This is because modern robotic vision, including visual place recognition, typically relies on power-hungry machine learning ...