An AI approach to lunar phase detection: enhancing the identification of the new crescent with astronomical data integration

Introduction The observation of the lunar crescent is significant in astronomy, cultural traditions, and religious lunar calendar determinations. However, earth-based imaging that captures all lunar phases, particularly the new crescent across multiple months, remains limited. This study explores the feasibility of using artificial intelligence (AI) techniques to detect and analyze the birth of the new lunar crescent using space-borne imagery from NASA’s Lunar Reconnaissance Orbiter (LRO), spanning over 13 years. Methods This study evaluates both deep learning and traditional machine learning approaches for new crescent detection. Convolutional Neural Networks (CNN), Random Forests (RF), and Support Vector Machines (SVM) were applied to orbital lunar images. A custom image preprocessing pipeline was implemented, including grayscale conversion, contrast-limited adaptive histogram equalization, and noise reduction. The CNN architecture was further enhanced by integrating lunar imagery with moon age data. Experiments were conducted using a temporally split dataset to simulate real-world conditions. Model robustness was also evaluated using synthetically generated noise and occlusion. Results The experimental results demonstrated high performance across all evaluated models, achieving precision, recall, F-score, and overall accuracy of approximately 98%. Among the tested approaches, RF and CNN models produced the best overall performance, outperforming SVM. The CNN model showed strong robustness under degraded image conditions, maintaining high accuracy when subjected to Gaussian noise and image occlusions of up to 50%. Discussion The findings indicate that AI-based techniques, particularly CNN and RF models, are effective for detecting the new lunar crescent from orbital imagery. The robustness of the CNN model suggests practical applicability in real-world lunar observation scenarios. This study contributes toward supporting traditional crescent identification methods and offers potential solutions for reducing calendar discrepancies across different regions.