Improved visual inspection for nozzle inner radius based on panoramic imaging

Abstract

Visual inspections of Nuclear Power Plant (NPP) reactors are important for understanding the current condition of the reactor components, such as nozzle inner radius. After addressing current deficiencies of the existing visual inspection methods, a panoramic image mosaicking method is proposed in this study to improve the efficiency and reliability of detection. However, unlike traditional mosaicking methods, the proposed method obtains the variation of the pixels between frames in a video by establishing a geometric mapping relationship between scenes and images. Then, the pixels which are varied between frames will be extracted and the frames will be combined to generate the mosaicked images. A function to discriminate overexposed pixels is also derived based on the redundant pixels in a series of frames, to reduce the impact of specular reflection on the resulting images. These above methods are applied to the detection of defects on nozzle inner radius of the mock-up. The resulting 360° panoramic images provide a large-field-of-view defect analysis approach to obtaining geometric measurements with pixel-level precision, and the image optimization method is found to mitigate the effects of specular reflection on defect identification by providing clear and reliable images. The results of this study demonstrate the promise of this new visual inspection method for NPP reactors, and establishes a potential foundation for the automatic detection of surface defects. Contribution-A method is presented to improve the reliability of defect detection, and it has been applied to the visual inspection for nozzle inner radius. Contribution-A method is presented to improve the reliability of defect detection, and it has been applied to the visual inspection for nozzle inner radius.