FR2565750A1 - Solid state imaging appts. - Google Patents

Abstract

The solid state imaging appts. comprises a charge-coupled device having photosites and a transport shift register. The appts. includes exposure control gate and sink diode for controlling the integration time of the photosites to be a constant selected as desired independent of the exposure period. The charge-coupled device may include a charge-well, between each photosite and the corresponding section of the transport shift register, the constant integration time being provided by only allowing charge to build up in the charge wells after an initial period of time during each scan period. The required initial exposure control period is determined by means of counters.

Description

 The present invention relates to an exposure control system for a solid-state imaging device or imager, comprising a charge coupled line or a network-shaped detector of a certain extent and a connected circuit to this detector in order to extract a signal representing the intensity of the incident light falling on the detector.

Such a device is used in tsslssclnssmaic systems that is equipment used to produce a video signal from a cinematographic film. Articles concerning the operating principles of charge coupled devices and their use for ia image detection appeared in the catalog of the so-called Fairchild company, concerning CCD type image forming and signal processing apparatus, published in 1981 by Electronic 2000 for Fairchild CCD Imaging, 4001 Miranda
Avenue, Palo Alto, California 94304, United States of America.

 A particular problem with such solid state imaging devices is to adapt to changes in the total level of incident illumination, that is, to provide means for controlling the exposure. Exposure control for linear grating telecine systems poses various problems for specialists in the practice of scanning spot or cathode ray tube telecine. These problems are due in principle to the fact that the linear array shaped charge-coupled device (CCD) cannot handle any condition involving exposure levels above its saturation level. itself which is referred to as "hyperluminosity". Hyperluminosity occurs when the size of the photoelectric charge packets is too large to be supported by the device's shift register wells. There is an overflow of overloaded wells in neighboring wells, which therefore receive information which does not emanate from the photosensitive sites associated with them. This has disastrous effects on the image provided.

There are two known techniques aimed at solving or solving this problem, but each of them has its own limitations. The first method uses a disc or wedge, forming a neutral density filter and which is inserted in the light path so as to increase or reduce the amount of light falling on the network.
CCD. The disadvantage of this system is its slow operation due to the mechanical inertia of the disc and the components associated with it and which control it, which excludes. effectively. this form of light control in applications that implement any form of pre-programming. Ideally, in a preprogramming operation, all corrections should be completed during an image deletion period. Since the neutral density filter requires an interval of time of several images to complete a desired correction, the effect of installing the filter can be observed during operation.

 According to the second method, a fixed neutral density filter is used such that, in the worst conditions expected, the network can never be over-exposed, which eliminates the hyperluminosity effect. . If this filter is used, the exposure control then becomes a gain control in the video processing channel connected to the output of the CCD network. This eliminates the response speed defect, but in reality means that under a wide variety of film speed conditions, the array does not operate near its optimum light entry level and therefore can cause the appearance of '' an increased level of interference in the telecine output signal.

 Another problem encountered in solid-state imaging devices or imagers is the effect of variations in the scanning period of the amplitude of the output signal from the CCD device.

 In itself a linear CCD network consists of a row or line of photosensitive sites, typically present in a number of 1024. These sites are exposed to light during the time interval during which they transform light energy into "packages" of charges under each photosensitive site. At the end of this period normally designated as the scanning period, these charge packets are transferred to a shift register, known as the transfer register, and are transmitted in a clocked fashion to the shift register until at the output, where each packet of charges is individually transformed into a voltage signal and is subsequently amplified. According to this operating method, the minimum scanning period is the clock period of the shift register multiplied by the number of elements present in the shift register (typically around 50 pus). When using a CCD network in a continuous motion telecine system, this means that the maximum film speed is associated with the number of lines of information required in each film frame (which is fixed by the standard number of lines in the film. television system used) and the time interval that an image requires to be transmitted by the CCD network.

 It will be noted that it is not the entire film which circulates, so that the speed of the film through the CCD device makes it possible to obtain precisely the minimum scanning period. That is, if the film speed is too high, c1 is insufficient line information which is read from the film, and if the film speed is too low, too many lines are then read from the film to satisfy the standard number of television lines. This is why it is necessary to control the rate at which line scans are required at the detector.

