Research on Deep Hole and Large Thread Defect Detection Based on Machine Vision Fusion

The threads between the dome cover and cylindrical body of the nuclear reactor pressure vessel have experienced corrosion and cracks during operation, causing problems such as seizure, sliding and connection failure, resulting in major safety production accidents. Regular defect detection is necessary. Due to the traditional manual investigation of this large thread, the working conditions are harsh, the risk factor is high, the technical content is high, and the investigation time is limited, making it difficult to accurately determine the damage inside the main bolt hole. This paper proposes a machine video detection method for deep-hole large thread defects. Using the large depth of field of CCD to scan the thread, automatic detection control and efficient mechanical structure and light path system are used to clearly observe the top and bottom of the deep-hole large thread, and imaging preservation is carried out for future reference. Three reasonably arranged cameras are used to provide remote real-time observation of the entire picture of the deep-hole large thread, which provides assistance for real-time detection. The method is convenient and effective after experimental verification.