Research On Landmark Tracking For MUH

The mini unmanned helicopter, with distinguished flight performance and a promising future, has been a hot research issue nationally and internationally. It is an irresistible trend to apply machine vision technology to MUH, including vision-based navigation, visual obstacle avoidance, visual landing and visual tracking, etc. Landmark tracking is the key point for a MUH's vision system. As to the research on landmark tracking for MUHs, the work this paper has done is mainly as follows:1. Systematically reviewing the development of MUHs and visual tracking, introducing the current research status of related fields at home and abroad, summarizing the existing research achievements, and analysing the key problems this paper facing.2. Considering reasonably the aircraft bearing capacity and the image quality requirements, designing and constructing the hardware and software framework of mini unmanned helicopter landmark tracking.3. Embedding online feature selection into particle filter for landmark tracking, and the feature with the best discriminability between the target and the background is selected for computing the likelihood image.4. Tracking landmark with particle filter method. Moving the particles to the region with a larger value during the tracking process to improve the effectiveness of the particles.5. Testing the performance of landmark tracking in video image sequences, contrastive analysis of tracking results between the proposed and common visual tracking methods shows the practicability of the landmark tracking algorithm presented in this paper.