In this paper a single camera vision guidance system for fixed wing UAV is developed. This system searches for and identifies a target object with known colour and shape from images captured by an onboard camera. HSV colour space and moment invariants are utilised to describe the colour and shape features of the target object. Position, area and rotation angle of the target object in the image plane are collected. This information is then processed by the Extended Kalman Filter to estimate the relative positions and attitudes of the UAV. The vision guidance system guides the UAV towards the target object automatically based on these estimated states by using a proportional controller. A Senior Telemaster aircraft model kit installed with an onboard camera and computer is used for flight test. The target object for the flight test is a white flag with a red cross. Flight simulations and flight tests results are presented in this paper, showing that the vision guidance system can recognise the target object and guide the UAV effectively.
All Science Journal Classification (ASJC) codes