Font Size: a A A

Research On Runway Identification And Verification For Airborne Vision-based Auto-landing System

Posted on:2011-05-24Degree:MasterType:Thesis
Country:ChinaCandidate:H LiFull Text:PDF
GTID:2212330338996038Subject:Navigation, guidance and control
Abstract/Summary:PDF Full Text Request
This paper focused on the runway identification technologies during the process of UAV autono-mous landing. Besides, a set of ground simulation verification platform was designed, which could help us to verify corresponding identification methods, and the fine results were got.Firstly, the paper generally introduces a present overview of the status of autonomous landing techniques at home and abroad. By analyzing the characteristics of digital images from forward look-ing camera which reflect the airport environment, the vision processing and object recognition process were determined. According to salient features of reflected in the different period of landing, an adap-tive thresholding method based on histogram local minimum was proposed, which could effectively segment runway target area with long-range. Meanwhile, with the advantage of an improved Canny edge detection, a method that could reliably extract edge information of runway target in the closer distance was proposed.Secondly, based on the area and edge information of the runway, a complete set of runway object recognition solution was determined according to the different application scope of former two image processing methods, and the whole landing process was definitely divided into far-stage and close-stage. For far-distance stage, a template matching method for runway target recognition could be used. And for close-stage, with a method based on line characteristics extraction, the vision system can also recognize runway target area reliably. Through the transition conditions determined, a recog-nition method which could switch automatically during the process from far and near area was real-ized.Again, in order to verify and test the image processing algorithm and the whole recognition method, an embedded system simulation verification platform mainly based on DM642 was designed. It can simulate parts of scenarios of UAV landing process on the ground and provide the runway iden-tification experiments.Finally, the corresponding experimental results show that the designed runway identification so-lution is valid during the whole landing process .In the addition, the content of the research is also summarized and prospects in this paper.
Keywords/Search Tags:Autonomous Landing, Adaptive Thresholding, Edge Detection, Runway Identification, DM642
PDF Full Text Request
Related items