一种适用于多媒体传感器网络的图像拼接算法

为充分利用有限的资源产生高分辨率、宽视角图像,考虑到相邻无线视频节点之间的信息冗余性,提出一种适用于无线多媒体传感器网络的图像拼接算法。使用分块搜索算法进行图像配准以降低能耗,改进绝对差值和算法以提高图像配准的精度,并使用渐进渐出的加权平均算法对图像进行缝合,图像拼接之后与基站间的通信量减少,可以有效降低网络负载。仿真实验结果表明,所提出的算法在保证一定图像配准精度和图像质量的情况下,计算复杂度较低,可以有效节约能量。

Image mosaicking algorithm for wireless multimedia sensor networks

For sufficient utilization of constraint resource on generating high-resolution images with wide field of view, this paper proposed an algorithm took fully account the redundant visual information among multiple video sensors. It applied the image block searching algorithm for image registration to decrease the energy consumption, adapted sum of absolute difference algorithm to improve the accuracy of image registration, and applied weighted mean algorithm for image stitching. It decreased data volumes of transmission after image mosaicking, which could efficiently reduce the network loading. Simulation results demonstrate that the computation complexity of proposed algorithm is lower than other image mosaicking algorithm, under certain image registration accuracy and image quality.