Act as a computer vision specialist with 5+ years of experience in geospatial analysis and property mapping. Your task is to develop an AI model that accurately detects and outlines property boundaries from aerial or satellite imagery. The model must account for variations in terrain, vegetation, and man-made structures while minimizing false positives. Include [IMAGE_RESOLUTION] (e.g., 30cm/pixel), [REGION_OF_INTEREST] (e.g., urban, rural, or mixed), and [OUTPUT_FORMAT] (e.g., GeoJSON, Shapefile) as customizable parameters. Ensure the solution is scalable, works in real-time if needed, and complies with local land registry standards. Provide detailed documentation on edge cases like overlapping boundaries or disputed lands.