基于目标检测与跟踪算法的边坡危岩监测系统设计.pdfVIP

中文摘要 摘 要 我国是一个地质灾害频发的国家，边坡危岩崩塌引发的灾害通常会带来巨大 的损失，因此如何在野外复杂环境下实现长期边坡危岩监测成为一项重点研究内 容。本文提出一种边坡危岩监测系统设计方案，采用图像检测法监测边坡危岩区 域，及时获取危岩的状态信息，并研制了一套边坡危岩监测系统设备，实现对边 坡危岩的高效、自动化监测。 本文的主要研究内容如下： ① 针对运动危岩的检测问题，提出基于自适应混合色彩空间信息融合的运动 目标检测算法。该算法能够提取完整的目标轮廓，对目标运动速度的适应能力较 强，检测性能突出。 ② 针对边坡危岩崩塌过程中的跟踪问题，提出基于改进Camshift 的运动目标 跟踪算法。该算法能够及时获取运动危岩状态信息，实现在复杂背景中对危岩的 可靠跟踪，具有良好的稳定性。 ③ 开展边坡危岩监测系统方案设计。在野外真实边坡环境下进行多场景测试， 实验表明本系统能够准确地提取出运动危岩并对危岩崩塌过程进行持续跟踪，达 到系统设计需求，验证了系统的可行性和有效性。 本文开展了边坡危岩监测系统设计方案研究，实现对边坡危岩的自动化检测 跟踪，在工程应用方面具有一定的实用价值。 关键词：边坡危岩；自动化监测；目标检测；目标跟踪 I 英文摘要 Abstract China is a country with frequent geological disasters, the disaster caused by collapse of slope dangerous rock usually bring huge losses. Therefore, how to realize long-term slope dangerous rock monitoring in complex field environment has become a key research content. In this thesis, a design scheme of slope dangerous rock monitoring system is proposed. The image detection method is used to monitor the dangerous area and the state information of dangerous rock is obtained in time. At the same time, a set of slope dangerous rock monitoring system equipment is developed to realize efficient and automatic monitoring of slope dangerous rock. The main research contents of this thesis are as follows: ① In order to solve the detection problem of dangerous rock, a moving target detection algorithm based on adaptive information merging in hybrid color space is proposed. The algorithm can extract the complete target contour, has strong adaptability to the change of the target moving speed, and has outstanding detection performance. ② In order to solve the tracking problem during the collapse process of dangerous