However, if the line period (and therefore the time interval for each line) varies, the output level varies since the detector is subjected to a longer exposure period (normally referred to as time integration).

 FIG. 1, appended to the present application, represents in (a) a (d), different waveforms illustrating the operation of a conventional CCD image detector. In (a) there is shown a series of scan request pulses, which are assumed to occur at intervals which are equal to the minimum scan period (MSP), i.e. the time used to unload the transfer register. The integration time (IT) available for the photosensitive sites has also been shown. The pulse train delivered by the CCD device depends on the light falling on the grating, but will typically take the form shown in (b).

 If for any reason the frequency of the scanning pulses is reduced, then the integration time is increased correspondingly. As shown in (c), the scanning period is doubled and the integration time is increased correspondingly, so that the CCD output signal is shown in (b) and is substantially doubled in amplitude.

 It is noted that if the exposure could be controlled so that the integration time is constant independent of the exposure period, then the output signal from the CCD device could be constant regardless of the scanning request period.

 This control can be achieved by adjusting the exposure control voltage, as shown in Figures le to g. Since the device is regularly controlled in a clocked fashion, the integration time can be defined in the form of a number of clock cycles and counting techniques can be used to control this period. Note that controlling the integration time should also eliminate the problem of hyperluminosity.

 The exposure control according to the present invention can be obtained using a modified charge coupling device, which includes a charge sink between each photosensitive site and the corresponding section of the transfer register. The data is moved along the transfer register as a result of the alternating application of voltages V1 and V2 at the scanning clock rate. However, the charge, which forms at the photosensitive sites, is not transferred directly to the transfer register, as is conventional.

Instead of the usual single transfer door, two transfer doors are provided. The charge sink depends on the voltage applied to the transfer gates so as to allow the charge to form or not to form in the charge sink below the photosensitive sites. In order to obtain a constant integration time, the charge can only form at the end of an initial time interval during the scanning period.

 More specifically, the above-mentioned problems are solved using a solid state imaging device comprising a charge coupled device having a plurality of photosensitive sites and a shift register of transfer, scan request means for producing pulses of scan requests at a frequency representative of the desired scan frequency, a clock for producing scan clock pulses for controlling the transfer shift register, and control means for applying control voltages to the charge coupling device, characterized in accordance with the invention in that the control means include counting means for counting count pulses delivered by the clock and for validate the charge coupled device in response to these pulses so as to allow the integration of the output signal from the photosensitive sites during a predetermined number of counting pulses, which makes it possible to adjust the integration time to a desired value.

Other characteristics and advantages of the present invention will emerge from the description given below taken with reference to the accompanying drawings, in which
- Figure 1, which has already been mentioned, shows different waveforms illustrating the operation of a conventional CCD image detector
- Figure 2 schematically shows the arrangement of part of a CCD device intended to be used in accordance with the present invention
- Figure 3 shows a circuit used to control the CCD network of Figure 2; and
FIG. 4 illustrates waveforms making it possible to explain the operation of the device of FIG. 2.

 FIG. 2 schematically shows the arrangement of a part of a CCD device intended to be used in accordance with the invention. The charge produced te at each photosensitive site 10 is transferred via a photosensitive grid 11. A well is created below the photosensitive grid 11 by voltages which exist at the level of both a grid transfer 12 and an exposure control grid 13, effectively mounted in parallel with the transfer grid. The capacity of the well can be controlled only by the potential of the exposure grid, with any excess charge entering the well diode 14.

 If, during operation, the voltage on the exposure grid is controlled so that it reaches only a more negative level than the voltage of the photosensitive grid for a fixed time interval before the transfer of the charge to the transfer register 15, then this fixed time interval is equal to the integration time and can be selected as desired.

 In other words, the output signal of the CCD device can be controlled so as to take account of variations in the level of the incident light, by means of an appropriate selection of the integration time. Naturally, the integration time must be chosen equal to or less than the minimum scanning period.

