Spacecraft damage infrared detection for hypervelocity impact based on multi-objective optimal clustering algorithm

Abstract

The complex defects from hypervelocity impacts (HVI) present significant challenges for detecting and assessing damage on spacecraft surfaces. A multi-objective infrared feature extraction is developed to classify defect types while preserving local data correlations. Using a multi-objective evolutionary algorithm with NSGA-III, the method optimizes objective functions to identify the most representative transient temperature response (TTR). For the damage characterization, an image segmentation strategy with active contour model (ACM) is introduced to obtain the quantitative result of infrared reconstruction images (IRRI), completed by the level set contours. Experiments show the classification based on multi-objective optimization plays an effective role in the HVI damage detection.