How useful are the Home Office’s i-LIDS video analytics tests? asks Frank Crouwel, Managing Director, NW Systems, pictured.
Following NW Systems look at the new Surveillance Camera Code earlier in the year, we decided to follow by taking a closer look at video analytics. That market has had compound annual growth rate (CAGR) of over 16 per cent for several years now and was predicted to have a value of over $88m this year (source: IMS Research). Traditionally video analytics software has required a great deal of processing power to make it work – hence nearly all of its focus in the early years was PC-based. PC-based analytics software still makes up over 75 per cent of total software revenues in the space (Source: IMS Research).
But analytics software is being embedded increasingly in into IP-based ‘edge’ devices, that is, network cameras. Several manufacturers, most notably Mobotix, have made impressive progress in this area in the last year. One key advantage of edge-based analytics is that it takes pressure off the network as all the hard processing is done at the edge and only data relating to alerts is sent through to the central monitoring units (CMUs).
Video analytics market prospects remains strong in the UK as well as overseas despite or in fact because of the economic downturn. One of the key drivers for this demand in the UK is local authority CMU consolidation which is a major trend now. The CMUs that are left in place, when this wave of consolidations has run its course, will be monitoring several local authorities’ camera systems each. This means that each CMU will be monitoring several hundred more cameras and they may well be doing this with the same or even less staff. UK control room staff are at risk of being overloaded.
We think sophisticated video analytics could be the answer because the software can be configured to alert staff when specific types of incidents or scenarios happen which may be the precursor to (or indicative of) criminal behaviour. In short, analytics can help separate the ‘wheat from the chaff’.
To help Government CCTV system procurement teams select effective analytics solutions, the Home Office Science’s Centre for Applied Science and Technology (CAST) partners with the Centre for Protection of National Infrastructure (CPNI) to maintain and develop the Imagery Library for Intelligent Detection Systems (i-LIDS): https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/115681/intro-to-cast.pdf
i-LIDS comprises a library of CCTV video footage based around ‘scenarios’ central to the government’s requirements. The footage accurately represents real operating conditions and potential threats. The ‘event detection’ and ‘tracking’ scenarios are:
1.Abandoned baggage detection
2.Parked vehicle detection
3.Doorway surveillance
4.Sterile zone monitoring
5.Multiple-camera tracking
6.New Technologies.
CAST runs most of the evaluations, although sterile zone monitoring tests are now outsourced. All the scenarios have a scheduled annual evaluation and manufacturers can submit their test performance scores to CAST to be considered for inclusion in one of the annual evaluations. The latest annual timetable is here: https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/143874/ilids-evaluation-timetable.pdf
These scenario tests are being steadily adapted and improved by CAST. Latest enhancements have absorbed changes associated with use of HD and infrared cameras.
Video analytics software can be evaluated by official testing bodies against official datasets established for each of the above. Some datasets contain up to 50 hours of real world footage. The test provides an accuracy measurement based on the percentage of false positives (false alarms).
Each dataset consists of two or three camera views referred to as stages, and is further segmented into shorter video clips of 30 to 60 minutes. The training dataset is further split into individual events. Each dataset is supplied with a user guide detailing the library structure, user interface and the procedure used to evaluate the systems against the relevant scenario. Analytics software firms wear their i-LIDS badges with pride. And because it is the only globally recognised certification system for video analytics it helps secure major contracts in the Far East and Middle East as well as in the UK.
There do however appear to be two specific drawbacks to the existing tests:
1.They are geared to test CCTV systems not IP-based surveillance systems which are being increasingly adopted, even in IP-adverse local authority CMUs.
2.They are designed to test PC-based analytics rather than edge-based analytics. Both IP and edge-based surveillance deployments are of course on the rise and most major camera manufacturers are now focusing hard on building more intelligence/analytics into their cameras.
i-LIDS is clearly a market-leading testing service which just needs further investment to keep it relevant to newer technologies as they emerge. Was Andrew Rennison not leaving his post as surveillance camera commissioner in 2014 perhaps this would have been something he would have pushed for. But without him who will champion this inside Government?
For further reading
Home Office CAST paper on helping small businesses specify CCTV Systems correctly: https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/142684/cctv-small-business-guidance.pdf
For more on iLIDS follow this link: http://tna.europarchive.org/20100413151426/http://scienceandresearch.homeoffice.gov.uk/hosdb/cctv-imaging-technology/i-lids/index.html
