Advanced object detection for heritage preservation: A case study of Ouaouizerth and the Palace of Sultan Moulay Ismail

Abstract. Ouaouizerth, a historic town located about 80 kilometers southwest of Béni Mellal, stands as a testament to Morocco's rich and diverse history. Founded centuries ago, the town is home to landmarks like the Palace of Sultan Moulay Ismail, commissioned around 1710 and also known as Kasbah de Boulaouane. This palace, built during the reign of Sultan Alaouite Ismail Ibn Sharif, showcases a blend of grandeur and architectural skill. Positioned strategically overlooking the Oum Er Rabia river and inspired by Byzantine fortifications, it offers a fascinating glimpse into historical architecture. Preserving these historical monuments is vital, yet traditional methods of documentation and categorization are often labor-intensive and prone to inaccuracies. This study tackles these challenges by developing an advanced object detection model to improve the efficiency and accuracy of degradation detection in these historical monuments. Utilizing the object detection method, a robust model was created for detecting and classifying various forms of deterioration. The dataset included meticulously captured and annotated images from Ouaouizerth and the Palace of Sultan Moulay Ismail, featuring a range of degradations. Annotations were performed using Label Studio software to ensure precise identification and categorization. Data augmentation techniques were employed to create a diverse and representative dataset, essential for effective model training. The model's performance was rigorously evaluated using metrics such as precision and recall, demonstrating its high accuracy in detecting specific deteriorations. The successful implementation of this model holds significant potential for aiding conservation efforts, providing a valuable resource for heritage preservation, urban planning, and tourism promotion, thus ensuring the enduring legacy of Morocco's historical treasures.