The video UAV flight control and visual system of measurement

• Main Authors: Chung-Hsien,Tasi, 蔡忠憲
• Other Authors: Chung-Cheng,chiu
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/66566035516910704584

碩士 === 國防大學理工學院 === 電子工程碩士班 === 98 === In this thesis, we use the video image captured by an unmanned vehicle to detect the skyline. The proposed method detects the angle and position of the skyline and then calculates the roll angle and the pitch value. The distance between the height of the skyline and the central level called the pitch value dp. These parameters will be processed by proportional control and integration control to get the output voltage in order to stabilize the flight. There are two new algorithms added in the previously study, they are adapted tracking and central block decision mechanism algorithms. The adapted tracking algorithms calculates the two terminals from the edge points according to skyline changes. It uses skyline as the basis to find the point of the biggest difference on skyline edge as the tracking points. By tracking the points continuously, we can reduce the calculation works effectively. The goal of the central block decision mechanism is to make the skyline detection more robustly. It records the information by sampling one of the two points in central block of the image, and determinate if the error detection of the skyline occurred. It is proved that the two check mechanisms can improve the precision of the skyline detection, and become more adaptable under various environments. The proposed system uses computer to detect the skyline and calibrate the offsets from wide-angle lens to reach the goal of automatic surrounding flight. We are now successfully controlled the roll, pitch, and yaw and overcome the complicated problems of images during surrounding flight. The experimental results prove that the vision-based flight control is practicable.