Camera Surveillance … An Experience

I know that it is pretty weird for an IT Manager to write regarding a complex issue like camera surveillance but after a number of complex projects I decided to transfer gained experience to the world including tips and tricks.

Before everything it is very important to know that when you go for a surveillance camera project you need to decide firmly the main purpose of the project, do you need it for security reasons, surveillance reasons or both.

The scope of this article will concentrate ONLY on digital IP PoE camera systems with NVR units connected via LAN, I think that older technologies are not advisable for usage mostly.

If you select the “security purpose” project I suggest to initiate the project to cover “the journey of human beings and any other moving entities inside your guarded location to clearly fully trace these movements and distinguish breaches instantly “ the following points ( in my opinion ) may aid in having a successful solid security surveillance system :

·        You need to guard all entrances / exits / windows ( specially in ground / first floors ) by cameras and these cameras must be installed in an elevated hard to reach positions

·        Every camera MUST be guarded by another camera’s scope for integrated security

·        Camera power ( in mega pixel – we will talk deeply later about this point ) must be able to catch human face’s details in an acceptable quality for local’s authorities recognition otherwise will be rejected as an evidence , this can be very hard to decide taking in consideration 0 LUX condition ( night time ) or for fast moving targets ( outdoor installations )

o  From my field experience an acceptable face recognition images can be recorded at different ranges depending on camera in-scope , camera producers can give more detailed information about this point as there is no standard among manufactured for this specific issue BUT still it is very advisable to have a demo for the camera units before buying ( demo must be done at least for continuous 24H and must test both day and night conditions )

·        All cables used to connect cameras to network MUST be out of reach ( deep pipes in walls , pipes can be steel not plastic in high value guarded assets to ensure sabotage resistance ) 

·        You can always mix number of camera forms to suite your need, for instance you can use bullet and PTZ cameras for outdoor operations while you can use mini-dome, micro-dome and fish-eye cameras for indoor operations, it depends on the need really.

·        Using vandal resistance cameras is vital whenever available

·        All cameras used in security systems MUST have (in my opinion) infra-red built-in illuminators for night vision, extra external infra-red illuminators can be installed at need if the built-in units in cameras do not fulfill the need solely.

If you select the “surveillance” purposes I suggest to initiate the project to cover “ office spaces , restaurants , balconies , roof tops and any other locations that may be shown as points to break discipline in surveyed area “ , the term “break discipline” means that an employee is supposed to stay at the company’s restaurant form 9 AM to 9:30 AM for breakfast time , exceeding this time can be seen as a breach from the management , entering certain restricted locations by non-authorized personnel can be a breach ( in case there are no other access control means in-place ) , the following points ( in my opinion ) may aid in having a successful solid surveillance system :

·        You can still have all the points mentioned for the security surveillance plus adding some more points as follows

·        It is always advisable to use small size cameras despite that it is not a must for indoor installations, small sized cameras are low profile that will not distort decorations especially for in-door installations.

·        You will not be in need to use special metal / plastic housing for cameras mostly for indoor installation thus will reduce project’s cost.

Generally and for the most of the projects you will not be able to distinguish clearly between being concentrating on security or surveillance purposes, this is due to the intersection of requirement between the management’s and the security department’s, a series of meetings need to take place to normalized needs before going to design phase or the project’s cost will erupt.

Step 1:

“It is always the hardest” , after both management and security department unify their goals from the surveillance camera project and have very clear concept and goal(s) for the project we need to start making a long exhausting tour ( mainly on foot ) in the location to be guarded by cameras to decide installation points of the cameras , this can be very hard to conclude because camera manufactures do not supply fixed line of products among the industry , this means that you will find a manufacturer supplying 2,3,5,8,15M pixels cameras while the competition is supplying 2,4,6,9,20M pixel models for the same camera type , in the field this may affect height and distribution of installation points in certain locations for instance .

