ESTIMATION OF THE RECOVERY TIME OF A CCD TELEVISION CAMERA AFTER EXPOSURE TO LIGHT OVERLOAD..
Vyacheslav Mikhailovich SMELKOV, PhD in Engineering, Associate Professor
ASSESSMENT OF THE RECOVERY TIME OF A CCD-MATRIX TELEVISION CAMERA AFTER EXPOSURE TO LIGHT OVERLOAD
The paper [1] presents the results of a study of the effect of intense illumination of a television camera on the CCD photodetector matrix. It is shown that light overload with an illumination of 1.4 x 108 lux from a solar radiation simulator does not lead to the appearance of irreversible defects (burns) in the CCD matrix, and the resistance of the photodetector to overload does not depend on the exposure duration. It is noted that the sensitivity of the CCD is restored after the end of an overload of any duration almost without inertia. Obviously, the latter conclusion is not flawless from a metrological point of view.
The author of this paper proposes a method for experimentally measuring the camera recovery time after overload, the peculiarity of which is the organization of high-intensity exposure of the CCD matrix in a pulse mode.
In previously published works [2, 3], the upper limit of the range of working illumination in sunlight for a security television camera was estimated at 100,000 lux. This light overload can be present for a fairly long time: from several seconds to several hours, so the camera must provide an acceptable level of distortion of the television image for the operator, i.e. an acceptable “smear” signal.
For a technical solution of a camera based on a single CCD photodetector, under conditions of such excess illumination, it is possible to guarantee a 10% “smear” signal for the most advanced matrix technology [2] with an exposure time of 1 μs.
On the other hand, using a technical solution of a camera based on two CCD matrices [3], it is possible to reduce the “smear” signal to 1% with an accumulation time of 200 μs, which allows one to additionally abandon the expansion of the range of digital adjustment of the electronic shutter of the photodetector towards “short” times.
Considering that in a number of applications of special-purpose television equipment, where the light overload time may be short, and the speed of presentation of undistorted video information to the automatic receiver is a decisive factor, it is advisable to have a metrological assessment of the time of exit of the television camera after exposure to light overload.
The structural diagram of the television channel for organizing the experiment is shown in Fig. 1, and Fig. 2 shows a timing diagram explaining its operation. The diagram contains a television camera based on a South Korean camera module for wide application on a CCD matrix with a 1/3-inch target format. The CCD matrix has a typical “line-frame transfer” (LFT) organization for imported photodetectors, and in terms of sensitivity it can be classified as a standard sensitivity device (standard CCD).
Fig. 1. Block diagram of the experiment
Fig. 2. Timing diagram explaining the structural scheme of the experiment
A pulsed light source based on two domestic F-326 LEDs connected in series, which emit energy in the near infrared region of the spectrum, which is the working section of the spectral range of the CCD matrix, was used as a source of light overload.
The duration of the light pulse (Fig. 2e) is 20 ms with a repetition period of 80 ms. It is assumed that the “ignition” time (tВ) of the LEDs, as well as their afterglow time (tП) are an order of magnitude shorter than the duration of the light pulse, and therefore do not introduce errors into the measurement of the camera recovery time. The emitter is controlled by a circuit consisting of a sync pulse selector that extracts frame sync pulses from the full video signal (Fig. 2a), the first frequency divider by two that forms a square wave with a period of 40 ms (Fig. 2b), the second frequency divider by two that produces a square wave with a period of 80 ms (Fig. 2c), an “OR-NOT” logical element that provides a positive pulse at the output with a duration of 20 ms with a repetition period of 80 ms (Fig. 2d), as well as a level converter designed to obtain a control output pulse with an amplitude of about 8 V with a load current of at least 1 A (Fig. 2d).
Before measurements are started, the working television camera is installed opposite the switched-off emitter at a minimum distance in order to prepare for organizing a strong light effect on the photodetector. Due to the wide-angle lens installed in the camera, both LEDs are in its field of view against the background of the emitter's front panel painted black.
Then the television camera and the emitter are isolated from the ambient light, and the emitter power is turned on. In this case, a strong light overload of the CCD matrix occurs, since the electronic shutter, which provides a typical minimum accumulation time of 10 μs (1/100000 s), clearly cannot cope with the charge spreading. The oscillogram of the full video signal at the camera output, transmitted by the first beam of the C1-96 oscilloscope, with the emitter turned off is shown in photo 1, and with it turned on — in photo 2. For a temporary representation of the processes occurring in these figures, the oscillogram of the logical pulse for controlling the emitter, transmitted by the second beam of the C1-96 oscilloscope, is shown at the bottom.
Photo 1. Video signal oscillogram (top)
at the camera output with the emitter turned off
Photo 2. Video signal oscillogram (top)
at the camera output with the emitter turned on
From photo 2 it follows that the “parasitic” signal from light overload occurs over two half-frames, i.e. its duration is 40 ms.
The occurrence of this signal can be explained as follows.
During the first half-frame, when the CCD photo target is strongly illuminated, the charge carriers completely fill the potential wells of most of its elements (pixels). At the same time, neighboring potential wells can “collapse,” creating the effect of charge spreading. Due to this, and also due to their greater mobility, infrared photons overcome the potential barrier and penetrate into the cells of vertical shift registers shielded from light. These charge packets are then read in the usual way (line by line and element by element) through the horizontal register of the photodetector, forming a “parasitic” signal of the first half-frame at the output.
During the subsequent (second) half-frame, there is no illumination from the emitter, but in the CCD matrix the processes of charge transfer and their reading in the output device do not stop. Therefore, the charge packets, held until this moment in the potential wells of the photo target, are, as usual, transferred to the vertical shift registers, and then read through the horizontal register. As a result, the photodetector is completely freed from the previously accumulated charges.
As can be seen from the oscillogram in photo 2, in the next two half-frames, i.e. during the remaining 40 ms before the new radiation pulse appears, there is no video signal at the camera output. The experiment shows that if the exposure duration is increased while maintaining its intensity, the recovery time does not change and is also one half-frame (20 ms). This means that the “pause” in restoring the camera image signal after the end of the light overload takes no more than one half-frame according to the standard.
Literature
1. Gridin A.S., Salin V.I., Sushchev G.A., Podgorsky E.G., Ratnikov A.N., Trofimov M.N. Evaluation of the stability of a CCD photodetector to light overloads. “Communication Equipment”, series “Television Equipment”, issue. 2, 1983, pp. 28–32.
2. Kulikov A.N. Television surveillance in bright sunlight.//“Special Equipment”, No. 1, 2001, pp. 11–20.
3. Smelkov V.M. Method for minimizing television camera distortion when operating under light overload conditions.//“Special Equipment”, No. 5, 2001, pp. 20–22.