基于超像素分割和图像配准的深度图像修复方法  

The depth map restoration based on superpixel segmentation and image registration

作　　者：杨飞 王俊元[1] 段能全[1] Yang Fei;Wang Junyuan;Duan Nengquan(School of Mechanical Engineering,North University of China,Shanxi Taiyuan,030051,China)

机构地区：[1]中北大学机械工程学院

出　　处：《机械设计与制造工程》2020年第1期20-24,共5页Machine Design and Manufacturing Engineering

基　　金：山西省自然科学基金面上项目(201801D121187)

摘　　要：针对Kinect V2深度相机采集的深度图像存在深度信息丢失(孔洞)问题,提出了一种融合超像素分割RGB图像和深度图像配准的深度图像修复方法。首先对RGB图像进行超像素分割;然后将分割后的RGB图像与原始深度图像配准;接着将孔洞像素按照所在区域划分为超像素区域内的孔洞像素和超像素边界上的孔洞像素两大类,并对不同类型的孔洞像素采用不同的方法完成深度图像的修复。最后进行了实验验证,结果表明,该方法与双边滤波算法、联合双边滤波算法相比,能够有效地修复深度图像中的孔洞,获得较高质量的深度图像。Aiming at the loss of depth information(holes)in depth images collected by Kinect V2 depth camera,it applies the fusion of RGB image and depth image registration to present a new method of depth image repair.Firstly,the RGB image is segmented by super pixels.The segmented RGB image is registered with the original depth image.Then the hole pixel is divided into two categories according to the region of the hole pixel and the hole pixel on the boundary of the superpixel,the different methods are adopted to repair the depth image for different types of hole pixel.Finally,the experimental results show that the proposed method can effectively repair the holes in depth images and obtain higher quality depth images compared with bilateral filtering algorithm and joint bilateral filtering algorithm.

关 键 词：KINECT V2 深度图像 孔洞修复 配准 超像素分割 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]