应用光学, 2016, 37 (2): 203, 网络出版: 2016-04-12   

基于结构特征引导滤波的深度图像增强算法研究

Depth image enhancement algorithm based on structure feature guidance filter
作者单位
西安应用光学研究所, 陕西 西安 710065
摘要
为了解决深度图像中存在的图像模糊、空洞和噪声等图像质量问题, 拟从软件的角度出发, 在不改变传感器成像系统物理结构的前提下, 基于结构特征并以彩色图像作为引导展开研究, 实现深度图像增强和空洞修补的目的, 提高深度图像的质量。通过对彩色图像和深度图像的结构特征进行提取, 得到共性的全局特征, 并对得到的结构特征进行联合双边滤波, 最后基于马尔科夫随机场的方法进行深度图像增强, 实现了低成本获取深度增强的图像。实验结果表明本文算法在保持图像边缘的细节性、平滑性和整体性上具有更好的效果, 与其他算法相比, 图像的均方根误差RMSE更低, 仅为0.506 93及1.169 30(针对Teddy及Art图像) 。
Abstract
There are many quality problems in depth image, for example, images may be combined with blurring, empty holes and noise. In order to solve these problems, from the view point of software, an algorithm guided by color images based on structural characteristics was studied without changing any physical structure of the sensor imaging system, which could realize depth image enhancement and empty holes repairing, as well as improve the quality of the depth image. Through the extraction for the structure of the color image and depth image, the common global features were obtained. After joint bilateral filtering for the structural features, the depth image was enhanced based on Markov random filed (MRF), and the depth enhanced image was finally obtained with low cost. Experimental results show that the algorithm has a better effect in keeping the detail, smoothness and integrity of image edge; moreover, the root-mean-square error (RMSE) of this algorithm is smaller than other algorithms, which is 0.50693 and 1.16930 for image Teddy and Art, respectively.

钱钧, 李良福, 周锋飞, 王超, 邹彬. 基于结构特征引导滤波的深度图像增强算法研究[J]. 应用光学, 2016, 37(2): 203. Qian Jun, Li Liangfu, Zhou Fengfei, Wang Chao, Zou Bin. Depth image enhancement algorithm based on structure feature guidance filter[J]. Journal of Applied Optics, 2016, 37(2): 203.

本文已被 1 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!