Megapixel cameras.

Megapixel cameras.

High definition television is the main trend in CCTV development for the coming years. No one has any doubts about this. Moreover, the consumer video equipment sector has already overcome the analog format. Megapixel photo and video cameras have become everyday devices for reasonable money. The consumer is waiting for a qualitatively new level in the CCTV sector as well. However, despite the emergence of specialized megapixel cameras for security systems, their use is still associated with a number of limitations. The nature of these limitations is in the very essence of a megapixel camera as a source of high-resolution images, as well as in the complexity and high cost of creating a system based on these devices.
In this publication, I will touch on some of them.

Sensitivity and image formation speed
Megapixel cameras use CCD (charge-coupled device) and CMOS (complementary metal-oxide-semiconductor) light-sensitive matrices as image formers. The fundamental differences between these two types of matrices are given in the table.

The main factor limiting the use of megapixel CCD matrices is the low charge reading speed and, as a result, the impossibility of ensuring a high image formation speed. For 1.5-2 MP TV cameras, this value is 15-12 fps. The higher the matrix resolution, the lower the image formation speed. CMOS technology, which combines a light-sensitive element and a processing chip, allows for a high frame refresh rate even for 3 MP sensors. At the same time, the cost of a CMOS matrix is ​​lower than that of a similar resolution CCD.
On the other hand, a higher signal-to-noise ratio and high sensitivity of CCD-based cameras make this technology more preferable for security systems. It is also necessary to understand that with increasing resolution, the sensitivity of the matrix decreases due to a decrease in pixel area. The typical sensitivity indicator for megapixel CMOS cameras is 30 times lower compared to traditional analog cameras based on CCD.
However, CMOS technology is improving, and given the growing demands for higher resolution, megapixel cameras based on CMOS matrices are beginning to dominate the market.

Data Transfer Interface
Megapixel digital cameras have been used for machine vision systems for many years, but, as noted above, they have only recently appeared on the CCTV market. The main limiting factor was not only the price of the device (photosensitive matrix, camera processor), but also the lack of a universal data transfer interface that allows connecting several devices and combining them into a system. We are talking about a digital interface with high bandwidth, since the image parameters of a megapixel camera go beyond the analog standard. Unfortunately, the FireWire (IEEE1394) interface used for industrial cameras is expensive and does not allow creating a distributed system. The development of IP networks and the emergence of IP systems led to the use of Ethernet as a standard universal interface for megapixel CCTV cameras. Of course, the image is compressed. It cannot be transmitted in RAW (uncompressed) format, since for a digital representation (4CIF, 704 x 576) even of an analog image the flow is more than 29 Mbps*. For a camera with a resolution of 1.5 MP (16CIF, 1408 x 1152), the uncompressed flow will be 118.8 Mbps. The most common compression algorithms used to compress images in megapixel cameras are MJPEG and MPEG4. They allow you to reduce the amount of data by tens of times. A typical flow of a 1.5 MP camera is 1–1.5 Mbps. This is 2–3 times more than a 4CIF (704 x 576) flow. Cameras with a resolution of 2 and 3 MP (they are currently the most common on the market) have a bit rate of 2–4 Mbps. The volume of transmitted data makes you think about creating a gigabit network for a system based on megapixel cameras, which seriously affects the cost of the solution. In any case, the task of correctly calculating and building an IP network for megapixel cameras is more pressing than for IP cameras with a resolution of 4CIF (704 x 576).
Recording and visualization
Data flow transmission and management is not the only pressing issue. Data must be saved, the image must be decoded for display on the workstation screen. When choosing a solution based on megapixel cameras, it is necessary to take into account the cost of creating an archiving system and the performance of workstation computers. Using camera bitrate data, it is easy to calculate the amount of disk space per day for a high-definition system consisting of, for example, 16 cameras with a resolution of 1.5 MP. The figure is impressive — more than 2 TB!!! Of course, you can reduce the number of frames and, as a result, the bitrate. At the same time, the information content of the system will also be reduced.
However, the question arises: why such sacrifices?
The flow calculations for the 4CIF and 16CIF formats, as well as the figure of the ratio of the two formats, are not provided by chance. The reader can evaluate the efficiency of using a megapixel TV camera in comparison with a camera of analog resolution and answer the question often asked on the pages of industry magazines: can a megapixel camera replace several analog cameras? Unfortunately, except for excellent image visualization on monitors with a large screen and high resolution, cameras with a resolution of 1.5–2 MP do not have any advantages. The resolution is definitely not enough to use the capabilities of digital scaling (ZOOM).
However, I would not like the reader to get the impression that megapixel cameras are not practical for use in security video surveillance systems. Currently, cameras with a resolution of 3 and 5 MP are appearing on the market, which are really capable of changing the principles of video security and the quantitative parameters of the system elements. Reducing the number of cameras with a resolution of 4CIF, the costs of digitalization channels, data transmission and directly installation work can compensate for the costs of building high-speed networks and data storage arrays.
Today, the most relevant use of megapixel cameras is not in IP systems, but in hybrid systems based on DVR. Specialized hybrid video recorders allow recording images from both analog cameras and megapixel IP cameras. Surveillance areas that require special detail can be equipped with megapixel cameras connected to a hybrid DVR, while other areas are equipped with conventional analog cameras. Megapixel cameras must be distributed between recorders so that no more than 1–2 cameras are connected to each of them. If it is necessary to organize remote workstations, it is important to correctly calculate the client's request traffic and select the configuration of workstations.
In the future, improvements in data processing, transmission and storage technologies will allow creating a truly effective high-definition system.

*Note: Calculation of the data volume from one camera with a resolution of 704 x 576 pixels: 704 (horizontal dots) x 576 (vertical dots) x 24 (digitization, bits per color)/8 (representation of bits per byte)/1024 (representation of bytes per kilobyte) x 25 (frames/s) = 29,700Kb/s.