结合SFS和双目模型的单幅图像深度估计算法
|
摘要
从二维图像中获取图像深度信息是计算机视觉领域的热点问题。经典的双目视觉方法需要相机内部参数和同一场景的多张图像,视觉参数不足容易导致计算错误,而单幅图像只能依靠自身的几何信息得到图像深度。文中针对未知相机参数得到的单幅普通二维图像,结合SFS(从阴影恢复形状)方法,运用图像明暗的几何信息和双目视觉模型获取图像目标深度值,再利用目标轮廓信息对不同目标区域进行赋值,得到图像不同目标距离观察者的远近关系。实验结果表明,所提方法得到的图像深度能较为准确地反映出场景的真实信息,较符合实际观测结果。
Obtaining depth information from 2 D images is a hot topic in the field of computer vision.Classical binocular vision methods require camera parameters and multiple images of the same scene.Insufficient visual parameters can easily lead to errors in calculation,while a single image can only rely on its own geometric information to get the image depth.This paper used the geometric information of the image and the binocular vision model to get the depth value of the object in a single ordinary two-dimensional image with unknown camera parameters.The experimental results show that the image depth obtained by the proposed method can relatively accurately reflect the real information of the scene,which is consistent with the actual observation results.
Obtaining depth information from 2 D images is a hot topic in the field of computer vision.Classical binocular vision methods require camera parameters and multiple images of the same scene.Insufficient visual parameters can easily lead to errors in calculation,while a single image can only rely on its own geometric information to get the image depth.This paper used the geometric information of the image and the binocular vision model to get the depth value of the object in a single ordinary two-dimensional image with unknown camera parameters.The experimental results show that the image depth obtained by the proposed method can relatively accurately reflect the real information of the scene,which is consistent with the actual observation results.
引文
[1] 王军华,李丁,刘盛鹏.基于改进RANSAC的消防机器人双目障碍检测[J].计算机工程与应用,2017,53(2):236-240.
[2] NING X,LI W,TANG B,et al.BULDP:Biomimetic Uncorrela- ted Locality Discriminant Projection for Feature Extraction in Face Recognition[J].IEEE Transactions on Image Processing,2018,27(5):1-1.
[3] JOANNIDIS C.Binocular Vision in the Total Situation[J].American Journal of Psycho-analysis,2018,78(1):1-15.
[4] 杨勇,闫钧华,井庆丰.融合图像显著性与特征点匹配的形变目标跟踪[J].中国图象图形学报,2018,23(3):86-100.
[5] 杜娟,梁睿,胡跃明,等.一种提高测量范围的双目视觉测距方法:中国,103776419 B[P].2016.
[6] 汪香香,张成,沈川,等.三维复杂场景的菲涅耳全息图频域压缩重建[J].光学学报,2017,37(7):75-82.
[7] NAH S,KIM T H,LEE K M.Deep Multi-scale Convolutional Neural Network for Dynamic Scene Deblurring[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).IEEE,2017.
[8] ANDRADE J.Defocus Map Detection Using a Single Image[C]// International Conference on Computational Science and Computational Intelligence.IEEE,2017:777-780.
[9] HORN B P.Shape from Shading:A method of obtaining the shape of a smooth opaque object from one view[D].Massachusetts:Massachusetts Institute of Technology,1970.
[10] 连晔.基于单目视频的高质量人脸动画生成[D].杭州:浙江大学,2017.
[11] 杨志明,赵红东.从阴影恢复形状的径向基函数反射模型研究[J].中国图象图形学报,2017,22(11):1565-1573.
[12] 曹芳,朱永康.基于SFS方法的三维重构及精度分析[J].计算机科学,2017,44(S1):244-247.
[13] 杨孝光,李运明,张虎军.基于松弛迭代方法的线性SFS算法分析[J].计算机与现代化,2013(1):40-44.
[14] JIANG L,ZHANG J,DENG B,et al.3D Face Reconstruction with Geometry Details from a Single Image[J].IEEE Transactions on Image Processing,2018,PP(99):1-1.
[15] ZHENG Q,CHELLAPA R.Estimation of Illuminant Direction,Albedo,and Shape from Shading[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,1991,13(7):680-702.
[16] SWIERCZYNSKI P,PAPIE,et al.A level-set approach to joint image segmentation and registration with application to CT lung imaging[J].Computerized Medical Imaging & Graphics,2018,65:58-68.
[17] 明德学院[EB/OL].http://vision.middlebury.edu.
[2] NING X,LI W,TANG B,et al.BULDP:Biomimetic Uncorrela- ted Locality Discriminant Projection for Feature Extraction in Face Recognition[J].IEEE Transactions on Image Processing,2018,27(5):1-1.
[3] JOANNIDIS C.Binocular Vision in the Total Situation[J].American Journal of Psycho-analysis,2018,78(1):1-15.
[4] 杨勇,闫钧华,井庆丰.融合图像显著性与特征点匹配的形变目标跟踪[J].中国图象图形学报,2018,23(3):86-100.
[5] 杜娟,梁睿,胡跃明,等.一种提高测量范围的双目视觉测距方法:中国,103776419 B[P].2016.
[6] 汪香香,张成,沈川,等.三维复杂场景的菲涅耳全息图频域压缩重建[J].光学学报,2017,37(7):75-82.
[7] NAH S,KIM T H,LEE K M.Deep Multi-scale Convolutional Neural Network for Dynamic Scene Deblurring[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).IEEE,2017.
[8] ANDRADE J.Defocus Map Detection Using a Single Image[C]// International Conference on Computational Science and Computational Intelligence.IEEE,2017:777-780.
[9] HORN B P.Shape from Shading:A method of obtaining the shape of a smooth opaque object from one view[D].Massachusetts:Massachusetts Institute of Technology,1970.
[10] 连晔.基于单目视频的高质量人脸动画生成[D].杭州:浙江大学,2017.
[11] 杨志明,赵红东.从阴影恢复形状的径向基函数反射模型研究[J].中国图象图形学报,2017,22(11):1565-1573.
[12] 曹芳,朱永康.基于SFS方法的三维重构及精度分析[J].计算机科学,2017,44(S1):244-247.
[13] 杨孝光,李运明,张虎军.基于松弛迭代方法的线性SFS算法分析[J].计算机与现代化,2013(1):40-44.
[14] JIANG L,ZHANG J,DENG B,et al.3D Face Reconstruction with Geometry Details from a Single Image[J].IEEE Transactions on Image Processing,2018,PP(99):1-1.
[15] ZHENG Q,CHELLAPA R.Estimation of Illuminant Direction,Albedo,and Shape from Shading[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,1991,13(7):680-702.
[16] SWIERCZYNSKI P,PAPIE,et al.A level-set approach to joint image segmentation and registration with application to CT lung imaging[J].Computerized Medical Imaging & Graphics,2018,65:58-68.
[17] 明德学院[EB/OL].http://vision.middlebury.edu.