Wouldn't it be handy to know how full your train carriage is? You might score a seat or stay out longer until its less crowded. Or how about learn when platform are overcrowded? Perhaps even detect trespassers in rail corridors before a train is in the area. My FYP utilises CCTV footage to monitor people levels in various environments to give data on when more services are needed and when safety is at risk. This project builds on the YOLO detection system which uses a convolutional neuronal network to recognise a range of objects, including people. My aim is to reduce the 'flickering effect' by adding temporal connectivity to video frames and give a simple user interface for an operator to monitor a set of cameras.