Weld Recognition of Pressure Vessel Based on Texture Feature

In order to accurately identify the welds of pressure vessels, this paper proposes a machine vision recognition method based on LBP operator texture recognition. Compared with the common used recognition method that uses the gray gradient and gray distribution of the weld image, this method uses the texture features such as weld moire, and the feature that the texture features of the weld are very different from other regions, and uses the Bhattacharyya distance to compare the LBP histogram. The graphs are compared, and the threshold segmentation method is adopted to clearly distinguish the weld area from other areas, which can solve the problem of inaccurate recognition of different types of welds on the pressure vessel due to interference such as rust and light spots. Experiments show that this method can accurately identify the welds of pressure vessels.