The Application of Artificial Intelligence and Deep Learning to Visually Identify Micrometeoroid and Orbital Debris Impacts

Abstract Recent advancements in artificial intelligence (AI) have made machine learning (ML) techniques readily available for practical applications while using a fraction of time that was previously required. In particular, the use of deep learning (DL) models has the potential to assist in the identification of micrometeoroid and orbital debris (MMOD) damage. The ability to detect MMOD damage with AI models has the potential to improve the safety of human space flight by reducing the time required for damage identification via automated methods. To demonstrate this, a binary classification dataset of 226 images was utilized to train and validate a MMOD damage detection model using a residual network architecture. The benefits of data augmentation were consistent and improved performance as expected. For performance, the MMOD damage detection model achieved 95.6% accuracy on the validation set and generalized to an accuracy of 87.5% on a test dataset of spacecraft images with MMOD damage. With the use of deep learning algorithms, MMOD damage can be identified with high accuracy and provides a new possibility to identify of hypervelocity impacts.