DORI (Detection, Observation, Recognition, Identification) how to measure how much detail and how far a surveillance camera can see?

The DORI (Detection, Observation, Recognition, Identification) standard is based on the IEC EN62676-4: 2015 International Standard. Unlike DRI (also known as the Johnson Criteria) for thermal cameras, DORI is designed for visible light surveillance cameras and uses specific measurements in PPM (Pixels Per Meter) to define what Detection (25PPM), Observation (62PPM), Recognition (125PPM), and Identification (250PPM) are for standard color/visible surveillance cameras . By using these PPM values, it is possible to select a specific camera sensor/lens combination and verify that it will provide the performance needed in each application based on the criteria given above. Dori White Paper Link

The DORI (Detection, Observation, Recognition, Identification) standard is based on the IEC EN62676-4: 2015 International Standard. Unlike DRI (also known as the Johnson Criteria) for thermal cameras, DORI is designed for visible light surveillance cameras and uses specific measurements in PPM (Pixels Per Meter) to define what Detection (25PPM), Observation (62PPM), Recognition (125PPM), and Identification (250PPM) are for standard color/Visible cameras . By using these PPM values, it is possible to select a specific camera sensor/lens combination and verify that it will provide the performance needed in each application based on the criteria given above.

The difficulties in image quality comparisons.

There are a great many variables that affect the image quality of long-range cameras. Some of these variables are straightforward and measurable, but many are not. Likewise, some of these variables are controllable while others are not.

First off, there are the physical components of the camera. Starting with the build of the lens, each lens will vary in how much it magnifies the image (which is always specified), how much light is passed through the lens (which is sometimes specified), and how sharply and accurately the lens focuses that light (which is rarely specified). These specifications not only vary greatly from model to model but even slightly from lens to lens of the same model. Then you have the image sensor that captures that focused light. These vary by resolution (which is always specified), sensitivity (which is specified but not measured equally), sensor size (which is vaguely specified based on approximate 1950s equivalents), image processing (which varies by manufacturer) and color accuracy (which is never specified). So just in the physical components of the camera there are a lot of compounding variables that can result in two cameras with exactly the same specifications having different levels of performance.

Then you have atmosphere and weather. Even in the best conditions, the air you’re looking through is made up of molecules that degrade the image by hindering the transmission of light. Humidity, rain, snow, dust and fog will reduce the image contrast, and atmospheric turbulence from heat variations throughout the scene will cause distortions in the image. The further you’re looking, the more these problems are amplified, and these factors differ not only from location to location, but even minute by minute.

Note that these atmospheric variables affect all surveillance cameras. All cameras are designed to capture a specific range of wavelengths (such as VIS, NIR, SWIR, MWIR and LWIR). Any two cameras using the same wavelength of light will suffer the same atmospheric limitations. It’s important to keep in mind that atmosphere and weather will affect real world performance more than what is seen in most demo videos, which are often picked from moments of ideal atmospheric conditions.

Pixels Per Meter

For comparing cameras and setting desired performance levels, we’ve found the simplest method is to use is pixels per meter (PPM), which is a measurement that defines the amount of pixels of definition across a 1m width at a specified distance from the camera. It’s a simple measurement that provides a fair comparison of the long-range magnification capabilities of each camera and avoids the essentially useless X-factor numbers that are misleading to customers (see sidebar). It is also an objective measurement of detail that can be agreed on by multiple parties. Pixels per meter is calculated using two measurements:

(The examples here simulate the amount of detail if you were to digitally zoom

into the image. Please note that these image simulations assume optimum

imaging conditions, however many factors such as atmospheric conditions,

heat waves, available light, subject motion or camera shake can degrade

image clarity, and most of these issues are amplified at longer distances.)

Field?of View

The field of view (FOV, also called angle of view)?is the width of the scene that a camera detects on its sensor. It is determined by the focal length of the lens in relation to the sensor size. Longer lenses?or?smaller sensors produce narrower fields of view, while shorter lenses or larger sensors produce wider fields of view.

