In this paper, an improved feature extraction algorithm in Eddy Current Pulsed Thermography (ECPT) is developed to realize automatic defect identification. The proposed feature extraction algorithm includes a data block segmentation, a variable interval search, a correlation value classification and a between-class distance decision function. The data block segmentation and variable interval search are firstly combined to reduce the repetitive calculation in automatic defect identification. The classification and between-class distance are used to select the typical features of thermographic sequence. The method is not only able to extract the main image information, but also can reduce the time of thermographic sequence processing to improve the detection efficiency. Experiments and comparisons are provided to demonstrate the capabilities and benefits (i.e. reducing the processing time) of the proposed algorithm in automatic defect identification. (C) 2017 Elsevier Ltd. All rights reserved.