After this tour(s) end I suggest to have this tour to be done again with every supplier to show in reality the requirements of installation location, after this tour the supplier can be able to recommend cameras from his product portfolio to fulfill the project’s need (I strongly recommend to study all the technical data sheets for every suggested camera / device from every manufacturer / supplier related to the project by responsible technical team inside the organization and not totally offload the issue to the supplier(s) only ).

You will see that for a certain installation location a manufacturer will recommend a 5M camera from his portfolio however the other supplier will recommend a 4M camera for the same location, now comes the major question “How to decide that the 5M is better compared to the 4M? .. both of suppliers says that everyone’s product will fulfill the need “ ,, at this point I recommend to have a real time live demo for proposed models by supplier IN YOUR ENVIRONMENT for a continuous period of time of 24H at least , we do not select cameras because ONLY of its mega pixel power , there are many factors to be taken in consideration like optical sensor’s power , dynamic / fixed focal length ( dynamic is better ) , built-in infra-red illuminator strength , PoE , ability to record on SD cards , number of video streams that can be out of the camera , frame-rate supported per resolution ( many of us can be fooled by having a strong 15M camera and when you inspect its datasheet you will find that it can only use this pixel power at 12 frame per second , this is very low frame rate specially for tracing high speed moving targets in outdoor environment ) , ability to work in low/high temperature environments and many other points that can be taken in consideration when selecting cameras.

I will assume that all tours finished and all meeting took place and now you have selected the supplier and the decided firmly on camera models to be selected for the project, by far I can say that you have accomplished 40% of the project at this point .

Step 2:

At this point you need to start ( with supplier ) making the storage calculations for storing recordings , in my opinion a surveillance camera system without strong dependable recording  systems does not worth the effort , every supplier ( nearly all the major players ) has a software to aid in this calculations but you will need to supply some information to be able to calculate storage needs precisely , this is very important as recording devices ( mainly NVRs ) are expensive , I do not recommend using old technologies like DVR or to use server based recording , server based recording will add extra costs to the project in terms of hardware and OS plus the need of securing it as a server not a single function device ( the NVR ) , you need to know that not all camera manufacturers manufacture NVR of their brand , some of the manufacturers recommend other third party products for recording of their cameras , this is a big supplier-selection factor to be taken into consideration too .

To calculate recording storage precisely I suggest taking the following information for every camera in consideration plus other factors that maybe suggested by camera supplier:

·        Active pixel power to be used

·        Frame rate / second

·        Codec used for recording H.264 , H.265 or any other codecs to be used

·        Bit rate

·        Using motion detection ( motion detection technology make the camera start recording if an object moves in its vision’s scope , if no object moved for a certain period of time it will stop recording , this is very useful feature for indoor office location for after work hours surveillance when there is no one in the office )

·        Recording in monochrome vs color

Now you have the total storage need for recordings per day for the whole system , at this point you need to decide how long you will need to keep recordings ( I did not see organizations need to keep recordings for a period exceeds 30 calendar days but still law requirements may apply in this point ) , in case you need to keep the recordings for 15 days ( to keep the last 15 day’s recording precisely ) we will multiply the total daily storage needs by 15 to get the periodic storage need .

We still not done yet for calculating the storage , we need to decide what RAID system to be used in storage device , most NVR ( Network Video Recorder ) have either RAID5 or RAID6 or support both systems , of course RAID6 is better but will make you buy more HDs or bigger ones but still it is more secure when 2 disks fail at the same time ( RAID 5 is OK when only 1 disk fails in the array at certain point of time , 2 disks fails in the RAID5 system at the same time and the array is gone ).

Now you have the final figure for storage needs for recording.

At this point you need to decide how many location(s) you need to keep copies of your recordings at , locations here means data center locations , YES a data center location , recording devices needs 24/7/365 electricity ON and cooling ON too , depending on the technology used you may also need a server ( other than NVR units ) or more running 24/7/365 for the cameras to access the database system of the cameras , you may need to store recordings in a primary data center location plus another / backup data center location for extra security , through time I found the following points to be important to be taken in consideration :

·        You need to have a single or more data center locations inside the location to be guarded by cameras, this data centers preferred to be totally independent of IT data centers and if not available you can make caged areas inside your IT data center to host NVR racks secured with keys and locks.

