The Math Behind Megapixels

Jan 1, 2008 12:00 PM, By Paul Bodell

Megapixel camera technology is a fast-growing option for users with video surveillance applications. Yet many users might find themselves in the dark when separating fact from fiction in regard to the technology. In order to make an informed choice, users must weigh many considerations including application goals, cost, storage capability and bandwidth concerns.

Video surveillance goals

Before considering cost, let's ponder why an end-user would even consider going megapixel. In some applications, megapixel cameras may not be appropriate. The first step for users is to define the goals for their video system, which can fit into three categories:

• Resolution and pixels-per-foot

  General Surveillance — These are applications, often live-viewing, in which users do not need detail on recorded video. For instance, the user may be watching a road and looking for traffic jams, but he or she does not need to read license plates. As another example, the user may be looking to see what a crowd is doing, but does not need to recognize faces. The user simply wants to detect intrusions and be alerted when someone is in a restricted area, so he or she can respond immediately.•

• Forensic — These are applications where the user needs to see, record and recognize images, such as license plates and faces, so he or she can go back “after the fact” and determine exactly what happened.

• High Detail — These are applications where the user needs to read a license plate and also read, for example, the model of the car. In a retail or banking application, the user would need to see customer and employee faces clearly, as well as identify the currency in their hands in detail.

By assigning one of these categories to the application, users are defining their resolution requirements. Once a user has defined the type of application they have, and they know how wide of an area they have to cover, they can figure out if megapixel technology is right for them.

Historically, the resolution of older surveillance technology was defined by Total Video Lines (TVL). Over the last few years, the market has transitioned to defining resolution by the total number of pixels, or the smallest single component of an image. This is more objective when comparing products that have different resolutions. So, when an image is listed as 640 × 480, that is actually 640 (horizontal or columns) pixels × 480 (vertical or rows) pixels, adding up to a total of 307,200 pixels, or approximately 0.31 of a megapixel (million pixels).

However, resolution can only be defined by pixels after determining how wide an area the pixels will be covering. How does one determine that? It's all about pixels-per-foot.

For general surveillance applications, users need approximately 20 pixels-per-foot. Forensic applications require about 40 pixels-per-foot, and high-detail applications demand at least 80 pixels-per-foot.

Are megapixel cameras worth the price?

For example, suppose a user wants to cover a parking lot with forensic detail and that the lot is 100-ft. wide. For simplicity, assume that we are discussing the width (the horizontal field-of-view). For a true analysis, users will need to factor in the width and depth.

The formula is 100 ft. × 40 pixels-per-foot, which equals 4,000 pixels. This is what the user will need to cover the 100-ft. area and to recognize license plates and facial detail.

Compression, bandwidth and coverage

The next step is to determine what resolution cameras to use. This is determined by dividing the number of pixels needed to cover the area (4,000) by the number of horizontal (columns) of pixels provided by the camera. If a user uses 320H × 240V resolution cameras (0.07 megapixel) they will need 4,000/320 = 13 cameras (The precise math works out to 12 ½. Then, you round up.) At 640 × 480 resolution (0.31 megapixels), the user needs 4,000/640 = 6 ¼, which works out to at least seven cameras. For 1280 × 1024 resolution (1.3 megapixels) it works out to 3 ¼, so four cameras will surely cover the area, and at 2048 × 1536 resolution (3 megapixels), two cameras will do the job.

Why are so many end-users switching to megapixel technology? The main reason is price. However, conventional thought tells us that a reliable, industrial-quality megapixel network camera is three or even four times more expensive than comparable low-resolution (4CIF) cameras. How can the megapixel camera be cheaper?

Let's compare a non-megapixel installation for a parking lot with some of the industry's highest-level megapixel cameras. According to online research, a high-quality camera made by an established manufacturer that delivers a 640 × 480 resolution image has a manufacturer-suggested retail price (MSRP) of about $350. A high-quality camera made by an established manufacturer that can deliver a 2048 × 1536 resolution image has an MSRP of about $1249. A high-quality outdoor heater/blower housing enclosure has an MSRP of $225, and 100 ft. of Cat-5 cable can be purchased for $20 (coaxial cable is more expensive). Let's also include $100 per unit of labor to install a camera. The charts on page 15 show the math.

While there are many different types of compression — some of the more familiar being H.264, MPEG-4 and MJPEG — there are basically two standard types: frame-by-frame and temporal.

MJPEG is the most popular frame-by-frame compression technique, compressing each image in its entirety. It is widely deployed and easy to integrate. The advantage of this technique is that it enables you to recreate images accurately, and the bandwidth use is predictable. The drawback is that handling each image is not very efficient in terms of bandwidth when there is little motion or activity.

Temporal compression is represented by popular compression methodologies such as H.263, H.264 and MPEG-4. These techniques are widely deployed in applications where the available network bandwidth is limited and low-quality images will suffice. Temporal takes an image called a “key frame” and compresses it in its entirety. For the next few images, the technology only compresses and transmits things that change within the image. With every few images, it takes another key frame and repeats the process.

Want to use this article? Click here for options!
© 2008 Penton Media Inc.

This month in Access Control

• Opening Up About Door Closers
• An Enterprise Approach
• The Framework For Open Systems
• On A Higher Plane
• More from April's issue