NOTE it is acceptable that the default frame rate could be reduced if an alarm trigger mechanism is deployed that would cause the frame rate to increase on activation.
Annex E
(normative)
Detection response testing and acceptability criteria
E.1 General
The detection test poses the question: can the observer easily pick out a suitably clothed target over the range of conditions defined in the OR? Tests are carried out by placing a suitably clothed average person height target, without the operator's knowledge, at locations in the areas under test, as defined in the OR. The operator is then alerted and the time he takes to find the target is measured. For fixed cameras, this could mean the time taken from the initiation of an alarm signal to the point where the operator has correctly acknowledged the presence and location of the target.
Testing PTZ cameras is more complex. The time could be taken from the initiation of an alarm signal to the point where the relevant camera has moved to an automatic pre-set position and/or the operator has driven the correct camera, including zoom and focus, and has correctly acknowledged the presence and location of the target.
The results of each detection test are assessed according to the following scale:
level 3 - target easily seen; it would be immediately obvious and no mistake possible;
level 2 - fairly easily seen; the target needs to be searched for but would not be missed;
level 1 - difficult to see; the target is only found after a careful and lengthy search that exceeds the allowed response time;
level 0 - target not seen at all.
Timing the response of observers from the moment of an intruder alarm to their rejection or acceptance of the alarm is a significant measure of system performance. The acceptable response time depends on the delay expected to be provided by the barrier associated with the alarm system and should be included in the system specification. This timing is the recommended method of setting a threshold for distinguishing between response levels 1 and 2. The results will be, to some extent, operator dependent.
Although, ideally, all target detection scores will be level 3, level 2 is acceptable for effective CCTV system operation. Where test results are not to specification the contractor should rectify the problem. In some cases, system managers might be able to relax the OR based on suitable alternative security measures or a review of the risk.
E.2 False and nuisance alarms
Tests should be carried out without any targets and using false targets to see how long the operator takes to declare a false or nuisance alarm with confidence. Nuisance targets might include items such as a plastic dustbin, dustbin bag, football, etc. The following scoring scale is suggested:
level 2 - when 'no target' condition is declared correctly within the OR response time;
level 1 - when 'no target' condition is declared correctly outside the response time;
level 0 - target declared when no target is present.
E.3 Setting the response time
The acceptable response time should be specified by the client. It should be practical and it may be influenced strongly by a range of factors including:
the delay provided by the barrier, ie, the time taken to breach it;
the area/volume viewed by the camera;
the number of false targets (eg, people who are entitled to be present in the area) which need to be recognised and accepted as present;
the number of screens to be searched;
the need to manually select cameras;
the number of different pictures to be examined, following the alarm, to ensure that the whole alarmed area has been checked;
the need to manipulate any of the cameras with a remote control unit in order to search an area.
Through careful design of the lighting system and camera siting, together with good control room design, large fixed camera systems with accurate and reliable operator cueing can reliably achieve response times as low as one or two seconds. With PTZ cameras the response time may be 30 s or more, depending on pan, zoom and focus speed and the area to be searched. For systems with barriers providing a very long delay for an intruder, a longer search time might be allowed. For one with a comparatively short delay in the barrier any time lost during the search might allow an intruder to pass unobserved through the detection and search zone. Where the required detection times cannot be reliably obtained with 10 % screen height targets then a larger target percentage may need to be specified.
E.4 PTZ response time test procedure
From the OR checklist and knowledge of the site, a location should be chosen where the system response time is to be measured. If, operationally, the camera covering that area may be parked anywhere, the test should be started with the camera set to one end of its pan range. If a camera has a pre-set or datum condition, the time to locate a target from this position should be the system response time. A suitably clothed target should be placed without the operator's knowledge at the agreed position. An alarm signal should then be initiated and the time should be noted for the operator to drive the camera and lens to a position where the presence and location of the target is correctly determined. This test should be initiated with the lens set at minimum focal length and the camera in its normal rest condition.
If tilting of the cameras is necessary to search the whole area, this may dramatically increase the response time. It should not be assumed that targets will be visible to the operator if the pan speed is too fast, the lighting poor or the scene busy. A slower pan or modified search pattern may be required to improve the probability of detecting a target.
E.5 Observer cueing and prompting
It will be necessary to decide the method of observer prompting to be used during the tests. This may require technical solutions such as adding a trip mechanism to the alarm monitoring and control system. This will allow an alarm sector to be triggered or held off while the target is set up.
In some systems, when the alarm is triggered, the observer is prompted to look at each of a number of camera pictures providing coverage of the alarmed zone. The specification may call, for instance, for both sides of a barrier to be viewed and searched. The response time should include selection and search of all of these pictures.
