Stanislav

There are other ready devices for controlling ambience transparence - Infrared beam smoke detectors. For example

Hi, Jimmy I get 1.8GB (Compression=50). May be you just used other compression ratio. I understand that the Axis Design Tool 2 doesn't allow to specify pixels directly. But the CIF is 384*288=110592 pixels is not equals to 480*360=172800 pixels. The 480*360 image has in 1.56 times more pixels therefore it must have in 1.56 more file size (in bytes). I get 0.44GB (Compression=50). May be you used other compression ratio and mistaken with number of days. This is the same as 0.44*1.56=0.69GB for 480*360. It is in 2.5 times less than in Axis Design Tool 1 (1.8GB). It seems like an error in Axis Design Tool 2.

I have obtained the following results: 1. /////////CIF (480x360 on the Axis design tool) H.264 12FPS Compression - 50 24HR Recording Stairway Profile (Only specific to Axis) Axis Design Tool 1 (320*240) - 80kBit/s 843.7MB/day Axis Design Tool 1 (480*360) - 181kBit/s 1.8 GB/day Axis Design Tool 2 - (CIF=352*288) - 40.9kBit/s 0.44 GB/day The Axis Design tool 2 gives result in 2.5 times less than the Axis Design Tool 1. It seems like an error in the Axis Design Tool 2: In the Axis Design Tool 2, 4CIF produces in 10 times more traffic (427 KBit/s) than CIF (40.9KBit/s), but the 4CIF has only in 4 times more pixels. In the Axis Design Tool 1, 640*480 produces in 4 times more traffic (321 KBit/s) than 320*240 (80KBit/s), that is correct. 2. ////4CIF (640x480 on the Axis design tool) H.264 24FPS Compression - 50 24HR Recording Station Profile (Only specific to Axis) Axis Design Tool 1 (640*480) - 1268kBit/s 13 GB/day Axis Design Tool 1 (800*500) - 1650kBit/s 16.9 GB/day Axis Design Tool 2 - (4CIF=704*576) - 1.71MBit/s 18.5 GB/day Approx the same result in the second case.

I suppose you need not microcontroller to make this device. It can be made by several CMOS chips, transistors and relay. The simplest way is using ready photosell pair for a gate or a bar. You should adjust (decrease) receiver's sensitivity. I suppose in this case you will need the separated modulated IR transmitter/receiver too. Using a mirror allows to place transmitter and receiver on the same wall. Interesting idea, but the problem is the exposure adjusting time as you mentioned. It will make terrible image from the camera.

i did state motion detect. but with a normal camera with built in IR there is not much you can do anyway (another reason why you dont use built-in ir) We can open the camera and build a relay for IR, but you'll write it is very technical . Generally it is disadvantage of built-in ir, I agree.

wow its getting very technical. you could always remove the subatomic hadron particle let the water dissolve the neutron frame work and block it with a nuclear filter. but i have found it much easier if you just use a motion light with pir sensor. if its bright enough it will switch your ir off. remember you are talking about fog. i dont see the point in modifying a camera just to deal with a problem that only happends with fog pir sensor will be not confidently switched by fog and snow. Also the illuminator will not confidently switch of the IR built in cameras. The illuminator will produce additional reflection. We also need to turn on the floodlight. Active IR seems the best. There are a lot of such schemes.

May be a scheme like an "active infra-red sensor" will be suitable. It consist in two parts: a transmitter of modulated IR beam and a receiver. The worse transparence between the transmitter and the receiver is, the less signal level on the receiver. When the signal is less than the specified level, a relay will turn off IR build in cameras and turn on the floodlight.

