摘要
针对Kinect获取的深度图像存在重影、空洞且与彩色图像存在偏差等问题,提出一种基于彩色图像引导的深度图像修复算法.采用归一化相关系数匹配方法重对齐彩色图像;利用改造的中值滤波方法消除孤立空洞点;利用重对齐的彩色图像引导大面积空洞点的修复.实验结果表明,该算法能够有效去除深度图像中的空洞和重影现象,具有较好的修复效果.
Aiming at the problems that depth image acquired by Kinect has the problems of ghosting and holes,and deviations from color images,a depth image restoration algorithm based on color image guidance is proposed.Firstly,the normalization correlation coefficient matching method is used to realign the color image;then the modified median filtering method is used to eliminate the isolated hole points;finally,the realigned color image is used to guide the recovery of large-area hole points.The experimental results show that the proposed algorithm can effectively remove holes and ghosting in the depth image,and has better restoring effects.
作者
毛雁明
MAO Yan-ming(College of Information and Electrical Engineering,Ningde Normal University,Ningde,Fujian 352100,China)
出处
《宁德师范学院学报（自然科学版）》
2019年第4期364-368,共5页
Journal of Ningde Normal University(Natural Science)
基金
福建省自然科学基金项目（2017J01774）
福建省科技厅自然基金项目（2016J01674）
宁德师范学院青年科技项目（2015Q04）
宁德市指导性科技计划项目（20160214）
