Edge adaptive image resolution enhancement in video sensor network
Edge adaptive image resolution enhancement in video sensor network
复制标题
视频传感器网络中的边缘自适应图像分辨率增强
DOI：
10.1016/j.optcom.2011.06.027
复制
发表时间：
2011-09-01
影响因子：
2.4
通讯作者：
Wei, Gang
中科院分区：
文献类型：
--
作者：
Cai, Kang;Shi, Jinglun;Wei, Gang
视频传感器网络在传输图像时，由于传输带宽的限制，通常使用分辨率较低的图像。利用图像分辨率增强技术可以将低分辨率图像增强为高分辨率图像。图像分辨率增强的关键是保持图像的边缘结构。本文提出了一种新的图像分辨率增强方法，即利用相邻像素的双边加权平均来估计未知像素的亮度。更具体地，具有较近距离的相邻像素具有较大贡献。此外，属于变化较小方向的相邻像素的贡献较大。实验结果表明，该方法优于几种传统的边缘导向图像插值算法。此外，所提出的方法产生低的计算复杂度，它是潜在的实时实现。(C)2011 Elsevier B.V.保留所有权利。
Video sensor network usually uses fairly low-resolution images due to the limited transmission bandwidth in transmitting images. It is potential to enhance the captured low-resolution images using image resolution enhancement technique that is able to produce a high-resolution image from its low-resolution counterpart. The key challenge of image resolution enhancement is to preserve the edge structure in images. In this paper, a new image resolution enhancement approach is proposed to estimate the intensity of the unknown pixel using a bilateral weighted average of that of its neighboring pixels. More specifically, the neighboring pixels with nearer distance have larger contributions. Furthermore, the neighboring pixels belonging to direction with smaller variation have larger contributions. Experimental results are provided to show that the proposed approach outperforms several conventional edge-directed image interpolation algorithms. Furthermore, the proposed approach yields low computational complexity; it is potential for real-time implementation. (C) 2011 Elsevier B.V. All rights reserved.