أThesis (M.Sc.) -Cairo University, 2024.

Bibliography: pages 60-65.

This research investigates Airborne Gamma Ray Spectrometry (AGRS) data to detect naturally occurring radioactive anomalies, including potassium, uranium, and thorium, in the Wadi-Biyam area and its surroundings in Egypt's Eastern Desert between latitudes 22° 39' 45'' & 23° 01' N and longitudes 33° 47' & 34° 21' 30'' E. AGRS collects raw, unlabeled data from background radiation and radiation emitted by radioactive elements. When applying statistical methods for anomaly detection to AGRS data, several challenges arise: (1) Inaccuracy: The statistical methods might not accurately identify anomalies due to noise and variability in the data. (2) Time Consumption: Processing and analyzing the extensive data sets generated by AGRS can be time-consuming, delaying results and decision-making. (3) Bias: The statistical methods might be biased, resulting in skewed outcomes that fail to reflect the true distribution of radioactive anomalies accurately. (4) Limited Scope: While useful, traditional statistical methods have limitations regarding the complexity and variety of AGRS data. These limitations can restrict the ability to detect all significant anomalies, necessitating a more comprehensive approach. (5) False Certainty: These methods might produce results with a high degree of apparent certainty not justified by the underlying data, potentially leading to incorrect conclusions. The current challenges in anomaly detection can lead to severe consequences, such as misidentifying areas with high concentrations of radioactive elements or failing to detect significant anomalies. Therefore, the development of a new approach that can effectively address these challenges and improve the precision and reliability of anomaly detection in AGRS data is of utmost importance. We employ two machine learning methods for anomaly detection, DBSCAN and BIRCH, using the Silhouette Coefficient to evaluate results and plan mineral zones. Among the nine rock types studied, we observed significant variations in the concentration of radioactive elements, which we assessed using a Silhouette Score greater than 0.75 to ensure reliable clustering. Notably, the younger granite and wadi deposits stood out for their higher concentrations of these elements. Specifically, the younger granite exhibited a marked anomaly, with 7.55% of its points classified as abnormal, indicating elevated levels of radioactivity. Similarly, the wadi deposits demonstrated a significant presence of radioactive elements, with 4.22% of the points identified as anomalies. These findings suggest that the younger granite and wadi deposits play a crucial role in understanding the distribution of radioactive materials within the studied region, potentially offering insights into geological processes and the environmental implications of these anomalies.

من بين الأنواع التسعة من الصخور التي تمت دراستها، لاحظنا اختلافات كبيرة في تركيز العناصر المشعة، والتي قمنا بتقييمها باستخدام درجة Silhouette Score أكبر من 0.75 لضمان التجميع الموثوق. والجدير بالذكر أن رواسب الجرانيت والوادي الأصغر سناً برزت لتركيزاتهم الأعلى من هذه العناصر المشعة. وعلى وجه التحديد، أظهر الجرانيت الأصغر سناً شذوذًا ملحوظًا، حيث تم تصنيف 7.55٪ من نقاطه على أنها غير طبيعية، مما يشير إلى مستويات مرتفعة من النشاط الإشعاعي. وبالمثل، أظهرت رواسب الوادي وجودًا كبيرًا للعناصر المشعة، حيث تم تحديد 4.22٪ من النقاط على أنها مناطق مشعة. تشير هذه النتائج إلى ان younger granite وwadi deposits تلعب دورًا حاسمًا في فهم توزيع المواد المشعة داخل المنطقة المدروسة، مما قد يوفر رؤى حول العمليات الجيولوجية والآثار البيئية لهذه العناصر المشعة.

Issued also as CD

Text in English and abstract in Arabic & English.