With the rapid growth of robot technology, robotic tiling to replace manual work gains increasing attention. However, existing studies may have oversimplified their working conditions in automatic tiling work, where the floor-tiling robot can only tile the middle area in a room, leaving the edge area to be tiled manually. This research aims to provide a vision-based solution for floor-tiling robots to determine the profile of the edge areas, aiming to improve the automation level of the tiling robot.