 ·        Every data center ( in case of independent data center ) preferred to have 2 doors , first then the second , first one to be opened by finger print access door machine for IT people , second door to be opened with conventional keys by security people , this will ensure that at least a team of 2 members from both departments to be together when entering the data center , a conventional key can be saved in a safe away from the technology world guarded by another human being with log book at the security department , it seems to be complicated but pays back well by raising trust in the system “ no one will be able to claim that a fraud took place easily”.

 ·        At least 2 independent electricity power sources come to every data center from the public grid ( not easy but can be accomplished ) plus third backup line from onsite backup power generator , a power distribution board for every data center have to be available to cycle power sources automatically in case of any power source is down , it is recommended that this board to be secured and guarded by special team members from the security team and to be checked every 6 hours, in case if any of the 3 power sources is down , an indicator ( light + siren ) has to work immediately to notify a 24/7 resident security team member or an engineer in the engineering department that we have a downed source to check what is wrong ( we may have a sabotage attack ) .

 ·        A multiple air conditioning units have be installed inside the data center location with backup air condition unit(s) to work in case that any of the primary units fail or go on maintenance mode, all air condition units have be connected to an automatic operation board to recycle air condition units every 6 hours, you need to keep temperature no more than 18C ( or as recommended ) in the data center’s location as NVR devices really get hot and these are expensive devices we need to keep in-service for long time.

 ·        Inside that data center we need to have UPS units that can keep the power load for at least 30 minutes in case we have a switch time between the 3 power sources , of course the more time we can have ( more than 30 minutes ) the better but will need more money to pay , I recommend to have ONLINE type UPS units from a high grade ups supplier to secure operations for both NVR units and LAN switches for cameras ,, using electronic ATS is very much advisable when you have dual power supplies in your NVRs and LAN switches to ensure non-stop operations even at maintenance time ( if you have another data center location the ATS may not be necessary because recording will continue at the other location even if the first location is down ) , do not go for one brand for UPS units , always blend from at least 2 suppliers to ensure more reliability , I always see that 2 UPS units of 3KVA are better than 1 UPS unit of 6KVA ( think redundancy ) , seek for power factor of .9 ( .9 is better than .7 ) to get more watts from the same KVA, it is no more expensive to find UPS units that can be connected to the camera network and send notifications in case the power input is lost ( extra warning system ) but this means that you need to ensure that required LAN interfaces is available at the UPS units nominated.

 ·        You will need racks to host NVR units and LAN switches , I saw that it is vital to have racks that are deep enough for future expansion , also go for the 42U size from day one to be future proof , having cooling fans in racks are very good plus net doors ( glass doors will reduce the performance of cooling ) , every rack comes with a set of keys for its front , back and side doors be sure to collect them as soon as you get the racks and save them from being theft of lost .

 ·        It is needed very much to have a temperature and humidity meters installed in every data center to keep reading the temperature / humidity inside the data center, keep a monitor outside the data center for a security officer’s check without getting inside the room and if possible you can install an SMS based device to send SMSs to person in-charge in case temperature / humidity raises unexpectedly.

 By this we have prepared the heart of the camera system and now we will go to build the network itself.

 Step3:

Despite that building a network for the camera surveillance system is nearly identical for building it for PC systems we still need to take a number of extra factors into account.

For big guarded locations we need to get a network engineer to design the system taking in consideration that we will have video streaming 24/7 and this is something tough as I experienced , use the best CAT6 cables you can afford ( cables are laid down nearly once in a lifetime and changing them will incur cost if found inappropriate ) , consider 1G link speed per camera ( despite lower power cameras needs 100M link speed ) as a future proof, when need rise for fiber optic links I recommend to go for single-mode fiber cable ( this will give you links for long distances inside your guarded locations ) , fiber optic speeds can be high which is very good for future (still your network designer will recommend the real need at this point) , ensure to have steel enforced fiber optic cables and if to be barren in soil ensure to use pipes ( plastic of metal ) depending on case ( pipes have to be barren on depth of no less than1 meter or as your network engineer advises ) .

