Exclusively Online

Producing an Advantage

Multi-megapixel IP cameras work on time, every time

• By Pierre Parkinson
• Dec 01, 2007

Have you ever looked at a surveillance image and wondered why the image quality is so poor when compared with the clear, detail-packed images captured by consumer digital cameras or even the newest cell phones?

There are two main reasons why consumer digital cameras and cell phones can take such good images when compared with conventional surveillance systems. First, people use consumer digital cameras to take images only on occasion. Because of this, inexpensive sensors and mechanical shutters to control exposure times are more than adequate. On the other hand, surveillance cameras must capture images all the time. To do so, they must be equipped with more expensive sensors and electronic shutters that enable continuous image capture and alleviate the heavy maintenance that mechanical shutters would require if they were to be used in continuous capture mode.

There is a second reason why consumer digital cameras can take such rich images. Once a picture is taken with a consumer digital camera, it remains stored in the camera—to be viewed, it does not need to be sent elsewhere. On the other hand, images captured via a surveillance system need to be sent to a recording device and then re-transmitted to a computer for live monitoring. This introduces bandwidth and storage constraints that degrade the images emanating from conventional surveillance systems.

New technologies, however, are transforming the surveillance industry. Off-the-shelf computer hardware and a new generation of multi-megapixel IP cameras and high-definition NVRs make it possible to deliver the great image quality that surveillance professionals have been asking for in an easy-to-deploy and cost-effective package.

There are advantages to the multi-megapixel IP cameras, and there are a few tips and tools that prove useful in selecting the right camera and lens for an application, as well as determining how much network bandwidth and storage will be needed.

The Advantages
The more pixels a camera is capable of capturing, the richer the image it can produce. A typical 640 x 480 analog surveillance camera can capture about 0.3 megapixels. A 4-megapixel IP camera, by comparison, supports imaging of more than 12 times as many pixels, and a 16-megapixel camera can capture more than 50 times as many pixels. More pixels translate into big advantages for security professionals.

It’s possible to aim a multi-megapixel IP camera at the same scene as a conventional camera and obtain far more detail. For example, monitoring the entranceway of a building with a conventional surveillance camera make it difficult to obtain enough detail in the images to positively identify people. On the other hand, use of a 4-megapixel camera in the same application gives 12 times the image detail. This allows the user to obtain much clearer images and identify people that are coming and going.

Aim a multi-megapixel IP camera at a wider scene and maintain the same amount of detail. For example, while monitoring a parking lot using a traditional surveillance camera, you would need to use a PTZ camera that can zoom in to pick up detail. The problem with PTZ cameras is they need to know ahead of time where an incident is going to occur so it can zoom in. With a multi-megapixel IP camera, simply record the scene and then investigate incidents post-event.

Picking the Right Camera
Picking the right multi-megapixel camera and lens for the job is just as easy as it was with analog cameras by asking help of a camera and lens estimator.

The new variable to keep in mind—which is different from the days of analog systems—is that you have multiple resolutions to choose from. A quick scan of available surveillance camera resolutions on the market today includes 1-, 2-, 3-, 4-, 5-, 11- and 16-megapixal cameras.

Step one in picking a camera and lens is asking how much detail is necessary. To determine the level of detail needed, use a reference image. The image below shows images with different levels of detail, measured in pixels per foot. If an image has 50 pixels per foot, then the image will have about 25 pixels across the face (the average head is usually a bit more than 6 inches wide).

Step two in picking a camera and lens means determining what detail is required and how far the camera is from the scene. It’s also important to note how wide a scene you want to cover.

To illustrate the point, refer to the image below. Here is an aerial image of a parking lot the owner wants to monitor. The key area of concern is highlighted in blue. By using Google Earth, you can measure the distance to the back of the parking lot, as well as the scene width to get a close estimate of the resolution you will have at the back of the parking lot.

The minimum resolution that the owners want for the parking lot is five pixels per foot.

To estimate the right camera and lens for the job, you can use a camera and lens estimator. The camera will be mounted on a 30-foot pole, so the height above the target is 30 feet. The back of the parking lot is 500 feet away, and the width of the scene is 750 feet.

The estimator then shows you that an 11-megapixel IP camera with a 24-millimeter focal length lens will do the job. The 16-megapixel camera can capture 6 pixels per foot—more detail than you need, and the 4.2 megapixel camera only delivers 3 pixels per foot, which isn’t enough.

