Defects Inspecting System for Tapered Roller Bearings Based on Machine Vision

Many bearing manufacturers take traditional manual methods to inspect defects of bearings, which are in low efficiency, costly and unreliable. In this paper, a defects inspecting system for tapered roller bearing based on machine vision (DISTRB) is introduced. DISTRB is used to realize online automatic inspection, which includes mechanical parts, electrical parts, lighting equipments, a CCD, a computer and its software. DISTRB actualizes automatic online inspection by CCDs and defects recognizing algorithms and takes the place of workers' eyes. Some implementation devices are used to get high-quality images. Defects recognition algorithms of DISTRB are mainly elaborated. The defects of bearings include roller-missing in tapered roller bearings (RM) and some rollers being installed opposite to the direction of standard need (ROD). The key of DISTRB is to use practical image processing skills to improve the accuracy of the defects recognizing algorithms. The accuracy and stability of DISTRB mainly depends on the algorithms with a precondition of a fixed lighting part and imaging part. The DISTRB involves multi-domains such as mechatronic engineering, computer science, image-processing and pattern recognition. In this paper, taking image of type 32011X bearing as a research object, defects recognizing algorithms for RM and ROD defects are illustrated. After theoretical calculation and experiments in laboratory and practical applications in factories, the defects inspecting system is reliable, stable and effective.