We all remember our high school photography class. Unlike many other high school classes, photography class gave you a formula to be successful. Before you were sent wandering the halls of the school in search of a Pulitzer wining photo, they provided several rules to follow in order to help ensure that your resulting photos were at least suitable to hang on the inside of your locker door. Wouldn’t it be great if a similar set of rules were available for security designers? As luck would have it, many of the rules you learned in high school photography class also apply to security design.
RULE #1 – Check your Background
Check your background to avoid clutter or distractions.
This rule seems deceptively easy: avoid busy scenes, keep cameras from facing directly into headlights, avoid privacy issues by not including residential views in the scene and try to avoid having the camera point directly into a sunrise or sunset. The difficult part of this rule comes in the upfront planning. The perfect mounting location from a design stand point may not be optimal when considering the background. The key to meeting this rule is taking it into account early in the design.
RULE #2 – Hold the Camera Straight
Ensure your horizon is level in the image.
In the video security world this is a rule that typically isn’t broken, but is often over emphasized at the expense of other rules. For those of you who have installed a security camera, you understand that this isn’t a trivial rule. Getting the horizon to be horizontal can be a fairly tedious job involving several people to properly accomplish. The problem is installers often spend a great deal of time with this rule, as it is very visible to the end user, and they ignore many of the other rules. So take the time to ensure that your camera view doesn’t have people tilting their heads to one side, but don’t overdo it at the expense of other rules.
RULE #3 – Hold the Camera Still
Hold the camera still to avoid a blurry image.
Attempting to monitor a camera which is vibrating or moving can be a daunting task. Reviewing video from a non-stable camera can also be unproductive, as important video frames can be blurred from the motion of the camera. To remedy this most security designers will use a hardware based approach. This involves the use of physical mounting techniques to keep the camera from moving.
In some cases, however, it may not be practical or even possible to avoid camera movement. In these cases, security designers have the option of hardware or software stabilization options. Software solutions are included in many video management systems, or as an add-on feature. In most cases, they can remove substantial deflections in each direction without any changes to the physical mounting of the camera. This is especially helpful for existing installs.
RULE #4 – Change your Perspective
Change your perspective or angle to greatly increase the effect of the photo.
In security, perspective is tightly aligned with the intended mission of the camera. For detection, it is typically better to have a perspective which is taken from a higher vantage point. This helps avoid occlusions and allows for better understanding of the target’s current track and potential trajectory. However, a high vantage point is not always conducive to identifying a face or unique clothing. When viewed from a high angle, a face may be blocked by a hat and identifying graphics on clothing may be unreadable. The opposite holds true for a low perspective which results in better identification, but potentially poorer tracking and detection. The rule is not whether to have a high FOV or low FOV, the rule is about understanding the benefits and drawbacks of each and using the most appropriate for the camera’s mission.
RULE #5 – Find the Point of Interest
Identify the point of interest and position it appropriately in the image.
We all like to be entertained, but that isn’t necessarily the best approach when setting the view of a security camera. A common mistake is to include “interesting areas” in the field of view (FOV), versus the “area of interest.” For example, consider a situation where a high traffic sidewalk runs next to a facility’s perimeter fence. Often times the camera’s FOV is set to include this high traffic region. However, if the area of interest is not the sidewalk itself, but the region between the sidewalk and the fence, the argument is why the sidewalk would be included in the camera’s FOV. Every movement on the sidewalk now becomes the responsibility of the monitoring guard or security software to detect and assess as a possible threat. The same can be said of roads or areas of water that occur in the scene. If the area is not of interest from a security standpoint, then care should be taken to consider whether the FOV should include the region.
RULE #6 – Allow Active Space for Objects to Move into
When photographing a subject that is moving, you should place empty space in font which the object can “move into.”
This rule is often an afterthought in security design. You may have the camera correctly position to detect an intrusion, but have you included enough “empty space for the subject to move?” This may be important if you need an understanding of the intruder’s intended mission should they breach the perimeter. The same logic applies if you would like to monitor events that are in the vicinity of the perimeter, but do not result in an actually breach. Both cases involve widening the field of view. Just be sure to remember that widening your FOV reduces your detection distances and can result in the need for additional cameras in the design.
