Consulting Engineers                      

CCTV Cameras used on Perimeters.

1.      Example for placing cameras on a Perimeter:

a) General: Herewith a hypothetical Site.  Please read the addendum and open the lens calculator (an Excel program) to understand why some of the choices are made:

Normally the image from a camera is usable from 30% to 100% of the design distance.  Further away the detail in the image gets too small to be useful and nearer than 30% too small an area is in view.  This must be taken into account when covering an area or perimeter. Therefore camera views must overlap with one camera covering the next camera.  E.g. If the design is for the cameras to each survey a 100m stretch of the perimeter,  each camera’s effective range is the 100 meters from 40 meters to 140 meters (30% to 100%) away from the camera.  The required details must therefore be detectable at 140 meters distance and at 40 meters distance the full required surveillance area must be covered.

For a viewer (security officer) to be able to estimate the position of an occurrence, the distance to between the occurrence and a cameras must be clear.  Cameras must be mounted on poles more than 3 meters high (this also makes them less accessible to possible intruders).  It helps to install clearly marked signs within camera view (e.g. “C 50m” which would indicate that the sign is 50 meters from camera C)

Corners can become a problem. A solution is for a set of cameras, 40 meters apart, looking at one another and then on to the corners.  See sketch above – please contact me if this sounds obtuse – Hendrik).

b) Method:

Step 1:  Read the “General Information on CCTV Cameras and Systems”. The addendum to this document. 

Step 2:  Obtain a scaled layout of the entire perimeter.  We will use the above sketch as an example.  Assume the perimeter consists of two fences spaced 4m apart.

Step 3: Select the cameras positions at convenient places, where possible so that each have a view of about 140 meters.  In an area with dense fog or rain the distances can be reduced to 50 or even 30 meters apart.  Longer ranges can also be used.  Generally 100m is a good compromise distance, which means the camera must cover the area up to about 140 meters and has proven successful over many years.  Remember to allow for the overlap i.e. cameras must overlap 30% of the distance.

Note:  Typically the cameras will be installed on 4 meter poles just inside the inner fence.  To economise the cables will be strung on the inner fence in conduit.  25mm conduit for the co-axial cables (1,5m above ground level) and 15mm conduit for the power cable (1m above ground level).  The cameras will be pointed so that the bottom of the image shows the ground at 30% of design distance and the outside border of the image shows the outside fence at 30% of design distance.

Step 4:  Choose a camera type for each position.  Draw them on the map showing the direction in which each is pointed.  The main decision is the size of the sensor – choose economically taking the resolution and sensitivity into account. The larger the sensor the better it will work in low light conditions.  Ask advice – cameras are now so reliable that the make is of lesser importance. 

Note:  BLC (Automatic back light compensation) is important.  B&W (Black and White) cameras is generally better for night vision and often is very sensitive in the infra red spectrum, which can allow better penetration of fog and rain as well as infrared lighting at night.

Step 5: Choose a lens:  The decision of which camera and lens combination to use for each of the cameras in the sketch above is by using the lens calculator available as an Excel program. 

The process is as follows:

Camera A:  View distance 138m.  Perimeter width to survey: 4m.

From the above discussion we know that at 30% of the distance we still want to survey a 4 m wide perimeter area:  138X0.3=41.4 meters.  Input these values in the lens calculator :

Supposing we chose cameras with 1/3” sensors, which are not very expensive and give a good picture both in low light conditions as well as in sunlight.   We decided to use an auto-iris lens with a large lens opening (wide aperture, See wikipedia for details) to use outside (i.e. to allow good images in low light conditions as well as in strong light conditions).  The calculator shows that for the 1/3” camera a 50mm lens is required. 

Insert the values in the bottom part of the lens calculator :

Then look at the lens sizes again, there is a great difference between the required width with the standard lens focal length lenses as calculated on the lens calculator , so choose another  lens size and see  whether this is now acceptable:

Using the 75mm lens gives more details at 138 meters. Even a face start being visible (probably not recognisable yet). To survey 3.8 meters at 30% distance is acceptable since another camera (C) is also overlapping the surveyed area.  You can now calculate on the lens calculator that camera C will survey a width of 8.7meters at 196.6 meters (the 30% distance from camera A) if the same camera/lens combination is used.

