Does the camera perform visual SLAM localization, lane line recognition, or any additional computer vision functionalities?

No, the Vision-RTK 2 is a real-time system that provides localization data based on its sensor inputs. The Fusion engine uses the in-built camera for visual tracking of extracted features. However, if the user wants to access the image data for own applications, then this is possible, by configuring a RTP camera image stream: https://docs.fixposition.com/fd/camera-image-streaming. The camera stream can be used for on-field monitoring purposes or more complex applications.