基于小型无人直升机的环境污染区域边界检测与跟踪算法的研究

发布时间：2018-05-03 07:09

本文选题：无人机 + 跟踪　；参考：《华南理工大学》2016年硕士论文

【摘要】：污染羽流是一类最为典型的环境污染表征形式:气浮颗粒(烟团、PM2.5等)放射泄漏污染、化学气体扩散、水体污染(油体泄漏、赤潮等),表现为空间上的区域性和污染区域内的不同浓度/强度分布,同时在自身扩散和环境(风、水流、地理环境、温度、湿度变化等)的作用下区域边界形体随时间变化,呈现一种不确定的类似羽毛边界的污染羽流的变化形态。对该类区域边界(或等浓度区域边界)的快速自主辨识、估计与预测是环境与应急监测中的关键环节,是预警、监测控制与快速响应的重要依据。为了减少事故造成的损害,本论文提出利用单架次直升机对环境进行监测,若出现环境污染,将对污染区域边界进行检测和跟踪。本文主要研究了四个方面内容:首先,构建一个环境模型来模拟真实环境中的油污扩散情形,对模型进行了详细的说明并作了仿真分析;其次,由于采用单架次无人直升机来获取的环境信息量非常少,使得为了得到局部的完整信息必须求助于空间插值技术,因此本文利用Kriging空间插值算法构建周围的环境,为边界检测做准备,并从多角度研究了无人机飞行轨线对Kriging预测的影响,通过仿真分析提出了比较合理的采样轨线;再次,运用模型预测控制(MPC)算法控制小型无人机循迹飞行,优化算法能够使无人机快速收敛到边界区域,使得路径最短,消耗资源最少,预测控制算法能够处理约束,同时对模型要求比较宽泛,因此较适合直升机的循迹控制,最后通过仿真分析验证了算法的可靠性;最后,借助地理信息系统(GIS)进一步对环境进行评估分析。GIS是地理统计方面比较专业的软件,能够将数据转化为可视化的3D图像,使结果更加直观清楚。
[Abstract]:Pollution plume is one of the most typical forms of environmental pollution: air floatation particles (PM2.5, etc.) radiation pollution, chemical gas diffusion, water pollution (oil leakage, etc.) Red tide and so on, manifested as the distribution of different concentrations / intensities in regional and polluted areas in space, and at the same time under the influence of self-diffusion and environment (wind, water flow, geographical environment, temperature, humidity change, etc.), the regional boundary shape changes with time. Presents an uncertain variation of the pollution plume similar to the feather boundary. For the rapid autonomous identification of the boundary of this kind of region, estimation and prediction are the key links in the monitoring of environment and emergency, and are the important basis for early warning, monitoring control and rapid response. In order to reduce the damage caused by the accident, this paper proposes to use single helicopter to monitor the environment. If there is environmental pollution, the boundary of polluted area will be detected and tracked. This paper mainly studies four aspects: first, build an environmental model to simulate the oil pollution diffusion in the real environment, explain the model in detail and make simulation analysis; secondly, Because the amount of environmental information obtained by single unmanned helicopter is very small, in order to obtain local complete information, we must resort to spatial interpolation technology, so we use Kriging spatial interpolation algorithm to construct the surrounding environment. In order to prepare for boundary detection, the influence of UAV trajectory on Kriging prediction is studied from many angles, and a more reasonable sampling trajectory is proposed through simulation analysis. Thirdly, the tracking flight of small UAV is controlled by model predictive control (MPC) algorithm. The optimization algorithm can make the UAV converge to the boundary area quickly, make the path shortest, consume the least resources, the predictive control algorithm can deal with the constraints, and the model requirement is relatively broad, so it is more suitable for the tracking control of the helicopter. Finally, the reliability of the algorithm is verified by simulation analysis. Finally, with the help of GIS (Geographic Information system), the environment is further evaluated and analyzed. GIs is a professional software in geography statistics, which can transform the data into visual 3D images. Make the result more intuitive and clear.
【学位授予单位】：华南理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：V19;X830

【相似文献】