A smaller field of view means that a camera is more “zoomed in” (to use a term that most people are familiar with). For example, a?camera with a 90° horizontal field of view (HFOV) will see a 1000m wide section of a wall that is 500m in front of it. If you then?adjust that camera’s HFOV to only?1°, it?will?fill the screen with?an 8.7m wide?portion of that same wall. This second “zoomed in” field of view?is what customers are looking for when they want a camera that can see a long distance. They want a narrow field of view.

8MP 49X Optical zoom + digital zoom: Video Link:

The other contributor to long-range camera performance is the sensor resolution. This determines the level of detail within a camera’s field of view. For example, using the 1° HFOV result from above, a newer HD?sensor and an old analog CCTV sensor will both produce images that fill the screen with 8.7m of the wall. The analog sensor has a horizontal resolution of 640 pixels, which means it displays 640 segments of detail across that 8.7m scene. This works out to 73.5 pixels per meter (640 ÷ 8.7). The HD sensor on the other hand, with a horizontal resolution of 1920 pixels, provides 3 times that level of detail with a value of 221 pixels per meter.

The images below show the importance of taking resolution into consideration.?Using the same hypothetical cameras from the previous example, a North American license plate would take up 3.5% of the screen width on both cameras (it would be the same field of view),?however?the older analog sensor would render that plate with only?242 pixels (22×11),?while?the HD sensor would render it with over 2,200 pixels (67×33). This distinction is the difference between a blur of pixels and a clearly readable plate.

Specifying PPM

One of the difficulties that arises when trying to specify detail requirements is that everyone’s idea of adequate detail is different. The problem with using terms like “recognize” or “identify” is that people get an idea in their head of what level of detail that would be and it often differs.

Since perceived detail levels vary greatly, we use the pixels per meter (PPM) measurement to specify the amount of detail needed for each of our clients. We highly recommend specifying your camera performance requirements in PPM, as it makes things clear for everyone involved. Shown below are examples of a few different PPM levels; we’ve also created a document that can be sent out upon request that displays a wider range of detail levels.

An International Standard

If you’re unsure of what level of what PPM levels to specify and are wondering what others use, we refer you to a widely used standard for detail level requirements in visible imaging surveillance systems.

The DORI standard was developed by the International Electrotechnical Committee (IEC), European committee for electrotechnical standardization (CENELEC) and British Standards Institution (BSI). The IEC EN62676-4: 2015 International Standard defines specific levels of detail for Detection, Observation, Recognition and Identification (see details on following page).

Whichever PPM values you decide on, using specified PPM values makes it possible to select specific cameras and calculate their PPM performance to verify that it will provide the performance needed for your application. The great thing about PPM is you are defining the net performance of the system without having to understand the indivudal variables and inputs such sensor size, resolution, pixel pitch, focal length etc. PPM like gas mileage provides a simple unified metric for how much detail you need or what a surveillance camera will provide at a given distance.

Just scratching the surface

Infiniti wrote this white paper in an effort to explain the rating systems we?use while also giving end users an idea of the actual performance they can expect from their cameras at the required ranges. That said, this information just scratches the surface of the many considerations that go into building long-range surveillance cameras.

In addition to standard visible light cameras, Infiniti also offers thermal infrared imaging, active IR illumination, SWIR, and other sensors such as LRF, Radar and more. We do not limit our customers to any one technology; rather we custom build solutions that typically use multiple sensors depending on the project.

There are also many additional aspects to consider when building a camera system, such as cooled vs uncooled thermal, size and weight, stabilization, connection and control, environmental and ingress protection, international trade regulations and more; all of which need to be considered alongside budget and availability.

At Infiniti we excel at designing customized systems to suit your specific needs. We can examine your situational requirements, location conditions and project budget and recommend the best system (or systems) available within that budget. We are also experienced in configuring multi-camera systems complete with multichannel recording, wired or wireless networking, access control, radar integration and more.

Note: Thermal IR Infrared cameras use DRI Johnson Criteria and rated very differently though they both use the words detection, recognition, identification Infiniti optics has a white paper on that topic as well.

https://www.infinitioptics.com/