I suppose it is just manipulations with the Contrast and the Black Level. Plus Saturation. Brightness of the fogged image is shifted down, thus the fog level becomes the Black Level, then the Contrast of the obtained dark image is increased. If these manipulations is performed by 12-bit video processor, in some cases they can increase visually quality of the output 8-bit image. But these manipulations can't increase camera sensitivity with fixed Signal/Noise ratio. I programmed similar image processing in VideoCAD but for the opposed purpose- modeling image distortions of real cameras in dependence of their parameters and scene conditions The video on youtube was compressed therefore it doesn't show real difference. Don't believe your eyes

There is a sufficient explanation in the KP-D5001(P)/D5000(P) manual. This is a contrast correction. Adaptive Fog Reduction AFR feature provides real-time image correction for improved visibility in dark, harshly backlit, or unclear/foggy environments, while optimized computational algorithms enable frame rates up to 30 fps. Through use of this technology the camera is capable of adjusting the contrast ratio of images that become pale due to foggy conditions and restoring the fading colors resulting in dramatic improvement in visibility.

Try to model your situations in VideoCAD Starter. During one month evaluation period you can use it for free. You can easy choose the best camera positions, heights, resolutions and lenses and see results in 3D.

I suppose it is not IR problem. Visible light produces the same effect, like headlight in fog. The problem is in the direction of light and known inverse square law. Fog/snow particles are near to the camera and the build in illuminator therefore the particles are much brighter than other far objects on the scene. Then camera automatic (AESC, AGC) adjust image contrast at these bright particles, at the same time other far objects on the scene become darker. To improve visibility in fog/snow you should place illuminators not near the cameras, but aside or from above. May be place additional illuminators and turn off build in illuminators during fog/show.

I am sure that fog, smoke and heavy rain must reduce the affect of IR illumination, as they reduce conventional light efficiency. IR is the same electromagnetic waves as the visible light and its behavior is very similar. May be very small particles less affect to IR because of longer wavelength, but I don't know exact data.

Also P radiant=47.1W is unreal for Raymax100. Its consumption is 70W, thus its expected P radiant can be no more than 20W minus optical loss. It seems new parameters are much better. Just seen spec of SFH 4232 Platinum Dragon Package, half angle ±60°, 850nm, typ. 530mW at 1A dc. Its total radiant flux=530mW with 1.8W power consumption. See an interesting IR illuminators design guide by OSRAM High Power Emitters for Illumination Applications See Page 5. There the difference of 60 deg LED SFH 4232 and the same LED with a 10 deg Lens is compared. Pay attention it is HALF ANGLE no FULL angle. "A suitable lens for the Platinum Dragon SFH 4232 to meet the ±10° FOV requirement is for example the Lisa-SS lens from LEDIL which will be used in this example. Test measurements show that the radiant intensity of SFH 4232 is increased by a factor of 7.6 to 1400 mW/sr at 1 A. Without lens we only get 180 mW/sr. " Thus the practical result is 7.6 times. (for 20 deg and 120 deg full angle) The difference from the theory arises because of optical loss.

The difference between indoor<>outdoor arises because of rereflection. Indoor there is much more rereflected light (or IR). Thus in our measures we should take rereflection into account. We should perform our tests at short distances between source and receiver. The distance should be much less than distances to other objects in the room. We can shade rereflected light using tube or other similar tools. We can also measure rereflected part of light separately by shading direct light. Then deduct rereflected part from the result.

For relative tests you can try to use a silicon photodiode and a multimeter, but for absolute measuring we need a standard (etalon) IR source to graduate it. Spectrum response curves of photodiodes are available in its spec thus we can use them for different wavelengths. For my tests I have ordered customized IR sources carefully tested in radiometric laboratory , they are expensive. In this laboratory it could be possible to order calibrated photodiodes too.

This specification seems correct. I got close results for such cameras. It is good that BOSCH writes right sensitivity values. White LED doesn't contain IR at all. It is correct. The difference between 0.09lx and 0.26lx is not very big at the most of levels of illumination. You are observant . IR sensitivity of TDN cameras can be close to the IR sensitivity of B/W cameras. But TDN camera sensitivity of visible part of light is worse than B/W because of color filters which can't be disabled.

