Body-worn cameras are increasingly being used by Police to record their activities and gather video evidence. The benefits promoted from using body-worn cameras range from reducing violence escalation to eliciting guilty pleas during suspect interviews.
Many forces are either initiating trials or going towards full scale deployments. As a
result, companies are increasingly offering camera devices, but is anyone seriously thinking about the back end?
So let’s look at some of the challenges that police will face as they scale up in their usage of body cams.
Firstly, the sheer volume of data which will result from the rollout of body worn cameras is daunting. For example, a medium sized European state with 30,000 frontline police could generate 90 million hours of video per year. This is approximately twice the rate of video generation that YouTube deals with globally. The issue of storage and its cost will be a big issue for forces to deal with.
So once you have the video recorded, how do you manage it for evidential purposes? Lots of people will need access to the video as part of the criminal prosecution system; the officers who recorded it, the arresting officers, the custody managers, prosecution, defence and so on. Video and associated evidential information must be made available quickly to elicit guilty pleas and enable prosecutions in a timely manner. The current delays evident with CCTV footage costs a lot, with body worn video thrown in, issues and associated costs will increase. A back end storage solution needs to allow for multi user access. The question then arises, if you have a shared platform for access, how do you make it available and searchable, cost effectively?
Video Intelligence Exploitation
A lot of thought has gone into the technical guidelines for body-worn-camera devices. The UK Home Office Body-Worn Video Technical Guidance highlights many essential device characteristics. Some of these will greatly help how the video is dealt with at the back end, such as timestamp information, book-marking functions, GPS information and export format. What data is required for optimising back end searching and indexing cannot be a secondary consideration. How to sift through recordings to find pertinent intelligence data has to be mapped out. If data from body-worn-cameras is stored in an intelligent manner, analytics can be used to exploit the intelligence within for predictive policing. For example, you can connect video data from different officers relating to the same incident based on GPS and timestamp location.
Going a step further intelligent video analytics could be used to analyse information within the video for intelligence exploitation.
Protecting Citizens Privacy
Finally and most importantly, privacy of the citizen must be protected. Body worn video also captures innocent bystanders who may not want video details of themselves distributed or shared. One increasing trend is that citizens who are aware that they are being recorded are requesting copies of the video under freedom of information acts. Currently, it is the job of specialist AV technicians to edit, blur faces and perform other video annotations. It again becomes a question of resources, will they have enough time? One solution is to deploy intelligent algorithms that detect faces and blur faces automatically. Not only would these ease pressure on resources, it would be doing the upmost to ensure personal privacy.
Hopefully, body-worn-camera trials will continue successfully. I just hope the back end exploitation is carefully considered!