CCTV Test: Image Stabilisation
The introduction of high resolution cameras has delivered significant benefits to those designing surveillance systems. Higher resolutions enable examination of areas of interest, but such an approach often relies upon telephoto views. The result can be image degradation and loss of detail due to motion or vibrations. Benchmark looks at how image stabilisation – designed to eradicate this – performs in real-world applications.
[dropcap]T[/dropcap]he issues created by vibrations and motion on camera mountings are not new, and to be honest they will probably never be solved. Ultimately, if a camera can be mounted so as to give no motion, then that’s simply the best cure. Indeed, it’s probably the only way to eliminate the issues. Sadly, it’s not always practical.
A balance needs to be achieved between delivering the level of performance that an end user demands, and maximising the camera to deliver stable and consistent images. Video surveillance is often specified in less than ideal environments, so the challenges that face installers and integrators are varied.
Many camera manufacturers now include images stabilisation features. There are different ways of achieving this, as well as a few ‘tweaks’ to other parameters which could give the impression of image stabilisation.
Stabilisation can occur at the lens or the chip, but such methods are typically too expensive for video surveillance. As a result most cameras use EIS (electronic image stabilisation) or DIS (digital image stabilisation). This can use redundant pixels outside of the displayed image, or a cropped element of the image, to allow a digital ‘shift’ to counter image movement.
As with any image processing, correct implementation is a question of balance, and to illustrate this Benchmark put three cameras with the functionality into a real-world application to see what results we achieved.
Axis Communications AXIS 1615
The AXIS Q1615 is an HD1080p camera which delivers streams at up to 50fps. The camera uses a 1/2.8 inch CMOS sensor, and video is compressed using H.264 (baseline, main or high profile) or Motion-JPEG. Multiple streaming is supported. There is a note about the faster frame rate: this drops to 25fps if WDR functionality is employed.
Sensitivity is claimed to be 0.36 lux (F1.3) at 50fps, or 0.18 lux (F1.3) at 25fps. The camera includes the manufacturer’s Lightfinder technology.
The camera supports electronic image stabilisation for use in applications where the camera may be subject to vibrations or occasional motion.
Other functions include WDR Forensic Capture, allowing critical detail to be captured in challenging conditions. There is also a shadow and highlight recovery function. The Q1615 has two operational modes: HD1080p at 50fps without WDR functionality, or HD1080p at 25fps with WDR functionality.
The camera also supports two-way audio, remote back-focus, BLC, AGC, privacy masking, lens barrel distortion correction, alarm I/Os, shock detection, video motion detection, audio detection and sabotage protection.
Edge recording is catered for with an MicroSD-XC card slot. Power is PoE; 12V DC power is also supported.
The EIS functionality is selected from a video processing menu, which also includes barrel distortion compensation. It’s an either/or choice, but this is no bad thing because EIS will be used with telephoto views and barrel distortion with wide angles.
The EIS option is selected via a radio button. There is no adjustment of the actual function. However, the menu does include a slider to set the focal length of the camera. It is also necessary to select the lens being used from a menu.
The Q1615 is supplied with a 2.8-8mm varifocal IR corrected P-Iris megapixel lens from Fujinon. This then is selected by default, but the menu does include a number of other specific lenses as options. There is also a setting for a generic auto-iris lens.
With the specific lens selected, you then need to match the focal length of the lens with the menu slider. It’s a bit hit and miss as you can only guess at the focal length being used. By opening a second window for a live view, the various settings can be tested easily enough. The hard bit is getting the camera vibration or motion to occur at will! In truth, if you can the problem occur, you can probably eliminate it too!
Once set up with the correct focal length selected, the Q1615 does cope well with slight vibrations and motion. There is a minimal degree of degradation in the image, and the effect that makes it look as if the image is ‘flowing’ which is often associated with EIS isn’t obvious.
With a higher level of motion caused by climatic conditions, or with bursts of higher level vibration, there is a slight tell-tale ‘jerk’ at the start of motion but the processing soon recovers and delivers a consistent image.
General set-up of the Q1615 is very good, and overall video performance is at a very high level too.
The DH-IPC-HFW8301EP is a networked day/night camera which delivers 3 megapixel, HD1080P, HD720p and standard definition streams. It utilises H.264 and M-JPEG compression, and triple streaming is supported. Sensitivity is claimed to be 0.1 lux. WDR has a 120dB range.
