The landmarks are often used in SLAM. What are the algorithms used to extract them, and how can a robot differentiate the landmarks, if they detect one in point A at X<sub>t</sub> and another in X<sub>t+1</sub>? How can the robot know if it's the same landmark or not?