Fuzzy Logic-Based Classification of Crescent Moon Images Using Contrast and Thickness

Accurate determination of the crescent moon (hilal) is crucial for establishing the start of lunar months in the Islamic calendar; however, observations are frequently hindered by daylight conditions, atmospheric disturbances, and subjective visual interpretation. This research proposes a fuzzy logic-based classification system to evaluate crescent moon images using contrast and arc thickness as input parameters, providing a transparent, rule-based alternative to black-box machine learning models for hilal visibility assessment. Images were collected on four distinct observation dates (May 28, 2025, August 5, 2024, September 16, 2023, and May 9, 2021) under varying atmospheric conditions and crescent appearances. Each image underwent pre-processing to extract quantitative measures of arc contrast and thickness, which were subsequently fuzzified using triangular and trapezoidal membership functions. A fuzzy inference system employing expert-defined rules was then used to compute a visibility score for each observation. The resulting visibility scores of 0.4691, 0.4604, 0.4689, and 0.4154, respectively, placed all four observations within the “partially visible” category. These findings demonstrate the system's capability to manage observational ambiguity in daylight conditions, showing potential for reliable classification while still requiring validation on larger datasets and clear non-visibility cases, and offering a transparent and interpretable framework to support more consistent and standardized hilal classification for calendrical purposes.