Please see my corrections of your calculations:

There are several factors. Generally, the more distance between your camera and your DVR - the more influence has cable type. The more resolution your camera has - the more influence has loss of hight frequencies of your cable. Color cameras require better high frequency transmission. The less voltage camera output has, the less sensitivity your DVR has - the more influence has cable resistance. The thinner your cable is - the less its resistance and the more its loss of high frequencies. The more radio interference on your place - the more influence has shielding of your cable. Thus is some conditions there can be no visible difference between different cables, but in other conditions there can be big difference.

There is a simplified way for manual calculation of the ratio between the Average Light intensities in dependance of the cone angle (without taking into account the form of Light intensity distribution curve and the difference in optical loss): 1. Solid angle As=2*PI*(1-Cos(An/2)) where An - the cone angle in degrees For 10 degree As10=2*3.14*(1-Cos(10/2))=0.023897 steradian For 60 degree As60=2*3.14*(1-Cos(60/2))=0.84136 steradian 2. Average Light intensity Li=F/Ab where F - light flux which is the same for 10 degree and for 60 degree angles For 10 degree Li10=F/As10=F/0.023897 For 60 degree Li60=Li=F/As60=F/0.84136 3. The ratio between the Average Light intensities of 60 deg illuminator and 10 deg illuminator with the same Light flux, without taking into account the form of Light intensity distribution curve and the difference in optical loss. R=Li60/Li10=(F/0.023897)/(F/0.84136) =0.84136/0.023897=35.2

Thanks, is the radiant intensity the same as the Ie term? edit: ok I answered that myself (yes) on Wikipedia Did you mean 36, not 32? Not trying to give you a hard time just don't have a lot of experience with these calcs and especially steradians... http://www.light-measurement.com/calculation-of-radiometric-quantities/ I got value 32.36 - ration between Axial Light intensities (!not equal Average Light intensities). The ratio depends also on the form of Light intensity distribution curve. I assumed the ratio of the light intensity on border of the cone by the light intensity on axis=0.5. I used the Illuminator calculation tool in VideoCAD. All necessary formulas with steradians were programmed in this calculator. But in practice the difference will be less because the less the angle of radiation is the more optical loss. You can measure this difference in practice using the CCTVCAD Lab Toolkit. See here detailed description of this procedure.

In theory sharpening angle of radiation from 60 deg to 10 deg can increase axial radiant intensity in 32 times (if the power of radiation will remain the same). In practice the differense will be less because increasing optical loss.

Thank you very much for your encouragement I have also a link for great independent laboratory testing of many widely known cameras by ProSystem CCTV magazine (formerly Russian edition of Australian CCTV Focus magazine). There are many other very interesting prof made tests. But these materials are in Russian. Here is the link I have also several interesting articles in Russian.

It is impossible in general like increasing image resolution frequently shown in cinema. Image processing can't increase amount of information which was produced by sensor. As I wrote above ...It can only improve some parameters (for example signal/noise ratio) at the expense of making worse other parameters (for example -resolution). Don't agree. You were involved in audio systems in the past. Your statement sounds like "Thus, MICROPHONE sensitivity does not directly correlate to SOUND quality or low-SOUND performance of the AUDIO SYSTEM." OK . I wrote my arguments, gave links, but you didn't read my links and remains on your opinion. It is your business

Ya xochu tebe soxranit tvou nervu Ostav Rory v pokoe sdes pochti ni y kogo net tools and obrazovania teby ponyat Thank you for your advice I know that you are right. But I am not nervous. I have already got accustomed to incomprehension. I hope someone understand me. I write for these people. I spent years for these investigations and I am sure of my rightness.

I have tested JVC TK-C925E carefully in 2007. Its really sensitivity in B/W mode (S/N 17dB, Halogen lamp, F1.2, Exposure=20ms) is 0.1lx. Is has a trick "Lo LUX". I don't remember exactly.. but all these tricks really are long exposure(-motion blur), noise reduction (-resolution) or merging neighboring pixels (-resolution).