All Cameras are selected in the same way:

Cameras B and C are the same as Camera A: 1/3” cameras 75mm lens.  Camera A overlaps the near view of Camera B and the 3.8 meter survey width is again acceptable.

Cameras D and E are the same as Camera A: 1/3” cameras 75mm lens.  View distance 140m.  Perimeter width to survey: 4m.

Camera F: 1/3” camera 50mm lens. View distance 95m.  Perimeter width to survey: 4m.

Cameras G, H, I, J, K, L, M, N & O: 1/3” cameras 75mm lens View distances 150, 165, 144, 144, 144, 154, 130 & 130m.  Perimeter width to survey: 4m.

Cameras P & Q: 1/3” camera 50 mm lenses. View distance 112m.  Perimeter width to survey: 4m.  Actually a 75mm lens would give much better detail.  As they are pointed at one another, the 56m distance is the narrowest survey point and here a full 5.6 m width is covered. And only 10 meters width at the other camera.  So choose a 75mm lens.

Camera R: 1/3” camera, 8mm lens. View distance 12m.  Perimeter width to survey: 9m (The gate area, the gate is assumed to be 7m wide).

Step 6:  Finalise the installation requirement by

a)      listing the cameras and lenses in a matrix,

b)      decide where the cameras will be monitored from,

c)      Position the nodes (see below)

d)     Show the cable runs and type (see below).

2.      Installation Methods:

The installation of cameras round a perimeter (or elsewhere), where the distances to be surveyed are long, requires some decisions on cost effectiveness to be made.

Installed cable cost is often greater than the cost of electronics.  The principle is therefore to reduce cable lengths by installing a number of “nodes” where data is captured and control instructions (if required) are passed on. 

The above hypothetical perimeter can be used as an example to illustrate the principle:

In the case of CCTV the node will typically be a stand alone DVR (Digital Video recorder) unit which can record images from up to 8 cameras.  These will typically also allow control of camera functions (if needed) and VMD (Video Motion Detection).  The size of an internal hard disk (or other recording device) can be selected for size and the storage requirements of images from each camera can be configured.

The node communicates with control computers by means of TCP/IP (a communication protocol also used by the internet) which can be done by wireless or cable (often fibre optic).  The security of the communication is protected by various means.

Power is distributed from the node where a small UPS and power supply is installed in an environment protected container housing all.  The housing is positioned near a camera pole central to up to 8 other cameras – see below.

The video connection between the camera and the node is by means of either twisted pair or co-axial cable (RG59 up to 100meters and RG 6 up to 500 meters and RG11 up to 1000 meters); Power is supplied by a daisy chain between the cameras (12volt or 24volt or 220volt depending on the cameras selected).

Seeing that the cost of the cabling tends to be the more expensive than nodes, the node will typically connect to either 5 or 7 cameras i.e. not use all 8 inputs.

In the example the 18 cameras require either 3 or 4 nodes (7 or 5 cameras per node on average).  They are positioned in possible configurations and the cost calculated for the most economical solution.       As an example:

It would seem that a first configuration immediately yields an excellent solution.  The cable and conduit installed cost can now be calculated and compared to node costs to obtain good budgetary figures for the system cost.

3.      Lighting:

Lighting is discussed in a separate document.

4.      Operation:

The system configuration can be done from central computers.  Many vendors will be able to supply the above type of system.  Although each vendor might use a different set-up procedure the following result is easily obtainable:

Although all cameras are electronically monitored continuously, when no alarm is triggered, less than one image per second is more than adequate for future evaluation. On alarm (whether false or nuisance and whether triggered by VMD or external alarm trigger), images of the cameras viewing the alarm area must record at up to 25 frames/second (5 frames per second are the minimum to be adequate).  Good systems will be able to do pre-alarm recording.  The duration of higher speed recording and the speed are dictated by the size of the hard disk and the duration that records must be kept.  These records are recorded at the node.  Only some of the images are transferred to the monitoring computer.  Only during evaluation are more images or streams of video downloaded from the nodes.

The bandwidth to obtain all required information, even on a worldwide bases, can be controlled and limited by judicious decisions of which images and video streams to download to monitoring control rooms.

Addendum:

General Information on CCTV Cameras and Lenses

CCTV cameras are produced to conform to one of several standards.  The most important are PAL (South Africa, Europe, Asia, Australia, etc.) and SECAM (France, Russia, parts of Africa etc.) both of which standards specify 25 frame/second, while NTSC (USA, Canada, Japan, etc.) specifies 29.97 frame/second.  These standards are not interchangeable but most equipment, not cameras, can usually accommodate at least NTSC and PAL.

