Research Of Method To Improve Measurement Accuracy Of The Deformation Of The Vehicle Missile Launcher

This paper designs non-contact optical measurement system based on CCD as the core, the vehicle missiles as the research background and the vehicle missile launchers as the target. This deformation measurement system consists of lighting system, measured object, optical imaging system, optical sensors, signal processing circuit, and computer ,etc. This system is aimed at measuring the small deformation of vehicle missile launchers, providing the angle compensation date for the launch of missiles, reducing the error of missile launch angle, and effectively improving the accuracy of missile launch.This paper firstly introduces the development of deformation measurement system at home and abroad, then introduces the basic principle of the car missile deformation measurement system. This paper gives out the block diagram of hardware components and diagram of hardware block. Finally, according to measurement principle, it sets up an experimental platform in order to collect experimental data and make research of software algorithms.In this system, the accuracy of central laser spot positioning is an important factor for the deformation measurement system. This paper focuses on the software algorithm, and further improvement of measuring accuracy of the deformation measurement system, by using the sub-pixel location algorithm for accurate positioning of laser spot center.This paper introduces and analyses the common sub-pixel location algorithms, then proposes the curve fitting algorithm for sub-pixel based on the barycenter. Firstly, we get the central position of collected laser spot images by the method of barycenter. Then, the curve fitting is used to locate the position of laser spot center on the high-precision sub-pixel method. After many experiments, it is proved that the standard deviation of distortion angle is less than 15" by the method of this algorithm, which improves the accuracy of deformation measurement system. At the same time, the algorithm only do some little computation on the collected image data, making it easy to describing in computer language, and conducive to project implementation.