Choosing LAN switches can be a hard task because LAN switch along with fiber optic SFPs are expensive components BUT still these components are the heart of your network and you need latency at its lowest to keep switching speed up to the point for not losing video frames and to keep camera monitoring experience of a professional grade (no hopping), selecting LAN switches for camera surveillance system is a process of selecting the most appropriate option not the most expensive one , recommendations from camera system’s supplier have to be taken in consideration for compatibility issues.

As much as possible use SFPs from the same brand of the LAN switches to ensure highest compatibility, use gigabit Ethernet LAN switches even if you do not need all ports to be 1G at the moment, ensue that LAN switch is supporting PoE+ (30W per LAN port) because some cameras may need more power than others specially outdoor cameras (PTZ, heater supported, strong IR illuminators) , make precise calculations to ensure that selected LAN switches can supply the needed power budget for all its connected cameras simultaneously at its highest demand ( this is very important ) .

When building LAN you will need to have distributed LAN points to aggregate LAN connections coming from cameras in certain locations, as much as possible keep this LAN points elevated / hard to access normally ( 5 meters of height or more , can be accessed by hydraulic lift and not easily accessed by ordinary ladders ) , every LAN point has to have its UPS unit that supports the load of aggregated cameras , use 15U racks instead of 9U racks for better ventilation in hot weather areas and for future expansion , ensure that LAN racks are always locked and keys saved , guarded and logged its events ,, it is very recommended to have 2 cooling fans per LAN rack specially in hot areas and always install the LAN rack at no direct sun light , LAN rack has to be connected to recording location(s) with independent link per location to ensure redundancy ( redundancy can be built on multiple levels but the more redundancy you build the more money you pay ,, really big money ) , all distributed LAN racks have to be connected to at least 2 power sources ( 1 from mains + 1 of backup power generator ). 

I saw a case when a company built its camera system and after years they changed the supplier of the camera but needed to use the same LAN infrastructure, this is OK, I recommended to use different IP range for the new system to ensue more security as the new system was been used by another security team and used another independent NVR system, both worked OK simultaneously eventually.

Of course when LAN is built either by internal team or by third part contractor I recommend having an independent testing authority (especially in big configurations) to test the whole LAN configuration for every LAN outlet either for copper or fiber optic and a test document should be created and delivered by testing authority.  

Now I assume your LAN is built and tested OK and connected to your data centers.

Step 4:

Either you are going to install cameras by your team or by third party supplier you need to lead the process’s organization tasks, yes you need to drive it at a very very strong precise way according to the following recommendation from my experience:

·        Co-operate with security department to divide surveillance location to security zones (A, B, C….)

·        Create a unique identifier ID for every camera based on criteria like ( Location ID || Zone ID || Unique number )

·        Create IP range and assign IP address for every camera ( if you have network engineer/team they can do it )

·        Create an Excel file to host all the data for every camera to save the following data :

o  OEM serial number

o  Brand , model number , form factor ( Bullet , Mini-Dome … ) and pixel power

o  Assigned IP ( it is recommended to use fixed IPs rather than dynamic allocated to eliminate the need for DHCP and increase security )

o  MAC address

o  Location ( surveillance zone )

o  Looking direction

o  Connected to which port number on LAN switch and which LAN point connected to

o  Day of installation and entering service

o  Supplier name , PO number , invoice number ( this will simplify maintenance later )

o  And any other information you think that you will need it later during the camera life time

·        Create an Excel sheet for NVR units with relevant information too.

When physically installing cameras with this sheet in hand you can install big number of cameras per working-hour and save extra man-hour cost.