Make sure the CCTV cameras and all equipment used for ESKOM are PAL compatible. 

Video can be interlaced or progressive. Interlacing was invented as a way to achieve good visual quality. The horizontal scan lines of each interlaced frame are partitioned into two fields. NTSC, PAL and SECAM are interlaced formats.

Interlacing means that a frame actually consists of two fields (slightly offset vertically) taken 1/50th of a second apart.  If there is movement in this period, the frame (picture) looks as if the picture is out of focus wherever movement occurred (parts of a stationary picture appear to be moving or flashing).  Therefore recording is sometimes done with single fields, doubled and offset to simulate a full frame.  Resolution of the standard 640X480 pixels then is effectively reduced to 640X240 pixels.  In digital imaging, a pixel (picture element) is the smallest piece of information in an image.

Video resolution specifications often include an i to indicate interlacing. For example, PAL video format is often specified as 576i50, where 576 indicates the line resolution, i indicates interlacing, and 50 indicate 50 fields (half-frames or fields) per second.  The difference between the 640 pixel per line standard and 576 is due to information using up some of the resolution.  

In progressive scan systems, each refresh period updates all of the scan lines. The result is a higher perceived resolution – the problem explained above is overcome. The size of a video image is measured in pixels for digital video, or horizontal scan lines and vertical lines of resolution for analogue video. 

The cameras capture images by means of an image sensor which is typically either a charge-coupled device (CCD) or a complementary metal–oxide–semiconductor (CMOS) active-pixel sensor.  These are now typically made in a number of sizes by a number of manufacturers (Sony, Sharp, Panasonic, Philips, etc).  The size of the sensor is measured in inches over the diagonal.  Before 1990 sensors were vacuum tubes of   typically 1” and 1/2” size with the larger allowing the higher resolution and better quality. 

The typical sizes of CCD or CMOS sensors currently are 2/3”,½”, 1/3”, ¼” and 1/6”.     Technology has allowed miniaturisation to the extent that full resolution can be available in all these sizes but each size require a different lens, even if using a standard C-mount attachment (lenses are often specified as C or CS-mount. The 2 standard cctv camera lens mounts. The difference between the two is simply the distance between the lens and the image sensor. C Mount - 17.5mm; CS Mount - 12.5mm).

The 1/6” sensor can be of higher quality than the larger sizes since there is less likelihood of dead areas on the chip, the smaller the chip. One quality measure is the number of dead pixels and sizes of dead pixel clusters on a chip – the quality improvement of chip manufacture has made this less important.    However the larger the chip the more light it can capture. With PAL a full resolution image sensor is 640X480 (640X240 per field).  Note: not all usable due to information being superimposed on the ends of lines and images.   This means that the image can not be “blown up” to see more details (as done in TV series) and if a face must be recognisable, the total image in view can not be more than the width of a double garage (6 meters wide, i.e. 4.5 meters high).  In this case the face would have a resolution of about 20 pixels in width – not much.

The lens size must be chosen to view a specific width and height at a specific distance.    The selection of lens size is dependent on the sensor size.  The size is measured on the diagonal and is actually a 4:3 rectangle and the actual dimensions can therefore be calculated as in the accompanying spreadsheet (the actual picture is about 5% smaller since the information transmission/signalling method cuts of the edges):

The extent of the image, which will be recorded and can be seen on the screen , depends on the camera sensor and lens combination used. Note: Using a lens of the same focal length with a smaller sensors gives more of a telephoto effect but captures less light.  A larger lens aperture captures more light and can be used to overcome smaller sensors. 

The sensors inside cameras (here given as 640X480) are the typical sizes available on the market.  One must know the sensor size and the resolution of any camera used.  The resolution can be lower than the typical 640X840 pixels on cheaper cameras or more for High Definition TV.  In the calculator the calculation can be adapted by inserting other values in the place of the 640 and 480 values.  The more pixels the more details can be seen.  Less than 15 to 20 pixels across a face are regarded as unrecognisable.  I.e. as the picture is enlarged the pixels become visible as squares (larger and larger) and things become even more difficult to recognise.

Back                                                 Download Document                                      Contact us