摘要
为解决高光弱纹理物体表面重建时,存在的孔洞和噪声,及信息缺失等问题,提出了光照补偿结合深度图像的重建方法.结合光照方向和光照强度的参数估计,确定高光区域并均匀光照,再以激光点的变化轨迹修正均值漂移算法的漂移区域,建立改进的均值漂移中心描述子,对深度图像的噪声和孔洞进行判定并修复,实现物体表面重建.结果表明,该方法既可以保持不同种类对象重建完整,避免信息缺失,还可减少外界环境和对象自身特征的负面影响.通过高光弱纹理标准图片和实拍物体重建实验,以及使用均方根误差、峰值信噪比和结构相似性等性能指标进行评价,验证了该方法的鲁棒性和有效性.
In order to solve the problems of surface reconstruction and high-light weak texture objects,such as holes and noise,and information loss,a reconstruction method of illumination compensation combined with depth image was proposed.Combined with the parameter estimation of illumination direction and illumination intensity,the highlight area was determined and uniform illumination was applied.Then the drift region of the mean shift algorithm was corrected by the change track of the laser point,and an improved mean shift center descriptor was established,and the noise and holes of the depth image were determined and repaired.Finally,the surface reconstruction of the object was achieved.The results show that the proposed method can maintain the complete reconstruction of different kinds of objects,avoid the lack of information and reduce the negative impact of the external environment and the characteristics of the object itself.The robustness and effectiveness of the proposed method are verified by the high-light weak texture standard picture and real-time object reconstruction experiments,and the performance indexes such as root mean square error,peak signal-to-noise ratio and structural similarity.
作者
乔玉晶
张思远
赵宇航
QIAO Yu-jing;ZHANG Si-yuan;ZHAO Yu-hang(Key Laboratory of Advanced Manufacturing and Intelligent Technology,Ministry of Education,Harbin 150080,China;Institute of Mechanical&Power Engineering,Harbin University of Science and Technology,Harbin 150080,China)
出处
《光子学报》
EI
CAS
CSCD
北大核心
2019年第12期166-176,共11页
Acta Photonica Sinica
基金
国家自然科学基金(No.51675142)~~
关键词
机器视觉
表面重建
光照估计
均值漂移
深度图处理
高光现象
弱纹理
Machine vision
Surface reconstruction
Illumination estimation
Mean shift
Depth map processing
Highlight phenomenon
Weak texture
