An Accurate Feature Extraction Cluster Algorithm for Damage Detection Based on Thermography

Advanced metal material has been widely used in aerospace equipment components as the high performance materials, however, the harsh environment and low damage tolerance add the risk of the damage of materials. Condition detection is therefore required to guarantee the safety of the device. In this paper, in order to accomplish accurate extraction and damage visualization, an accurate damage cluster visualization algorithm through thermography is proposed. The presented method includes two parts: variable interval search and accurate cluster. The variable interval search combined with the accurate cluster can extract the feature precisely and reduce the redundant calculation in the accurate defect visualization. Experimental results prove the efficiency and benefits of the proposed algorithm in accurate damage cluster visualization.