摘要：近年来四旋翼无人机领域飞速发展，对四旋翼无人机导航系统的研究也逐渐成为热潮，目前主流的导航方法有GPS导航和惯性导航两种，但随着无人机活动范围和飞行时间的扩张与延长，GPS导航受环境影响较大和惯性导航误差随时间变大的缺点凸显，视觉传感器由于其廉价特性和对环境信息采集量丰富的特点，在四旋翼无人机导航方面有较好应用前景。本文研究了基于视频图像处理的四旋翼无人机自主巡线技术，主要内容包括：

首先搭建无人机硬件平台，选型无人机硬件选型。

然后研究机载摄像头对地面路径信息的提取问题，主要包括：图像压缩与二值化，图像分割与填充，通过路径搜索框解算出无人机与路径的横移量偏差与航向角偏差，以及通过卡尔曼滤波对采集到的导航信息进行处理。

最后对无人机进行动力学建模，设计基于PID的飞行控制器，并搭建Simulink模块进行仿真。并设计上位机软件，通过实际飞行试验，验证PID控制器实现无人机巡线飞行的可行性和有效性。

关键词  四旋翼，机器视觉，路径跟踪，PID

毕业设计说明书外文摘要

Title    A Study on Quad-Rotor UAV Path Navigation Based on Visual Processing                                                                       

Abstract：In recent years, with the rapid development of four-rotor unmanned aerial vehicles, the four-rotor UAV navigation system research has gradually become a hot, the current mainstream navigation methods are GPS navigation and inertial navigation . But with the expansion and extension of the UAV range of activities and flight Time, GPS navigation is greatly influenced by the environment and the inertia navigation’s error increases with time going on. The visual sensor has more advantages in terms of its cheap characteristics and rich information on environmental information collection, which make it has good application prospects in the four-rotor UAV navigation field .

In this paper, quad-rotor UAV’s route guidance based on visual processing is studied, ground path information from onboard camera is used for UAV path following as navigation information.The main comtents are as follows:

First of all, we design UAV navigation system , and build UAV hardware platform.

Then, we studies the problem of extracting the ground path information from the airborne camera. It mainly includes: image compression and binarization, image segmentation and filling, using the path search box to calculate the amount of transverse deviation and heading angle deviation between UAV and path , and using the Kalman filter to filter the collected navigation information.

Finally, the dynamic modeling of the UAV is carried out, the PID controller based on PID is designed, and the Simulink module is simulated. Then we design the host computer software, and, through the actual flight test, verify the PID controller’s feasibility and effectiveness on UAV patrol flight.

Keywords  Four rotor, machine vision, path tracking, PID 

目   次

1   引言 1

2  无人机硬件平台搭建 Y

2.1  四旋翼无人机机械结构 Y

2.2 无人机硬件平台 Y

2.3 本章小结  Y

3   基于图像处理的路径识别与跟踪  Y

3.1 图像金字塔压缩 Y

3.2 路径填充分割 Y

3.3 路径导航信息提取 Y

3.4卡尔曼滤波 Y

3.5 上位机设计 Y

3.6 本章小结 Y

4    四旋翼无人机飞行控制策略 Y