E.6 Detection test locations
The Security Manager with his advisers should define the exact test points using his knowledge of likely attack points and likely weaknesses. It shall be remembered that these tests and the results may represent a key to unlock a security system. They shall, therefore, be treated as sensitive information. For a commissioning test programme, the contractor should be notified in advance of the conducting of the tests but the right should be reserved by the client to carry out further tests if necessary. The more notice given to the contractor the more likely it is that the client's expectations will be understood and taken into account in the design. The contractor should understand that tests will be carried out under worst case conditions for size and contrast, where the lighting is poor, the scene is cluttered or busy or where the background provides little contrast for targets. This gives the opportunity to either re-design in order to produce a larger target or to advise on changes to the lighting in order to raise the contrast of the target.
E.7 Target camouflage
The target should be dressed in suitable clothing to take account of local conditions and what intruders might be expected to wear. Target clothing for commissioning detection tests should be specified by the customer in the tender documents. It may be necessary to try a variety of camouflage materials in order to test fully the system performance. The standard material for the test should be disruptive pattern material(DPM), camouflage. Other colours which should be considered are black green and tan. Test results should be noted for each type of clothing used.
E.8 Tests with moving targets
An effective security system shall be capable of performing well with both moving and stationary targets. Some cameras will perform less well, others may allow a moving target to be more easily seen but perhaps only under certain conditions. Detection tests should, therefore, be carried out with the target moving as well as with it stationary. The results for both sets of tests, using the grading scheme given in 3.5.1, should be noted with the worst determining the limit of performance, unless otherwise stated in the specification. For instance, in an area where it would be unlikely for an intruder to remain stationary for any length of time a pass mark may be acceptable only with the target moving.
E.9 Test conditions
Before carrying out commissioning tests, the whole system should be set up correctly. To provide a performance benchmark, tests should be carried out under conditions that most closely reflect normal everyday use. As far as possible, the operator or observer should occupy his usual position carrying out normal duties. All components of the system should be in their "normal" mode. For instance, cameras which are adjustable should be in their rest position with zoom, focus and iris in whatever is the normal pre-set condition; monitors should be adjusted to their normal settings. Any 'adjustment' to the system performance, even a simple task such as cleaning the monitor screens, which is not part of normal or prescribed procedure, may significantly affect the result of the test. Any factors which might, during normal operation, have an impact on the system performance should be noted. For example, the guard might have to leave the picture display monitors unattended in order to fulfil other duties; the monitor controls might be adjusted to suit particular conditions, e.g, the effect of stray lights from buildings, traffic or sunlight entering the control room. The fact that the observers are alerted to the test will undoubtedly affect the results. Other important parameters will be the weather and the time of year of the test. If the weather is good, allowance will have to be made for loss of performance under poor conditions. The time of year will affect the angle and direction of sunlight. This might make observations difficult during critical periods of high activity such as the rush hour.
E.10 Testing a 'live' system
For various reasons it may be impractical to meet all of the test conditions. For example, in a live system it is assumed that the perimeter intrusion detection system cannot be disabled without the knowledge of the guards. Placing the target in a detection zone without tripping the alarm would then be impossible. In these circumstances, the observer's view of the monitor could be temporarily obstructed while the target is deployed.
E.11 Detection test results tables
Having decided the type and extent of the tests to be carried out, test procedure and results table can be drawn up based on the test specification table. The table and supporting test recording system shall allow for all of the results that will be needed for the performance analysis. General statements such as test conditions might be made at the head of the test sheet or included in the columns. Any special test conditions to be taken into consideration during the test will have been specified in boxes 9 and 10 of the OR checklist.
Table E.1
|
Task area |
T arget |
Observer |
Observer |
Is whole |
Notes and comments |
|
Reference |
Height |
Response |
Response score |
Area |
|
|
Code |
|
Time (ORT) |
|
Covered? |
|
|
And test |
In % screen height or |
In seconds |
Level: 3to0 & |
Yes/No |
Night/day and lighting, |
|
Number |
'not |
& how |
Comments: |
Show |
Weather, |
|
Noted on |
Displayed' |
Observer is |
Very clear, |
Details on |
Clothing, |
|
Site plan |
|
Cued |
Clear, indistinct, |
Plan |
Moving or stationary |
|
|
|
|
Not discernible |
|
|
If there is a requirement for coverage of the same area for more than one purpose then there shall be a separate test row for each.
Where the CCTV is used to verify an alarm state from a PIDS, each zone will need to be identified to ensure that all are tested.Bibliography
http://scienceandresearch.homeoffice.qov.uk/hosdb/cctv-imaqinq-technoloqy/
R82 guideline, for videosurveillance system installation rules, CNPP France (www.cnpp.com)
Image Test Target, VdS, www.vds.de
EN 50174 series, Information technology — Cabling installation
EN 50310, Application of eguipotential bonding and earthing in buildings with information technology equipment
CLC/TS 50398, Alarm systems —Combined and integrated alarm systems —General requirements
EN 62305-3, Protection against lightning —Part 3: Physical damage to structures and life hazard
EN 62305-4, Protection against lightning —Part 4: Electrical and electronic systems within structures
ISO 31000:2009, Risk management—Principles and guidelines
1 Where it is impractical to test all cameras due to the size of the system, 5 individual views or 20 % of the views, whichever is the greatest, should be tested.