RULE #7 – Get in Close
Get in close, using a zoom or a closer vantage point, to get the detail of the subject.
When installing a security camera, “getting in close” is equivalent to “target identification.” This means, in addition to detecting the target, you also want to identify the target: is it male or female, blonde or brunette or Bob from accounting? When using a video surveillance system, you can’t physically move to a closer vantage point, but there are other approaches to “get in close.” Two popular methods are the use of high resolution cameras, which can provide more details to help identify distinguishing target features or the use of PTZ camera algorithms which can automatically zoom in on a target and then follow it. Both options provide for a closer look at the target.
RULE #8 – Take Lots of Photos
Take lots of photos to insure you have a good shot.
One of the best ways to insure you have the perfect shot is to take lots of photos. This same principle applies to security surveillance. Back in the era of film cameras, this could be an expensive rule to follow. How many times did you get back your roll of photos only to find out that you couldn’t see junior’s face or there was a tree growing out of your aunt’s head? Enter the world of digital photography where you can get instant feedback and the cost of taking additional photos is immaterial.
Storage is cheap. Let me repeat that….storage is cheap. In video surveillance, the “take lots of photos” rule is violated way too often. Far too often installations limit themselves to only recording on motion or choose to record at a lower frame rate until an alarm occurs, then switch to a higher frame rate. In terms of insuring you have the perfect image, this is probably the worst thing you can do. Changing a parameter, like frame rate, during an event is a risky proposition. We’ve all played back video frame by frame to get the perfect image of a person’s face or a license plate. In those situations, you want to insure that you have many images from which to choose. A better approach is to record at the highest desired frame rate and quality, then after a period of time has passed, delete unimportant video or decimate the frames to reduce the storage foot print. With today’s cheap storage, it’s better to have many images and delete the ones you don’t need, versus not collecting enough images and missing an important frame.
RULE #9 – Review your Images
Review your images and identify techniques that resulted in great images, as well as, those that resulted in poor images.
One of best rules of high school photography class was probably the requirement that you sit down and critically review each of your photos and take note of what worked and what did not. This rule takes time, but is also a critical rule for security deign. During the design and installation phases, both installer and end user should be reviewing the field of view of each camera. For each view, what are the potential security events that this camera is intended to capture and how will the resulting images from this camera be utilized? A process should be in place by which each camera view is reviewed and approved for potential issues and intended functionality. Ideally, this should happen before poles are put in the ground and trenches are dug. Worse case, this should occur as soon as video images are available. After an event occurs, it’s too late to talk about the fact that that the camera had too much vibration , the view didn’t allow space for the intruder to move into, video frames were missing or you didn’t have enough detail to identify the intruder.
…and finally, the rule you should avoid:
The Rule of Thirds
An image is most pleasing when its subjects or regions are composed along imaginary lines which divide the image into thirds — both vertically and horizontally. Horizons should be positioned along one of the horizontal lines.
Although the “rule of thirds” is pleasant to the viewer’s eyes, allowing as much as two thirds of the image to be dedicated to fluffy clouds in a blue sky does not apply in the security world. There is almost no security application where viewing the sky has any value. However, there are many reasons why you should not look at the sky. These include items such as glare blinding the camera during sunrise and sunset. Another consideration is that you are wasting 33% of your video storage and network bandwidth on the interesting cloud formations.
The only reason to have sky in your image is to give a point of reference and guarantee you have enough room for any vertical detection zones. As a general rule, 10% of the scene is more than enough to give a frame of reference. When considering vertical detection scenes, for example insuring you can monitor for intrusions that come over a fence, even less view of the sky is required
Designing a video security system is not a trivial task. Having a set of rules as a reference to aid in your design may be just the ticket to ensure a successful installation, but where would you find those rules? Perhaps you don’t need to look any further than your high school photography class.