Abstract

The digitization of cadastral maps plays a vital role in modern land management, facilitating efficient land administration, property transactions, and urban planning. However, many existing cadastral maps are still in paper form or stored in outdated digital formats, which are often difficult to access, process, and update. This paper presents an innovative solution utilizing Artificial Intelligence (AI) to automate the digitization of cadastral maps. By applying machine learning, computer vision, and image processing techniques, the proposed system is designed to accurately extract property boundaries, classify map features (such as plot numbers, ownership details, and administrative boundaries), and convert legacy map data into standardized digital formats compatible with Geographic Information Systems (GIS). The system addresses key challenges such as poor map quality, inconsistent boundary definitions, and varying map scales. The result is an AI-driven approach that significantly enhances the efficiency, accuracy, and scalability of cadastral data management, enabling real-time updates, minimizing human error, and improving the overall accessibility of cadastral information for land administration and policy-making. This paper presents an AIdriven approach for automating the digitization of cadastral maps, converting paper- based or outdated digital maps into accurate, standardized digital formats. Using machine learning and computer vision, the system extracts property boundaries, recognizes map features, and integrates them into modern Geographic Information Systems (GIS). The proposed solution addresses challenges such as poor image quality, irregular boundaries, and inconsistent annotations, streamlining land management, reducing human error, and enhancing the accessibility of cadastral data for urban planning and property administration