Stanislav

Your analogy is incorrect. The heat source is the ambient air and the sun, not the camera. This is convection and radiation heat transfer that is entering the camera housing, not exiting. Plastic is actually a better insulator to help prevent heat transfer. I don't agree. The camera electronics produce the heat. The temperature inside the camera box is always higher than the environment temperature. The higher environment temperature is the higher temperature inside the box. Then the temperature inside the box exceeds max allowed one, the camera will be overheated. The difference between box temperature and environment temperature depends of camera power consumption and heat resistance, which depends of box material, form, wind speed. Additional heat is produced by the sun. The darker camera color is the more sun heat it collects. To prevent overheat it is better to place cameras in shadow, use anti-sun shields, use heatsinks, white color cameras in metal boxes. Also you should decrease power consumption, don't use build-in IR, use minimum allowed supply voltage..

Thank you for your appreciation. Here is a way how to calculate the motion blur of an object with known velocity with known shutter speed without software. 1. Calculate a path (in meter, foot ) passed by the object during the exposure time (S=V*t) where S - the path V - object's speed t - exposure time 2. Calculate projection of the path on the plane of the field of view Sp=S*cos(a) Sp - the projection of the path a - the angle between the velocity vector of the object and plane of the field of view. 3. Then with known real object sizes, we can compare the Sp with object sizes on the screen. We can also compare Sp with spatial resolution (pixel density) and estimate how many pixels will be brurred. We can also model the motion blur in some photo editing software. But for this calculation we need to know the exposure time which depends of many factors: camera sensitivity and camera setting, lens aperture, illumination.

Please see my article Video surveillance of moving objects. I hope it will be useful.

Camera noise at low light can be a problem for motion detectors. To reduce noise you should increase light, use more sensitive cameras or cameras with noise reduction function, but they are expensive. Try also cameras with low maximum AGC gain. In low light such cameras produces dark images but without noise. Some cameras has adjustable AGC so you should set LOW AGC. After changing camera you should readjust motion detector sensitivity.

Camera noise at low light can be a problem for motion detectors. To reduce noise you should increase light, use more sensitive cameras or cameras with noise reduction function, but they are expensive. Try also cameras with low maximum AGC gain. In low light such cameras produces dark images but without noise. Some cameras has adjustable AGC so you should set LOW AGC. After changing camera you should readjust motion detector sensitivity.

Rory, as your next step you should buy a LUX meter and measure objective parameters of illumination and sensitivity as well as IR light intensity. Using objective parameters you can compare different cameras and IR illuminators with good accuracy.

The Perspex 962 light transmission curve seems interesting. It transmits 90% in IR range and close visible range. Did you test it with 850nm and 940nm IR LED? It must decrease red glow, but practice experiment is needed. You can use several layers to increase the filter effect. 6 layers will decrease 840 nm power in two times, thus it becomes like 940 nm in efficiency. It would be interesting to compare red glow between 840nm+6*Perspex 962 and 940nm clear with the same power consumption. These curves don't cover visible range. The side peak in visible range is too small therefore it is invisible in the spectral distribution curves in real scale.

I just checked UF100 spec. It has halogen bulbs, no LED. There is another path in theory to make illuminator less visible - we should increase shining area. The more the shining area is the less brightness of the area when radiating the same power. IR LED chips are small but irradiate high power.

Actually 850nm and 950nm IR are completely invisibly for human eye. The "red glow" is visible because of little side peak of radiation of IR LED within the visible range. Thus, in theory, using special IR filters which cut visible range it would be possible to reduce the red glow. Such filters are used in photography, but they are expencive. There are filters of different density. I suppose the same filters are used in IR illuminators with incandescent lamp (like UF500). Did anybody try such filters with IR LED illuminators?

The difference depends of spectral response curve of image sensor and difference in reflection factor of the target. My accurate laboratory test with special calibrated IR LEDs (image sensor=SONY ExView HAD, target= human body) shows what 950nm is 43% in efficiently relative to 850nm. With other image sensors the factor is in the limits of 40-50% Well Im only letting you know field tested results, for 10 years. I am glad that concerning ratio 850/950nm our results are practically the same But concerning White light/IR there is a difference Thank you for your appreciation. Please contact me in case of any questions. Cena odinakova v Rossii i za rubejom.

