Development and Performance Evaluation of an Image Recognition System for Unmanned Aerial Vehicle Automatic Landing

• Main Authors: YANG-CHENG-LI, 楊筬理
• Other Authors: 曾百由
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/jc7t9d

碩士 === 國立臺北科技大學 === 機電整合研究所 === 104 === In recent years, unmanned aerial vehicle(UAV) has been applied in various fields because of its advantages, such as high mobility, vertical takeoff, landing and hovering. Moreover, embedded system is ideal for the development of UAV flight system because its features, for instance, small size, light weight, low power consumption, highly processing speed and stability. On the other hand, machine vision is used in many systems, which can retrieve the information of environment with a non-contact and non-destructive methods. All the forementioned benefits are helpful to calculate the attitudes and position of a UAV during landing and take-off. The purpose of this study is to develop an image recognition system which can provide the information of the attitudes and relative distance to the landing position of a UAV for automatic landing. In this study, a digital camera is used to identify the landing mark with an embedded system hardware architecture. In order to obtain supplementary information in real-time, a multi-core embedded system is used for multi-threading tasks to increase operating efficiency to achieve an average image processing rate of 30 frames per second (33 ms /frame), and validate the image identification accuracy of this developed image-based guidance system.