• 研究论文 • 上一篇    下一篇



  1. (西安电子科技大学 智能控制与图像工程研究所,陕西 西安 710071)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-04-20 发布日期:2007-04-20

A novel image mosaic method

YANG Zhan-long;GUO Bao-long

  1. (Research Inst. of Intelligent Control and Image Eng., Xidian Univ., Xi′an 710071, China)
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-04-20 Published:2007-04-20

摘要: 提出一种生成合成图像的碎片图像镶嵌算法.这种算法是将许多任意形状的碎片图像进行镶嵌生成最终需要的图像.首先利用图像融合技术将原始图像提取出来的边缘信息图像与一幅相同尺寸大小的沃龙诺依图像进行合成,然后采用优化近似查询中关于数据结构的“反极策略”来加速搜索过程,根据模式识别中的形状相似技术在数据库中查找与需要填充的区域形状最接近的碎片图像进行镶嵌.试验结果证明该算法在保证图像质量的前提下比JIM算法快近50倍.

关键词: 图像镶嵌, 沃龙诺依图, 反极策略, 边缘检测, 图像增强

Abstract: This paper introduces a new mosaic, called Tile Image Mosaic(TIM), where image tiles with arbitrary shapes are used to compose the final picture. Firstly we merge the edge detected image and the same-size voronoi-diagrams, then adopt the recent Antipole strategy technique about data structure aimed to optimize proximity queries to speed up the search process, and finally search one tile image with the similar shape in the data base to mosaic the voronoi region. Experimental results prove that our algorithm is about 50 times faster than JIM with high image quality.

Key words: image mosaic, voronoi-diagrams, antipole strategy, edge detection, image enhancement


  • TP391.41