面向矿井复杂环境和机动应用的无线视频编码技术研究

项目名称： 面向矿井复杂环境和机动应用的无线视频编码技术研究

项目编号： No.51204175

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 冶金与矿业学科

项目作者： 程德强

作者单位： 中国矿业大学

项目金额： 25万元

中文摘要： 矿井应急救援可视化得到国家和行业的重视，但面向巷道应用的无线视频编码技术还未有效的解决。本项目借鉴人眼双目机制对图像的增强和容错性能，研究面向矿井复杂环境和机动应用的无线视频编码方法。内容包括：矿井狭长巷道环境下，双目视点的对极线几何约束及其相关性估计方法；融合时间边信息和空间边信息的高精度解码边信息重建方法；根据图像内容相关性自适应修正参数的虚拟依赖信道模型。通过研究，拟构建基于双目视点的分布式视频编码框架，提出基于双目视点的边信息重建方法和虚拟依赖信道模型，构建适合矿井巷道特点和物理场的高容错性无线视频编码基础理论。本项目是安全检测与监控、计算机视觉和多媒体信息处理与编码的有机交叉，紧密联系国家煤炭工业的规划需求和技术发展现状，其令人振奋的研究成果不但可进一步丰富分布式视频编码理论，也可为矿井应急保障装置研制提供依据，因此，具有重要理论意义和实际应用价值。

中文关键词： 分布式视频编码；矿井；边信息生成；关键帧提取；无线传感器网络

英文摘要： Researches on the technology of underground emergent visualizing rescue are paid more attentions by national and mining authorities. Techniques for the wireless video coding of coal mine are still not solved effectively yet. In light of the capabilities of enhancement and error tolerance for images of human eyes, the techniques of wireless video coding for underground conditions and mobile deployment are researched here including the following contents. The epipolar geometry constraints and correlation estimation methods of binocular points of view are first developed. Then, side information reconstructing algorithms are studied by syncretizing the temporal and spatial side information. Finally, a virtual dependency channel is modeled, which can adjust its parameters according to the content correlation between frames. With the studies, we will establish a distributed video coding framework based on binocular points of view, and propose novel side information reconstructing methods for underground tunnel as well as a virtual dependency channel model. This project is a novel challenge under interdisciplinary efforts such as mine safety and monitoring, computer vision, multimedia processing and coding, and meets the requirements of national mining industries. The corresponding researching achievements can not only

英文关键词： Distributed video coding；Coal mine；Side information generation；Key frame extraction；wireless sensor network