Just curios Did you use white level (IRE) of video signal as reference point ? Can you share more info about testing procedure / Equal bright of the same pixel of the image on the screen when illuminated by different calibrated LEDs was used as a reference point. AGC was turned off. The difference was calculated as a ratio between irradiances (watt/m2) of 850nm calibrated LED and 950 nm calibrated LED, when the same pixel on the screen had the same bright. CCTVCAD Lab Toolkit was used to measure bright of the pixel. Here are detailed descriptions of my testing procedures.

The difference depends of spectral response curve of image sensor and difference in reflection factor of the target. My accurate laboratory test with special calibrated IR LEDs (image sensor=SONY ExView HAD, target= human body) shows what 950nm is 43% in efficiently relative to 850nm. With other image sensors the factor is in the limits of 40-50%

I'm building some IR devices, just to play around. Found some nice tight optics and the same IR units that Bosch uses in their newer units. Should be fun. The numbers above sounded maybe a little theoretical, it can be adjusted based on published numbers from Bosch Raytec etc. Going to try to get there, but I want to do it with IP cams so we shall see. I'll report back after I get something built. Decided to lay out some circuit boards. The only thing that nags at me is the HID unit, since if they can do a steep cut at 850 with no red glow as they claim then they might utilize the sensor a little better. Some of the plots I've seen show the CMOS is closer and 940 vs 850 than CCD. Time to see what can be done. Yes This factor "4" is based on theory. In practice other factors influence on result too. Such as dependence between optic loss and beam angle (the less angle the more loss), repeated rereflection on the scene, which more visible with wide angles. These factors decrease influense of beam angle. Therefore I wrote "up to 2 times". But the theory is good too. We compared modeling in VideoCAD with field tests of 7 different IR illuminators of different manufacturers and obtained very close results for 4 of 7. See the image. Blue markers indicate modeled distances, red markers indicate measured distances in field. 4 illuminators from 7 were modeled very good. If you are really interested see our discussion of these tests (in Russian!!) here you can find detailed description

http://www.theiatech.com/calculator.php Here is the thread, you are looking for, but the calculator is unaccessible viewtopic.php?f=6&t=20101

But in many reality cases you can't achieve long distance of illumination in spite of enough IR power and camera sensitivity. If some bright object arises in front to the camera, then camera sensitivity will be limited to prevent white level limiting. In this case far objects will be dark. Actually it demonstrates lack of dynamic range. You should provide balanced illumination of the whole scene. Some improvements can be achieved by changing camera setting (level).

Some laws which can be useful The max range depends of square of IR intensity, wavelength. lens aperture and type of video sensor (IR sensitivity of video sensor). Axial IR intensity depends on consumed power and greatly depends on beam angle. For example, changing beam angle from 20 deg to 10 deg can increase max IR intencity up to 4 times, that increases max distance up to SQRT(4)= ~ 2 times. The same IR intensity of 850nm is in 2 times more effective than 950 nm, it increases max distance in SQRT(2)=1.4 times. Using box camera with F1.2 lens instead of mini camera with F2.0 lens increases sensitivity in 2.8 times and maximal range in 1.6 times. B/W camera with SONY ExView sensor is the most sensitive for IR. However TDN cameras with the same sensors are good for IR too, its IR sensitivity is worse in 1.5 times approx. To get the maximal distance you need to maximize all these factors. In case of the resulted distance is not good, you need more than one illuminator. Additional illuminators can be placed in other places not obligatory near the camera.

See my discussion with Rory about White LED vs IR. There is some inequality between theory and practice. I wrote the question to RayTec but didn't receive any answer.

The difference between 540 and 700tvl is not big in reality. Right camera positions, focal length of lenses, lighting are much more important.

Visit your local electronic store. For example in our stores there are tens of different car recording devices with different resolutions, outputs and other options. Some of them can record speed, GPS data, acceleration, sound. As a rule they are mounted on the wind screen or rear-view mirror. They are newfangled auto accessory, some my friends use them. But I plan to make special bracket to mount hight-quality HD camcorder with good optical zoom in my car for traveling. Yesterday in a store I seen a special glasses with 5MP camera for active sport and another waterresistant one for diving:)

