Schematic of the automated tip preparation tool. (i) The process starts by scanning an area to obtain an image of the surface. The image is processed to make it appear flat and to remove the bottom 20 lines from the scan to avoid visible creep before it is then classified. (ii) A cross-correlation-based analysis script compares the input image to a reference and outputs an estimated binary classification of Good or Bad. (iii) If the tip is classified as Good, the script then exits. (iv) If the tip is classified as Bad, the script will move away and attempt to reprepare it. (v) If the tip has already been through a set number of shaping events at this point, the script will instead reposition the scan area away macroscopically using the coarse motor, under the assumption that the area being scanned is not suitable to classify the tip.