When installing cameras physically it is very good to take the following in consideration:

·        Installing camera on metal surfaces / columns will need ( maybe ) to have isolating non-metal bases for the camera to eliminate power leak problems ( I saw an issue like this that cameras leaked power from its metal body )

·        Installing on marble surfaces is not easy and requires skilled worker to keep the marble safe

·        Ensure that all cables are hidden , this not only to look good but also to reduce sabotage probability

·        In outdoor environment you need to take in consideration weather conditions, for hot weather and direct sun use extra protective enclosures to elongate camera’s production life and also in cold weather to protect it from rain or snow, ensue to have anti-humidity preparations to keep clear non-foggy vision from camera.

Now I assume that you have your cameras are installed, NVRs working OK, everything is documented and tested OK.

Assuming that all cameras, NVRs and LAN switches installed OK and all powered on we need to get back to configure the system logically ( apply permanent IPs , configure camera database , adjust IR systems , adjust motion detection if needed ….. ) , you will need a strong workstation / server ( according to camera system supplier ) to adjust the system and you will need a powerful workstation(s) / server(s) for the surveillance system’s operators / team(s) , adhering to suppliers recommendations regarding devices used for monitoring is vital to have a professional grade monitoring experience and we need to concentrate at the following points to be taken in consideration generally :

 ·        I experienced that to view a big number of cameras simultaneously you may need a big number of cores in the processor (workstation’s processor).

·        According to camera supplier you may have a limit for viewing certain number of cameras at the same time on one monitor.

·        You may need to have strong VGA cards to secure high memory bandwidth for viewing the high quality video streams smoothly

·        You may need to have more than one LAN card and combine them in order to have more network bandwidth for better viewing (you need to ensure that your operating system supports this).

·        Using big screens for monitoring is recommended for 4K and 8K but you need to ensure compatibility with VGA cards in-place.

Despite that you can always use a blend of cameras , NVRs ( servers / appliances ) and VMSs ( Video Management System software ) I still recommend to have all the system from one supplier to ensure full compatibility and get the most of the system at the lowest capex and opex .

Using analytics is one of the magnificent features of the digital camera systems despite that it is rarely used but still it saves time and effort when used the right way.

I found that management should understand that to make the ROI from any surveillance system you need to have a surveillance center working 24/7 and not only to put a monitor at the conventional security office, the best always is to keep the surveillance center covering all surveillance zones in location and make every responsible security officer in a named zone to monitor also at least his fixed cameras working in his zone , we can always blend between fixed cameras and PTZ type cameras for better zooming and concentration needs , a connection between the surveillance center and zones security offices has to be kept via radio / walkie-talkie devices , using wired communications like conventional PBX can be not effective in case of mobile officers or if the PBX is down for any reason .

Camera surveillance system is not the security system, it is ONLY a deterrent tool for the security system so having the adequate number of human security officers is essential, adequate arming level for the security officers has to be taken seriously into consideration, using strong spotlights will secure full color vision for cameras overnight if needed, high fences , barbed wires are good to close and narrowing open zones , using access doors, bio-metric devices will give more access control and reduce efforts on human team in place.  

After the system is fully functional we need to put down a maintenance plan for cleaning cameras , dust , dirt and water droplets can continuously accumulate on the outer body of the camera and reduce the quality of the camera’s vision , a dedicated team must be trained to do this job according to a schedule , every time cleaning to take place the cleaning team has to be escorted by a security officers to ensure that the camera is working OK after the cleaning process ends ( it happens that after the cleaning ends the camera can be moved slightly or the connection to the network to be disconnected , this can be fixed easily be an on-board technician ) .

Checking on NVR devices ( recording devices ) is very much needed at least twice a week , modern NVR devices contains hard disk devices and these devices can fail and needs replacement , you can check on-site or via remote access for the NVR systems .

It is always a very good step to have your security system works efficiently but the more important is to train your security team to learn using the technology to get the most of it, good luck.