| When building a nuclear power plant,a layer of steel plate concrete structural module(referred to as the SC module)is built as a safety barrier.The so-called SC module is a steel sheet connected to the wire and injects the concrete,the SC module can be shielded.The role of protection.The most important process for manufacturing the SC module is to pull the bisector,but there are more than 3,000 holes on the SC module.Each hole needs to be welded to fix steel plates and tiles,and the welding efficiency in traditional manual welding will be extremely low.Therefore,it is necessary to study the positioning of the machine vision technology to identify the plug welding and guides the robot to prepare,and the degree of tribese welding visual recognition system is high,which can greatly improve processing efficiency.This paper studies the technique development,using Gaussian filter,image binarization,refinement,Sobel operator,and morphological processing,and structural light bar extract.The image is linearly detected by Hoff conversion,and the grades obtained by grayscale gravity will be divided into four parts: four parts of the upper and lower lines,and seek the intersection of each part.The combination of the plug welding holes is combined,which in turn is identified.From the effect,the algorithm can stably identify the plug welding holes and provide accurate positioning information,which indicates that the algorithm is highly applicable to the visual identification of the stunzzer welding hole. |
