摘 要无人机凭借着低廉成本、容易便携等优势在军事和民用领域受到青睐。在民用领域的测绘行业之中，相较于传统的手段，无人机航空测量无疑是目前获取地理信息的最佳的技术手段。因为无人机无需依赖机场，成本更是低廉，拍摄周期也极短。79769

  在如今的救灾领域中无人机同样扮演着必不可少的角色，比如投放物资，灾区情况勘察；为无人机配备灭火器，通过声波工作，进入危险的火灾地区进行灭火工作。在福岛核泄漏之后，利用T-鹰无人机以及全球鹰无人机进行灾后的信息收集工作。

  无人机自主导航技术的关键在于具备高精度性，作为无人机的灵魂——UAS，即无人驾驶系统，俗称导航技术是目前无人机研究的重中之重。

  结合无人机自主控制对导航系统高精度和高可靠性的迫切需求，本文主要展开基于图像处理，多特征融合匹配的无人机关键技术研究：UVA的图像处理技术，多种特征信息融合匹配，UAV自主导航规划路径的方案。

  本文主要研究无人机可用特征选取、高效的特征提取、存储、识别匹配算法，力求解决在无人机的机载设备有限的计算处理能力的局限下实现图像处理，特征匹配的关键技术问题，减少无人机导航系统对于GPS系统 的依赖。

毕业论文关键词：无人机；自主导航；鲁棒性，多特征融合，图像处理，显著图像特征提取。

Abstract  UAV by virtue of low cost, casualties, easy to portable and other advantages in the military and civilian areas are favored。 In the field of civilian surveying and mapping industry, compared to the traditional means, UAV aerial survey is undoubtedly the best access to geographic information technology。 Because UAVs do not have to rely on the airport, the cost is low, shooting cycle is also very short。

  In the disaster relief industry, UAVs also play an important role, such as delivery of materials, disaster situation investigation; for the unmanned aerial vehicles equipped with fire extinguishers, through sonic work, into the dangerous areas of fire fighting。 After the Fukushima nuclear leak, the use of T-Eagle UAV and the Global Eagle UAV for post-disaster information collection。

  UAV to achieve the key to self-flight is that it must have high-precision autonomous navigation capabilities, as the soul of UAV - UAS, that is, unmanned system, commonly known as navigation technology is the most important research on UAVs。

  This paper mainly studies the key technologies of UAV based on image processing and multi-feature fusion matching, and puts forward the application of UAV-based image processing, multi-feature Fusion matching, path planning and navigation of the program and algorithm for UAV autonomous navigation system research and application to provide a theoretical basis。

  This paper mainly studies the feature selection and efficient feature extraction, storage and identification matching algorithms of UAVs, and tries to solve the key technical problems of image processing and feature matching under the limited computing power of UAV's onboard equipment。 Effectively overcome the UAV autonomous navigation system on the GPS dependence。

  Key words: unmanned aerial vehicle; autonomous navigation; robustness, multi-feature fusion, image processing, and significant image feature extraction。

Contents

2017年 5 月 3 日 1

摘 要 2

Abstract 3

注释表 4

符号/缩略词 4