Study On Binocular Fusion Technology In 3D-PIV System

PIV(Particle Image Velocimetry) is a new fliud field measurement technology which developes with information science, image processing and several other subjects. Compare with traditional methods, PIV has many advantages because it measures the whole instantaneous fliud field without disturbance. At the present day, 2D-PIV has been used as an important experiment method in hydrodynamic and aerodynamics.3D-PIV system based on binocular vision employes computer stereo vision theory and image processing means. It can measure 3D velocity vectors in a certain volume of the fliud field.Being sustained by National Natural Science Foundation of China, this paper researches 3D-PIV system detailedly, including theory and component of system, camera and binocular model, and image processing such as camera calibration, feature extraction, 3D reconstruction. Binocular fusion which is the most important and difficult technology in 3D-PIV system is the key of this paper.According to the characters of the particle images, a new binocular fusion algorithm is proposed, which is based on radius directions fitting and luminance sorting. The result from experiments showes that this method can eliminate the noise in the particle images, and contributive for particle extraction in sheltering condition. It can also improve the rate of particle fusion which satisfies 3D-PIV system.In consideration of demand, 3D-PIV system is designed for experiments. The hardware is established with appropriate equipments and the algorithms in this paper have been programmed. The considerable quantities experiments demonstrate the stable performance of the hardware and effectiveness and practicability of the image processing method, which can provide good measurement service for users.