一种由明暗恢复形状的改进变分算法

杨磊; 韩九强

您当前的位置：

首页 >

文章列表页 >

一种由明暗恢复形状的改进变分算法

专题研究 | 更新时间：2025-12-08

- 一种由明暗恢复形状的改进变分算法
- Improved Variational Algorithm of Shape-from-Shading
- 西安交通大学学报 2008年42卷第4期页码：418-422
- 作者机构：
  
  西安交通大学电子与信息工程学院,西安,710049
- 作者简介：
- 基金信息：
  
  : 2007-08-08. 作者简介: 杨磊(1976-),男,博士生;韩九强(联系人),男,教授,博士生导师. 基金项目: 国家自然科学基金资助项目(60502021);教育部高等学校博士点专项科研基金资助项目(20050698025)
- DOI：
  中图分类号： TP391.4
- 网络首发：2008-04-10，
  
  纸质出版：2008
- 稿件说明：
移动端阅览
杨磊, 韩九强. 一种由明暗恢复形状的改进变分算法[J]. 西安交通大学学报, 2008,42(4):418-422.

杨磊, 韩九强. Improved Variational Algorithm of Shape-from-Shading[J]. 2008, 42(4): 418-422.
杨磊, 韩九强. 一种由明暗恢复形状的改进变分算法[J]. 西安交通大学学报, 2008,42(4):418-422. DOI：

杨磊, 韩九强. Improved Variational Algorithm of Shape-from-Shading[J]. 2008, 42(4): 418-422. DOI：

摘要

针对变分算法可能丢失表面细节信息的问题

提出了一种由明暗恢复形状的改进变分算法

使用接近实际的混合反射模型描述的反射图方程和图像梯度方程构造了新的目标泛函

由于图像梯度反映了二维图像的细节信息

新目标泛函比仅使用反射图方程的目标泛函蕴涵了更加丰富的三维形状细节信息.然后

使用变分原理求解该目标泛函

得到由明暗恢复形状的改进变分算法.所提出的算法比基本变分算法三维形状恢复的精度有显著提高.合成半球图像的实验表明

新算法在相同迭代次数时恢复三维形状的平均相对误差比基本变分算法减小了13.45%.实际复杂图像的三维形状恢复结果表明

新算法在恢复表面三维细节时比基本变分算法更加准确.

Abstract

An improved variational approach of shape-from-shading is proposed. A new objective functional is formulated using a reflectance map equation and an image gradient equation. The reflectance map equation is described by a hybrid reflectance model which is prone to reality. Since the gradients of images exhibit the detailed information of two-dimensional image

the new objective functional contains more information of three-dimensional surface than the objective functional formulated by only reflectance map equation. Then

an improved algorithm of shape-from-shading is proposed by using the variational principle. The accuracy of the proposed algorithm is improved evidently compared with the basic variational algorithm

and experiments on synthetic images of hemisphere show that the relative average error of the proposed algorithm is reduced by 13.45% under the same iteration times. Reconstruction results from real complicated images show that the proposed algorithm is more accurate than the basic variational algorithm in dealing with the details of the surface.

关键词

Keywords

references

SONG Limei, QU Xinghua, XU Kexin, et al. Novel SFS-NDT in the field of defect detection [J]. International NDT & E, 2005, 38(5): 381-386.

BORS A G, HANCOCK E R, WILSON R C. Terrain analysis using radar shape-from-shading[J]. IEEE Trans Pattern Analysis and Machine Intelligence, 2003, 25(8): 974-992.

DU Quanying, CHEN Shanben, LIN Tao. Inspection of weld shape based on the shape from shading [J]. Int J Adv Manuf Technol, 2006, 27(7): 667-671.

HORN B K P. Height and gradient from shading [J]. Int J Computer Vision, 1990, 5(1): 37-75.

ZHANG R, TSAI P S, CRYER J E, et al. Shape from shading: a survey [J]. IEEE Trans Pattern Analysis and Machine Intelligence, 1999, 21(8): 690-706.

HORN B K P, BROOKS M J. The variational approach to shape from shading[J]. Computer Vision Graphics Image Process, 1986, 33(2): 174-208.

LEE K M,KUO C C J. Shape from shading with a linear triangular element surface model [J]. IEEE Trans Pattern Analysis and Machine Intelligence, 1993, 15(8): 815-822.

CHO S Y, CHOW T W S. A new color 3D SFS methodology using neural-based color reflectance models and iterative recursive method [J]. Neural Computation, 2002, 14(11): 2751-2789.

RON K, JAMES A S. Optical Algorithm for shape from shading and path planning [J]. Journal of Mathematical Imaging and Vision, 2001, 14(3): 237-244.

PRADOS E, CAMILLI F, FAUGERAS O. A unifying and rigorous shape from shading method adapted to realistic data and applications [J]. J Math Imaging, 2006, 25(3): 307-328.

ZHENG Q, CHELLAPPA R. Estimation of illuminant direction, albedo, and shape from shading[J]. IEEE Trans Pattern Analysis and Machine Intelligence, 1991, 13(7): 680-702.

ZHANG R, MUBARAK S. Shape from intensity gradient [J]. IEEE Trans on Systems, Man and Cybernetics: A: Systems and Humans, 1999, 29(3): 318-325.

TAGARE H D, DEFIGUEYREDO R J P. A theory of photometric stereo for a class of diffuse non-Lambertian surfaces [J]. IEEE Trans Pattern Analysis and Machine Intelligence, 1991, 13(2): 133-152.

LIN C T, Cheng W C, LIANG S F. Neural-network-based adaptive hybrid-reflectance model for 3-D surface reconstruction[J]. IEEE Trans Neural Networks, 2005, 16(6): 1601-1615.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

电子级超纯氮深冷空分流程节能优化研究

采用机器学习的大规模并行程序I/O性能建模与预测方法

多工况下离心压气机叶轮-机匣一体联合设计优化

水蒸气掺混对氨裂解气旋流燃烧与排放特性的影响研究

电动车辆间接热系统快速除湿响应迟滞机理及优化控制策略