阵列工业图像快速拼接算法
DOI:
作者:
作者单位:

淮北师范大学计算机科学与技术学院

作者简介:

通讯作者:

中图分类号:

TP391.4;TP301.6

基金项目:

安徽高校自然科学研究重大项目(KJ2017ZD32)、安徽省自然科学基金资助项目(1808085MF174,1808085QF181)、安徽省高校管理大数据研究中心项目(AHDSJ20170120)和医学物理与技术安徽省重点实验室开放基金资助项目(LMPT201706)


Fast Mosaic Algorithms for Array Industrial Images
Author:
Affiliation:

College of Computer Science and Technology, Huaibei Normal University

Fund Project:

Major Natural Science Research Projects of Anhui Universities (KJ2017ZD32), Natural Science Foundation of Anhui Province (1808085MF174,1808085QF181), University Management Center of Anhui Province (AHDSJ20170120) and Key Laboratory of Medical Physics and Technology of Anhui Province (LMPT201706)

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    相机阵列获取的多目图像拼接常依赖特征点的数量和空间分布情况,采用传统Harris角点检测算法提取特征点时,会产生伪角点和角点簇,影响拼接速度和精度。本文自定义Harris角点检测阈值,保留明显特征点,利用自适应非极大值抑制(ANMS)优化角点簇问题;再生成特征描述子,实现特征点的初步匹配,接着用RANSAC算法去除误匹配;最后采用渐入渐出融合算法,实现了拼接图像的平滑过渡。实验结果表明,本文算法能够快速、精确地实现工业图像拼接,具有优良的工业实用性。

    Abstract:

    Multi-view image mosaic acquired by camera array often depends on the number and spatial distribution of feature points. When traditional Harris corner detection algorithm is used to extract feature points, false corners and corner clusters will be generated and the speed and accuracy of image stitching will be negatively affected. In this paper, a Harris corner detection self-defined threshold was defined to reserve the obvious feature points, and Adaptive Non-maximum Suppression (ANMS) was used to optimize the corner clusters. Then the feature descriptor was regenerated to achieve the initial matching of feature points, and the random sampling agreement algorithm (RANSAC) was utilized to remove the mismatching. Finally, the smooth mosaic image was achieved by using the gradual-in-gradual-out fusion algorithm. The experimental result shows that the proposed algorithm can achieve industrial image mosaic quickly and accurately, and excellent industrial practicability is realized.

    参考文献
    相似文献
    引证文献
引用本文

高向军,王 倩,洪留荣,等. 阵列工业图像快速拼接算法[J]. 科学技术与工程, 2019, 19(22): 249-253.
高向军,wangqian,,et al. Fast Mosaic Algorithms for Array Industrial Images[J]. Science Technology and Engineering,2019,19(22):249-253.

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2019-01-04
  • 最后修改日期:2019-03-06
  • 录用日期:2019-03-22
  • 在线发布日期: 2019-08-28
  • 出版日期:
×
律回春渐,新元肇启|《科学技术与工程》编辑部恭祝新岁!
亟待确认版面费归属稿件,敬请作者关注