《计算机应用研究》|Application Research of Computers

基于超像素的深度图修复算法

Depth image inpainting algorithm based on superpixel

免费全文下载 (已被下载 次)  
获取PDF全文
作者 潘波,范祺红,曹雪玮,刘骥
机构 1.重庆大学 计算机学院,重庆 400044;2.国网天津城南供电分公司,天津 300201
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2020)06-054-1863-04
DOI 10.19734/j.issn.1001-3695.2018.11.0913
摘要 随着计算机视觉领域不断的发展,用于描述场景深度信息的深度图受到越来越多的关注。针对深度图中由于深度信息缺失导致的图像空洞、深度值准确度不高及图像噪声等问题,提出了一种融合超像素和基于方向的联合双边滤波器的深度图修复算法来改善深度图的质量。该算法引入了基于超像素的自适应滤波窗口,并对不同类型的空洞像素点采用不同的滤波算法,从而对深度图进行修复和优化。定性对比实验和定量评价结果表明,其可以有效地修复深度图空洞噪点,获得高质量、高准确性的深度图。
关键词 深度图; 图像修复; 超像素算法; 基于方向的联合双边滤波
基金项目 国家自然科学基金资助项目(61502060)
本文URL http://www.arocmag.com/article/01-2020-06-054.html
英文标题 Depth image inpainting algorithm based on superpixel
作者英文名 Pan Bo, Fan Qihong, Cao Xuewei, Liu Ji
机构英文名 1.College of Computer Science,Chongqing University,Chongqing 400044,China;2.State Grid Tianjin Power Chengnan District Supply Company,Tianjin 300201,China
英文摘要 Depth images used to describe the depth information have attracted tremendous attention due to the continuous development of computer vision. In order to solve the problems caused by the lack of depth information, such as holes, inaccuracy and noises, this paper proposed a depth image inpainting algorithm which combined superpixel with directional joint bilate-ral filter. The method introduced an adaptive filtering window based on superpixel algorithm, and employed different filtering strategies on different types of holes to repair and optimize depth map. Qualitative experiments and quantitative evaluation demonstrate that the proposed method is capable of repairing holes and noises effectively and obtaining depth image with high quality and accuracy.
英文关键词 depth image; image inpainting; superpixel; directional joint bilateral filter
参考文献 查看稿件参考文献
 
收稿日期 2018/11/29
修回日期 2019/1/31
页码 1863-1866,1870
中图分类号 TP391.41
文献标志码 A