摘要
针对TOF(Time of Flight)相机深度图像在超分辨重建过程中易出现边缘模糊、纹理映射问题,在联合双边上采样滤波器的基础上提出一种基于深度图像自身边缘特征引导的超分辨重建方法。通过低分辨深度图像的边缘特征引导,将深度图像分为不同的区域,根据滤波区域性质的不同,对联合双边上采样滤波器模型中的颜色相似项进行不同加权。同时为了进一步保持图像边缘,在深度图像边缘部分加入一个结构保持项。最后利用联合双边上采样滤波器模型重建出高分辨深度图像。实验结果表明,该方法不仅提高了TOF深度图像的分辨率,而且很好地保护了深度图像的边缘结构,取得了较好的效果。
It is prone to edge blurring and texture copying in the depth map from TOF camera super-resolution process,a new super-resolution reconstruction method based on the depth map ＇s edge feature-guided is proposed.According to the edge feature of low resolution depth map,the depth map is divided into different regions and then the different color similarity weighting is calculated along with the different filtering regions. In order to protect the image edge further,a structure preserved term is added in the edge region. Finally,the high resolution depth map is calculated by using the joint bilateral up-sampling filter model. The experimental results show that the proposed method not only improves the resolution of depth map from TOF camera,but also protects the image edge structure and achieves a well result.
出处
《计算机应用与软件》
2017年第2期220-225,共6页
Computer Applications and Software
基金
国家自然科学基金项目(614031160
61273237)
中国博士后基金项目(2014M560507)
关键词
超分辨重建
深度图像
联合双边上采样滤波器
边缘特征
Super-resolution reconstruction
Depth map
The joint bilateral up-sampling filter
Edge feature
