Megapixel video surveillance in practice

IP video surveillance has as many supporters as opponents. This is due to many reasons. These include: the price of the equipment, requirements for the transmission medium, image quality, mobility, possibility of building distributed systems, stability of operation etc.
To overcome the differences between the two types of monitoring systems, the designers can employ available on the market IP cameras with built-in recorders or DVRs with network functions often more advanced that those typically embedded in IP cameras. Such mutual complementarity means that both the systems based on analog cameras and DVRs, and those based on IP technology, may be used interchangeably, depending on investor preferences.

ULTIMAX DVRs operating in multi-server - multi-client configuration
Megapixel IP Camera: IQinVision IQ755 <br />(5.0 Mpix, Day/Night, PoE)
Wireless MPEG-4 IP Camera: Pixord P405W
IP camera IQinvision K1433
recording on CF card
IP camera Pixord K1134
recording on CF card
There is, however, a feature not available in the systems based on analog cameras and DVRs. This is a much higher resolution of the images. Analog television systems (PAL, NTSC) limit it to maximum 720 x 576 pixels (PAL). It is the maximum number of points of the picture.
With digital technology there is no such restriction. Any image size is possible, if the camera supports it. This way the basic resolution of the image depends only on the parameters of the camera.

Comparison of images with VGA and SXGA (1.3 Mpix) resolution
Adjustment of image size to the system requirements
As an example let's analyze a parking lot. To be able to read the registration numbers of the vehicles, we need about 130 pixels per one meter of the real size of the observed object, i.e. a car. It follows that, if the parking lot is 50 meters long and the system is analog (720 pixels), there are needed 9 cameras. In the case of cameras with 1.3 Mpix resolution (1280 x 1024) just 5 will do the same, for 2 Mpix cameras - only 4 are required. It is important to note that higher-resolution cameras also cover a larger area vertically.

Comparison of images of SXGA and CIF resolution
A smaller number of cameras has many advantages. First of all, it lowers the cost and time of installation and maintenance. After all, each camera requires a lens, housing and bracket. Another argument is demand for transmission channels. In digital systems each device forms a channel regardless of its type. Less cameras also means lower operational costs - it is easier to conduct surveillance with a smaller number of points.
Of course, there is another alternative for megapixel cameras - very popular speed dome/PTZ cameras. With adequate lens, it is even possible to cover the area mentioned above with a single PTZ camera. However, such a solution has many drawbacks. First of all, the camera can be set only in one of two ways:
- The first one provides an overall view, without possibility of recognition of some important details, e.g. the license number of a car. Using optical zoom, we can ensure the capability of identifying details, but at the expense of losing the rest of the scene.
- The second way involves activity of the operator who must decide what is interesting at the moment and focus on that area. In the case of fixed cameras, there is always available a complete picture (set of images), with all events recorded by the system.
Another problem is the bandwidth occupied by the video stream and the required disk space. If the monitored area comprises 10 Mpix, this corresponds to 25 D1 cameras or 5 cameras with resolution of 2 Mpix. One D1 frame is approximately 60 kB in size, whereas a compressed image from 2 Mpix camera about 120 kB. As can be seen, in the case of 5 frames per second, the band needed for the set of D1 cameras is:

25 x 5 fps x 60 kB = 7.5 MBps = 60 Mbps

For the 2-Mpix cameras the band is:

5 x 5 fps x 120 kB = 3 MBps = 24 Mbps

In the case of megapixel cameras an important issue is the maximum number of frames per second. Analog systems can typically provide 25 frames per second. But in practice this full speed is not needed or used. It is essential in TV transmissions, but not in CCTV systems. In addition, comparing the number of frames generated by a standard-resolution camera and a megapixel device, we should take into account the total numbers of the pixels transmitted during one second. For a standard CCTV camera it is:

720 x 576 x 25 fps = 10 Mpix/s

For a 1.3 Mpix camera and 8 fps:

1280 x 1024 x 8 fps = 10 Mpix/s

For a top-quality 5 Mpix camera and 10 fps:

2560 x 1920 x 10 fps = 49 Mpix/s

Looking at these results we see that fewer frames does not mean fewer surveillance potential. On the contrary - the information is more detailed than in the case of a standard camera.

In conclusion, megapixel cameras slowly but steadily change the direction of development of video surveillance systems. They "cut the umbilical cord" with the resolutions based on TV standards. They can replace both fixed and PTZ cameras, and the systems are economically competitive with solutions based on "standard" CCTV solutions.