计算机工程 ›› 2019, Vol. 45 ›› Issue (2): 233-239,244.doi: 10.19678/j.issn.1000-3428.0049958

• 图形图像处理 • 上一篇    下一篇

基于图像区域像素重构的聚焦形貌恢复

赵洪盛1,丁华1,刘建成1,2   

  1. 1.太原理工大学 机械工程学院,太原 030024; 2.太平洋大学 工程与计算机科学学院,美国 斯托克顿 95211
  • 收稿日期:2018-01-03 出版日期:2019-02-15 发布日期:2019-02-15
  • 作者简介:赵洪盛(1991—),男,硕士,主研方向为图像处理;丁华,副教授、博士;刘建成,教授、博士。
  • 基金项目:

    山西省自然科学基金(201601D011050);山西省煤机重点科技攻关项目(MJ2014-05-02);山西省研究生联合培养基地人才培养项目(2016JD13)。

Shape from Focus Based on Image Regional Pixel Reconstruction

ZHAO Hongsheng 1,DING Hua 1,LIU Jiancheng 1,2   

  1. 1.School of Mechanical Engineering,Taiyuan University of Technology,Taiyuan 030024,China; 2.School of Engineering and Computer Science,University of the Pacific,Stockton 95211,USA
  • Received:2018-01-03 Online:2019-02-15 Published:2019-02-15

摘要:

为提高三维形貌图像的恢复精度,提出一种基于最大最小算子与区域像素重构的聚焦形貌恢复方法。利用最大最小算子判定图像序列中的噪声像素点,并采用中值滤波方法消除图像噪声。对图像序列进行区域像素重构,根据重构图像相邻像素的灰度差值设计聚焦评价函数,确定区域聚焦等级,实现聚焦形貌恢复。实验结果表明,该方法恢复的三维形貌图像精度优于传统形貌恢复方法,并对含噪图像具有较好的鲁棒性。

关键词: 聚焦形貌恢复, 图像噪声, 最大最小算子, 区域像素重构, 聚焦评价函数

Abstract:

In order to improve the accuracy of 3D morphology image recovery,a novel method for Shape from Focus(SFF) based on max-min operator and regional pixel reconstruction is proposed.The max-min operator is used to determine the noise pixels in the image sequence,and the median filtering is used to the noise points,so as to eliminate the image noise.It performs the regional pixel reconstruction on the image sequence.According to the gray difference of the adjacent pixels in the reconstructed image,the focus measure is designed to determine the focus level of the region to realize SFF.Experimental results show that the accuracy of the three-dimensional topography restored by this method is superior than that of the traditional one,and it has better robustness to noisy images.

Key words: Shape from Focus(SFF), image noise, max-min operator, regional pixel reconstruction, focus evaluation function

中图分类号: