Glasgow CCTV Object Count Data Edit

Object Count data collected from over 40 CCTV cameras owned by Glasgow City Council. Since 2019, Urban Big Data Centre at University of Glasgow, Glasgow City Council and Glasgow Centre for Population Health have been collaborating on a project to combine Data Science with the city’s existing CCTV network to generate object count data.

Selected CCTV cameras take a single frame image every 30 minutes across the 40 plus cameras and the images are processed using machine learning algorithms to detect and count various objects – the objects include person, car, bicycle, motorcycle, bus and truck. The counts of objects in the frame are collected and returned via an API created to access and query the data.

Users can query the API to return data on object type, location and time periods since November 2019.

As the data detects objects only once every 30 minutes the data cannot be used to determine the absolute number of objects (such as persons) in one place over time. Rather the data should be used to compare the difference in object counts between places at the same time or the same image frame over two or more time periods.

Note that the camera images do not store or track individual images of people, rather the algorithms are trained to recognise the outline shape of an object only and count it every time it detects the object in the image frame. One of the image count data are retained or stored after they have been counted.

Resources

• Web Page ARCGIS HUB DATASET (Details)
• Esri REST ARCGIS GEOSERVICE (Details)

Additional Info