Research On Multi-view Vision Measurement Method Based On Halcon

Engine thrust vector nozzle, is a kind of multi-function exhaust pipe,which is used tocontrol the flight posture of the aircraft. It develops from the originally fixed nozzle up to nowthe thrust vector nozzle which is more and more powerful, and the structure is more and morecomplex. So studying on both the space deformation and movement parameter measurementof vectored nozzle is of great significance to research and design the engine thrust vector andto control the flight posture of aircraft. Because of the complex movement of engine nozzle,the limited space,and the effect of working under high temperature airflow, conventionalmeasurement methods can’t be applied. The development of modern three-dimensional visionmeasuring technique has provided a new technical solution to solve this problem. With thecomprehensive utilization of multiple camera calibration technology, the stereo visionmeasurement technology and three-dimensional reconstruction technology, a multi-viewstereo vision system can be established to realize the motion measurement of vector nozzle.In recent years, due to so many advantages such as non-contact, high accuracy, strongreal-time performance, vision measuring technology has been widely researched and appliedin the modern industrial manufacturing, aerospace, and people’s daily life. Based on theexperiment platform of binocular vision measurement, this paper mainly studies on the basisof theory and experiments for camera calibration, feature extraction, feature matching andthree-dimensional information reconstruction of the binocular stereo vision system and so on.After the analysis of the system structure parameters and the selection for cameras, lens andscaffold, this paper sets a binocular measurement experimental platform and acquires theimage sequence. To start with, a mathematical camera imaging system model is established,then on the basis of Zhang Zhengyou camera calibration method, this paper studies a quickbinocular calibration method based on HALCON software platform; As for image featureextraction and matching. Harris corner detection algorithm is studied; A fixed windowmatching method based on image region is studied. The basic knowledge of stereo matchingis introduced and a feature points matching method with polar constraint is also proposed. Finally, through above work, the three-dimensional object points reconstruction and thetarget’s size measurement is completed by related experiments.