Sub-pixel Techniques For System Of Diagnosis Of Conical Thread

As channels to transport petroleum, the tubing is widely used in oilfield, and the quality of tubing thread's connection has great effect on normal oil production. At present, the conical thread measuring method is still staying on the stage of manual method, which is very inefficient. It is urgent to develop a more effective thread measuring system suitable for stable production.This paper introduces a system for auto-measuring parameters of conical thread based on sub-pixel techniques. The system obtains 2-D data of the thread axis profile section by CCD with optical projection and machine driving. The median filtering method is adopted to avoid destroying edge information while filtering the noise in the image. Based on the analysis and comparison of main methods for edge detection the method of binary edge detection is applied in edge detection of conical thread edge. After obtaining the edge points series, the geometry parameters of conical thread can be accurately calculated by using the sub-pixel locating algorithm based on interpolation polynomial and the curve fitting algorithm with the contour of conical thread, then demonstrate how to compute the geometry parameters of conical thread. It is testified the system can meet the diagnosis of conical thread.The main content of this thesis includes: the principle of the system, image acquisition and preprocessing, edge detection and sub-pixel technique, how to compute the geometry parameters of conical thread.