Vision guidance is a challenging component of robotics, but there are definitely tools available to make it work. Check out OpenCV for starters, which would probably work with a RasPi and webcam, or if you want to start down the rabbit hole, look up ROS and possibly the Xbox Kinect. ROS is a framework for controlling high level robotic behaviour and can connect to the Kinect for vision. It's definitely not a beginner topic, but is immensely powerful. I've only gotten started but hope to make my robotic project have Kinect-vision.