The camera features image stabilisation – Dahua bills this as ‘anti-jitter’ – and it is designed to eradicate any blur or image degradation caused by vibrations or motion of the camera.
Other features include integral infrared LEDs for illumination; these have a smart function which adjusts power to eliminate potential white-outs. The smart codec also includes smart detection. The camera supports alarm I/O, privacy masking, two-way audio, noise reduction and edge recording via a MicroSD slot. Power is PoE or 12V DC.
The anti-jitter functionality is selected from a camera functionality menu. Selecting it does reduce the image area, as it does with most cameras, but it seems a little more noticeable on the Dahua unit. There is no other required adjustment.
Once activated, the image doesn’t take on a processed look. However, even with slight vibrations and motion the camera seemed to over-compensate. It was almost as if the processing assumed that movement would continue and therefore shifted the image too far in the opposite direction, with the inevitable effect that the shift back would also therefore be over-emphasised. The result was that even smooth motion was amplified and became jerky and harsh. This was specifically noticeable in scenes with defined edges.
In the location with a higher level of motion, it was debatable whether the image quality was better with the ‘anti-jitter’ function disabled.
General set-up of the DH-IPC-HFW8301EP was hindered by a lack of documentation or utilities. These are not available from the Dahua website, and eventually found reference to the set-up process on a DIY website.
Samsung Techwin SNB-6004P
The SNB6004P is an HD1080p camera which features the WiseNet III processing engine. This incorporates the Smart Codec, which combines IVA, video motion detection, tamper detection, face recognition and audio detection.
The camera streams at rates of up to 50ips, using H.264 and Motion-JPEG processing.
The camera makes use of DIS (digital image stabilisation) to eliminate image degradation which can occur if vibrations or camera movement occur.
The intelligent video analytics functionality includes elements such as line crossing and object appear/disappear. When a face is detected, the smart codec uses variable processing to ensure that the quality delivers a positive identification.
The camera uses a 1/2.8 inch CMOS sensor to deliver HD1080p, HD720p and 1.3MP streams. Frame rates of up to 50ips are achievable when using H.264.
Other video-based features include noise reduction, defogging, privacy masking, focus control, two-way audio, edge storage via SD cards and alarm I/Os. There are also all the usual suspects such as AGC, BLC, etc.. Power is PoE or 12V DC/24V AC.
The DIS feature is selected from the camera functionality menu. It’s a simple radio button and there is no other required adjustment. Once selected it’s important to ensure that you scroll to the bottom of the menu page to apply the change.
Once activated, the image – when still – retains its sharpness and detail, and the processing isn’t obvious. With slight levels of vibration and motion the camera generally compensates well, although you do get the impression that the image is slightly ‘fluid’. It sounds odd, and certainly does not make the footage unusable.
With the camera moved to the location with a higher level of motion, the SNB-6004P still coped well. However, the image retained the fluid look when movement occurred, and whilst still usable it was obvious that you were looking at a processed image. The footage took on a slight shimmer; this was both uniform and controlled. Whilst detail was preserved, the effect was slightly off-putting.
General set-up of the SNB-6004P was good, with simple menus and a supplied utility. Also, overall video performance was good. As with the other cameras, using multiple processing features with DIS can lead to some unpredictable results.
The Power of Demonstrations
It is not uncommon for manufacturers to put on impressive demos in order to highlight how their image stabilisation performs. Many of these demos are impressive – we know, we’ve seen many of them – but the real question is how well do they relate to the real world?
The first thing to consider is what you expect the camera to be facing. Typically, vibrations and motion will be caused by an outside effect, and will occur randomly and sporadically. They could be generated by environmental conditions or by passing vehicles. Road traffic, low flying aircraft and trains can all cause vibrations in a building’s structure, as can heavy machinery, civil works, building maintenance, etc..
Many demonstrations use continuous motion, typically increasing in speed and magnitude in a fairly uniform manner. This obviously has no similarity to what will occur in the field. Others use constant uniform vibrations. Again, this doesn’t reflect actual installations. If an installer or integrator knowingly mounts a camera where constant vibrations or motion occur, then they’re probably not up to the job in the first place!
It is also important to realise that EIS, DIS or whatever terminology is being used can help reduce the amount of image degradation. It will never eliminate it.