 FIG. 3 represents a circuit used to control the CCD network of FIG. 2. The circuit includes a progressive counter 20 which is loaded with data corresponding to the number of scanning clock pulses necessary for the desired integration time at the end of the scan request pulse. The scan clock pulses are (or are derived from) the pulses used to control the transfer shift register in the CCD device. The data is loaded negatively by the fact that the counter is loaded with the difference between its counting number maximum and the number of scan pulses required for the integration time. The counter 20 then counts progressively at the scan clock frequency until the new scan request pulse 1 when the donation is charged again and the counting cycle is repeated. During this counting cycle, beyond the number of clock pulses corresponding to the integration time, a pulse is delivered via its line "OUTPUT OF REPORT" to a flip-flop control circuit 22. This validation pulse is used, together with the next scan request pulse, to produce a "flip-flop control" pulse used to activate a flip-flop circuit 24. During the next scan request pulse, the counting state of counter 20 represents the exceeding of the interval between successive pulses of scanning request during the desired integration time, that is to say the time interval during which the CCD device must be invalidated to provide correct exposure. The flip-flop 24 transmits the counting state to a regressive counter 26, at the instant when the next pulse of scan request is present in the form of a preload data value. This counter 26 is loaded in response to a pulse from the scan request pulse, and performs a regressive count at the scan clock frequency so as to derive the end of an exposure control pulse. An exposure control pulse generator 28 creates the required exposure control pulse from a start pulse, taken from another location (but precisely timed with respect to the request pulse scan), and the final pulse is taken from the carry-over output "of the regressive counter 26. The integration period then begins - at the end of the exposure control pulse and continues until next scan request pulse. This can be seen more easily on the waveforms shown in Figure 4. In 4 (a), the scan request pulses are shown and (b) the d pulses are shown. scanning clock, which have a significantly higher frequency. The progressive counter 20 provides a transfer pulse as shown in (c) after the integration time TINT. The control output signal of the flip-flop is represented in ( d).

The regressive counter 26 performs the counting of the remainder of the scanning period so as to deliver pulses as represented in (e) and which precedes the next scanning request pulse, for a duration equal to the time of desired integration. The exposure trigger pulses are shown in (f) and the exposure pulses produced by the circuit in Figure 3 are shown in (g).

 The exposure pulses are sent to the second transfer gate so as to erase the contents of the charge well and to allow subsequent integration during the correct integration time.

 The system shown works by maintaining a constant integration time for repetition frequencies of the scan request pulses, which are higher than the minimum scan frequency, provided that the two counters 20 and 26 count down a sufficient number of "bits" to cover the scan repeat period as a number of scan clock pulses. If the maximum period for the scan request pulses is equivalent to 2048 scan clock pulses, then the two counters must have a counting capacity equal to at least 2050, but preferably greater, for example equal to 4096 .

Claims (4)

 1. Device for forming solid-state images, of the type comprising a charge coupling device having a plurality of photosensitive sites (10) and a transfer register & shift (15), scanning request means ( 28) for producing scan request pulses at a frequency representative of the desired scan frequency, clock means for producing scan clock pulses for controlling the transfer shift register (15), means control means (20-26) for applying control voltages to the charge coupling device, characterized in that the control means (20-26) comprise counting means (20, 26) for counting pulses of counting delivered by the clock and used to validate the device with charge coupling in response to its pulses so as to allow the integration of the output signal from the photosensitive sites (10) during a predetermined number of pulses counting ions, which allows the integration time to be adjusted to a desired value.  2. Device according to claim 1, charac terized in that the charge coupling device comprises a charge well located in each photosensitive site (10) and the corresponding section of the transfer shift register, the integration time desired being obtained only. allowing a charge to settle in the charge sinks after an initial time interval during each scanning period.  3. Device according to claim 2, characterized in that charge wells are created by voltages on transfer grids (12) located between photosensitive grids (11) for photosensitive sites and the transfer shift register (15 ) and on exposure control grids (13) mounted in parallel with the transfer grids (12), the capacity of the wells being controlled by the potential of the exposure grid (13) with a penetrating excess charge in a well diode (14), and the voltage on the exposure grid (13) is controlled such that it reaches only a more negative level than the voltage of the photosensitive grid (11) during an interval fixed time after the transfer of charges to the transfer shift register (15) has been completed.  4. Device according to claim 1, characterized in that the counting means (20, 26) comprise a first counter (20) used to count counting pulses in order to produce a counting state representing the difference between the interval between successive scan request pulses and the desired integration time, and a second counter (26) for counting count pulses for a duration dependent on the output signal of the first counter so as to define a delay period before that integration does not begin.