What do you imply by the "light output"? The "light flux"="Lumen Output" is shown in any Raytec illuminator's spec. For example, RL200 has the Lumen Output =4700 lumen. A Tungsten Halogen lamp of 300Watt has close Lumen Output. But RL200 consumes no more than 80Watt according to the spec. Thus the Light Efficiency Factor of the RL200 is better than the Halogen lamp in 3.75 times (300/80) and this Illuminator must be good. For color cameras with IR cut filters the "light flux"="Lumen Output" in lumens is important only. For B/W, TDN and Color cameras without the IR Cut filter not only the lumens have influence. Please see details What is the Spectral efficiency factor if you are interested. However the white LED illuminators are widely used in spite of your negative experience. I see the following features: Visible light for human eyes and color image (as against IR), higher Light Efficiency Factor (as against Tungsten Halogen), narrow beam as IR LED ( as against Discharge lamps), fast start ( as against Discharge lamps), low voltage ( as against Discharge lamps)..... In some cases these features become advantages. IR illumination is preferable in some cases, but in some other cases White LED is more suitable (especially with color cameras with IR cut filter).

There is additional reason of why the scene illuminated by IR illuminator might look more brightly. Most of materials have different reflection factors in visible light and in IR. For example, if we assume that a human body has closer reflection factor, then green grass reflects IR radiation better than the visible light of a halogen lamp in 2.2 times! (according to my own tests with CCTVCAD Lab Toolkit). In comparision with White LED the difference should be 3 times and more (becouse of the Halogen lamp light contains a lot of IR too), but I didn't measure comparision with White IR. Thus a scene with a lot of vegetation looks more brightly in IR when a human body looks darker.

Not exactly what I had in mind. True D/N camera with removable IR filter has close sensitivity to White LED in COLOR and B/W modes. Moving off the IR filter has no influence on sensitivity to White light LED becouse the spectrum of the LED has no IR components. Moving off the IR filter increases sensitivity to tungsten lamp, to IR illumination, but doesn't increase sensitivity to White LED, discharge lamps, luminiscent lamp.. Thus if you could switch your TDN cameras to B/W mode, it didn't improve sensitivity to White LED. I suppose there is no reason to use TDN cameras with White LED.

White light can't be painful to the CCD sensors I continue to think about this... It is very important which camera type is used for measuring max range. For b/w cameras visible light LED must excels IR LED in efficiency in several times as I wrote and measured. But for D/N cameras with removable IR filter and for Color cameras without the IR filter the difference can be less. It is possible becouse such camera has colour filter which can't be removed. This color filter causes losing sensitivity in visible range in several times. This is the reason of less sensitivity in visible light of D/N cameras against B/W camera. However the colour filter doesn't reduce sensitivity in IR range equally, but the special IR filter is removed in low light. This can alter proportion between IR and visible light sensitivity for benefit of IR. Using Exview HAD sensor can additionaly increase sensitivity in IR range. In result Visible light efficiency can excel IR with color cameras not in 10-20 times but only in 2-10 times. Acording to inverse square law the max distance must change as square root, in 1.4-3 times only. This difference is not big and non-accurate measuring can ignore it. Other factors such as electric voltage or current, beam angle, internal circuit features, reflection of the scene have influence on the result.

Please see the spectral response curve copied from Sony HAD CCD image sensor ICX205AL user manual, Page 8 This curve shows that the CCD sensor has in several time more sensitivity in Visible light range than in IR LED range. The difference can vary with other image sensors, but sensitivity to Visible light is always greater than to IR in several times. Thus with the same power emmitted a visible light LED must be more effective than an IR LED. Not in all cases in 10-20 times but no less than in 2 times. I checked the Raytec web site and parameters of their illuminators. The parameters of White IR are good. For example, the Light efficiency of White illuminator RAYLUX 200 is 4700Lm/80W=58.75Lm/W. It is well value for LED. Unfortunately, the User manual of IR RAYMAX 200 doesn't contain value of its optical power. I could not find also definition of the "maximum range" and how it has been measured. I suppose there is a difference in definition of "maximal range" for IR and White LEDs. Besouse White LED is used as a rule with color cameras, but IR is used with more sensitive B/W and D/N. As a rule when using visible light installers expect better quality. It can be another reason of more prudent value of maximal range of Visible light. The third reason is that historically manufacturers of IR illuminators overrated the "maximal range" of their products. There is no power standards in IR field. It is like camera sensitivity value which is overrated in the most cases and all know this fact. But in the Visible light engineering field the strict standards are in force. I wrote email to the Technical director of Raytec with this question and link to this topic. I hope he will answer.

It seems very strange. I didn't test Raytec. I don't know the reason of such results.