This works presents a comparative evaluation of visual SLAM algorithms in the context of precision agriculture. The authors assessed the performance of ORB-SLAM3 and RTAB-Map using two datasets: the publicly available ROSARIO dataset and a custom field dataset collected with a ZED stereo camera and RTK GNSS mounted on a Jetson Orin NX. All experiments were conducted within the ROS2 framework to ensure synchronized and consistent testing conditions. The results demonstrate that visual SLAM is a viable approach for deployment in real-world agricultural environments, provided that SLAM algorithms are continually refined and validated using domain-specific datasets. Enhancing these systems through improved sensor fusion will be key to developing more robust and reliable autonomous farming robots.