The distance measure, calculated from the apparent width of a pole in the processed image has a very low update rate compared to the framerate that is transmitted by the drone. It also seems to be lower than the update rate of the detected pole-heading. This is weird, could have to do with multithreaded approach for processing the vision data and needs investigation (and fixing of course).