Below is the resulting image you would obtain when using the camera and lens specified:

From this image, you can digitally zoom into the image to see the level of detail contained in the image.

Detail 1 - This is the full-resolution detail of the back end of the parking lot. The resulting images deliver about five pixels per foot.

Detail 2 – This image is the zoom-in on the closest car in the image. It features detail at about 18 pixels per foot, because it is closer to the camera.

As these images demonstrate, a single multi-megapixel IP camera can monitor a very large area effectively.

Bandwidth, Storage Times and Compression
Calculating bandwidth and storage times for multi-megapixel IP cameras is the same exercise as it is for conventional analog cameras. Determine bandwidth and storage requirements in order to select an appropriate DVR or NVR for back-end storage.

The key variables are the number of images per second, the amount of compression and the percentage of time the cameras are recording.

Images. Generally, the higher a camera’s resolution, the fewer images per second you will need to capture. If you take the example of the parking lot above, the area being monitored is large. Having one or two images per second from the recommended 11-megapixel camera is all that is needed to provide good forensic evidence. A person or vehicle will not move very far across the field of view in half a second or even a full second.

As resolutions come down, the maximum image refresh rate goes up. Currently, 4-megapixel cameras can generally deliver up to 15 images per second, and some 2-megapixel cameras now offer up to 30 images per second.

Compression. One of the primary performance impediments preventing most NVRs and DVRs from properly recording and displaying multi-megapixel images is the compression algorithms that are used. Technologies such as MPEG and H.264, in the current implementations, have an upper image size limit of 640 x 480 pixels, making the use of multi-megapixel resolution cameras with these systems virtually pointless.

Before investing in any multi-megapixel IP cameras, it is a good idea to conduct tests with the planned recording platform to make sure that image quality transmitted from the camera is not being destroyed by the compression used. Leading high-definition NVR vendors are using JPEG2000 for compression of multi-megapixel surveillance images. This is advantageous because JPEG2000 offers the streaming advantages of MPEG and H.264 without destroying megapixel image quality.

A second key advantage of JPEG2000 is that it supports a storage technique called data aging that allows the resolution of older surveillance footage to be reduced to allow more of the total storage capacity to be used for the latest surveillance footage.

Recording. The amount of time that multi-megapixel IP cameras spend recording each day is controlled in the same way conventional surveillance cameras are controlled— through the use of record schedules, motion detection and sensor triggers. It is important to ensure that the NVR or DVR is tested, confirming that the system can support multi-megapixel IP cameras.

Bandwidth and Storage
Once the variables of resolution, images per second and compression are determined, it is then possible to determine the bandwidth required by the system. Leading vendors of multi-megapixel IP cameras and high-definition NVRs provide tools that make it easy to calculate the bandwidth contribution of each camera by simply plugging the variables into an online tool.

Once the bandwidth is calculated, the amount of time that the cameras spend recording can then be added to the mix to determine the amount of hard-disk space required to meet the necessary record times.

Conservative industry estimates show that the majority of surveillance cameras sold five years from now will be multi-megapixel IP cameras, because they allow a user to see more detail with fewer cameras at the same cost. The challenge is figuring out how to make prudent security investments today so money is not wasted on obsolete technologies.

To determine when to start deploying multi-megapixel IP cameras, review the operational requirements to identify where better images are needed in the facility or where a single high-resolution camera can do a better job than multiple existing analog cameras.

In areas with high-security requirements, such as border crossings, airports and ports, the move is already underway to standardize on multi-megapixel IP cameras for almost all applications. Commercial applications for multi-megapixel IP cameras are strongest for monitoring choke points such as entrances and exits, as well as for monitoring large open areas such as lobbies and parking lots.

Having determined where the multi-megapixel cameras make sense, attention then needs to be focused on the back-end DVR or NVR to ensure that that the existing equipment is able to record and manage the higher-quality images. Unfortunately, claims from DVR and NVR manufacturers sometimes exceed actual performance with multi-megapixel cameras. It is therefore crucial that the complete system that will be used for recording the higher-resolution cameras be tested before buying to make sure the image quality is what is expected.

NVRs and DVRs designed specifically for multi-megapixel IP cameras can easily handle conventional analog video streams through the use of video encoders. So, the first step in deploying multi-megapixel IP cameras is to invest in high-definition NVRs or DVRs on the back end that can then be used with both the new multi-megapixel IP cameras deployed in key locations, as well as the legacy analog cameras that can be replaced over time as the need arises.