This work is devoted to improving the reliability of thermal defectoscopy of pipelines and steam pipelines using integral filtration. The study addresses the problem of signal distortion caused by surface emissivity inhomogeneities, which complicate defect detection. A theoretical and experimental approach was applied, combining the development of a thermophysical model of a pressure pipeline with practical testing on real defective samples. Two main methods of interference suppression were considered. The proposed method of averaging the temperature field in regions with different emissivity coefficients reduces interference by more than three times while slightly distorting the temperature field. The results demonstrate the potential for transitioning from a visual to an automated defect identification system in industrial conditions.
