Research Of Multi-view Match Method In Measuring The Nozzle

The vector nozzle of the aircraft is a multi-function exhaust pipe, which can make the aircraft roll, yaw and pitch. The flexibility, mobility, survivability and combat capability of the airplane has been made great improvement because of its appearance. The work environment of the vector nozzle is relatively poor and its movement is complex, which brings a lot of inconvenience for measurement by traditional research methods.Through the multi monocular vision measurement system, we successfully obtain geometry parameters of the vector nozzle. This method was first used to measure research under the tail motion vector geometry parameters.In the process of measurement, multi-view matching is a key technology. The success of matching affects directly the realization of the follow-up measurement.After studying traditional matching methods, this paper have improved matching method based on feature points. In this method, firstly we need to solve the affine transformation matrix between the images by multi-view calibration. According to the matrix, the feature points extracted from images were transformed to get right matching points. Based on the relationship of image coordinates and world coordinate,we can obtain right matching points’ value in three-dimensional space. Finally, the geometric parameters of the engine exhaust nozzle are measured.We set up different platforms for different filed. From binocular vision measurement system in simulated environment to four binocular vision measurement system in real environment of vector nozzle, we did lots of experiments to verify that the proposed matching method is feasible. The error of radius is within 2mm. The error of declination is less than 0.2 degrees. The measurement results are very satisfactory.