to quantise colour components in non-equal interval and to form the eigenvector. Then it uses histogram intersection distance algorithm to match pieces' feature, and according to the characteristic of stitching pieces of image its own, the image segmentation technique is adopted to cull the background of the piece, and simul- taneously the approach to match the pieces by direction is implemented. Experiment results show that the algorithm can match the stitching pieces of image effectively.

关 键 词:自动拼图 颜色直方图 HSV颜色空间 图像检索 

分 类 号:TP242[自动化与计算机技术—检测技术与自动化装置] TP391.41[自动化与计算机技术—控制科学与工程]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象