Mixed analog and digital pixel for high dynamic range readout
View Patent ↗An improved CMOS pixel with a combination of analog and digital readouts to provide a large pixel dynamic range without compromising low-light performance using a comparator to test the value of an accumulated charge at a series of exponentially increasing exposure times. The test is used to stop the integration of photocurrent once the accumulated analog voltage has reached a predetermined threshold. A one-bit output value of the test is read out of the pixel (digitally) at each of the exponentially increasing exposure periods. At the end of the integration period, the analog value stored on the integration capacitor is read out using conventional CMOS active pixel readout circuits.
1. An improved CMOS pixel having a photo-detector configured to generate a photocurrent in response to incident radiation, and an integration capacitor for accumulating a charge associated with the generated photocurrent over an integration period, and an analog output circuit for generating an analog signal representative of the accumulated charge over the integration period, wherein the improvement comprises:
a comparator circuit configured to compare the accumulated charge in the integration capacitor with a reference level at discrete sub-integration intervals within said integration period;
a digital output circuit operatively coupled to an output of said comparator circuit, said digital output circuit configured to generate a digital signal representative of the comparison of accumulated charge in the integration capacitor with said reference level; and
wherein a sequence of said digital signals associated with said discrete sub-integration intervals of the integration period is used to scale said analog signal at an end of the integration period to output a scaled pixel output.
2. The improved CMOS pixel of claim 1 wherein said digital output circuit is further configured to stop an accumulation of charge in the integration capacitor in response to an output of said comparator circuit indicating said accumulated charge exceeds said reference level.
3. The improved CMOS pixel of claim 2 wherein a sequence of said digital signals over said integration period identifies a discrete sub-integration interval at which an accumulation of charge in the integration capacitor exceeds said reference level.
4. The improved CMOS pixel of claim 2 wherein said digital output circuit is configured in an initial state to generate a first digital signal representative of said reference level exceeding said accumulated charge; and
wherein said digital output circuit is configured to generate a second digital signal responsive to a signal from said comparator circuit indicating said accumulation of charge in the integration capacitor exceeds said reference level.
5. The improved CMOS pixel of claim 4 wherein said first and second digital signals are one-bit digital signals.
6. The improved CMOS pixel of claim 2 wherein said discrete sub-integration intervals increase exponentially over said integration period.
7. The improved CMOS pixel of claim 2 wherein said digital output circuit includes a flip-flop, an input to said flip-flop coupled to an output from said comparator circuit, and an output from said flip-flop coupled to a digital bus and to a photocurrent switch associated with said integration capacitor.
8. The improved CMOS pixel of claim 2 wherein said comparator circuit is a clocked comparator configured to receive an external sample signal, said external sample signal triggering said comparator circuit to compare said accumulated charge received on a first input to said reference level at each of said discrete sub-integration intervals; and
wherein said comparator circuit is configured to generate an output signal in response to said accumulated charge exceeding said reference level during said comparison.
9. The improved CMOS pixel of claim 1 wherein said digital output circuit is configured in an initial state to generate a first digital signal representative of said reference level exceeding said accumulated charge; and
wherein said digital output circuit is configured to generate a second digital signal responsive to a signal from said comparator circuit indicating said accumulation of charge in the integration capacitor exceeds said reference level.
10. The improved CMOS pixel of claim 9 wherein said first and second digital signals are one-bit digital signals.
11. The improved CMOS pixel of claim 1 wherein said discrete sub-integration intervals increase exponentially over said integration period.
12. The improved CMOS pixel of claim 1 wherein said digital output circuit includes a flip-flop, an input to said flip-flop coupled to an output from said comparator circuit, and an output from said flip-flop coupled to a to a digital bus and to a photocurrent switch associated with said integration capacitor.
13. The improved CMOS pixel of claim 12 wherein said photocurrent switch is coupled between the photodiode and the integration capacitor, whereby opening of said photocurrent switch isolates the integration capacitor from the photodiode.
14. The improved CMOS pixel of claim 1 wherein said comparator circuit is a clocked comparator configured to receive an external sample signal, said external sample signal triggering said comparator circuit to compare said accumulated charge received on a first input to said reference level at each of said discrete sub-integration intervals; and
wherein said comparator circuit is configured to generate an output signal in response to said accumulated charge exceeding said reference level during said comparison.
15. A method for extrapolating a high dynamic range scaled pixel output from a pixel having a low dynamic range analog output signal, comprising:
at each of a plurality of discrete sub-integration intervals over an integration period, comparing an analog output signal at an integration capacitor which is representative of incident radiation on a photo-detector of the pixel with a reference level to generate a sequence of one-bit digital output signals, wherein
(a) responsive to said analog output signal being less than said reference level, generating a first one-bit digital output signal, and
(b) responsive to said analog output signal exceeding said reference level, generating a second one-bit digital output signal and holding said analog output signal at a current level; and
modifying said analog output signal at an end of said integration period using said sequence of said one-bit digital output signals over said integration period to output a high dynamic range scaled pixel output.
16. The method of claim 15 wherein said step of modifying includes generating an exponent qualifying said analog output signal at the end of said integration period from said sequence of one-bit digital output signals.
17. The method of claim 15 wherein said step of holding said analog output signal at a current level includes disconnecting said integration capacitor from said photo-detector.
18. The method of claim 15 wherein said sequence of one-bit digital signals identifies a discrete sub-integration interval within said integration period at which said analog output signal exceeded said reference level.
19. The method of claim 15 wherein each of said discrete sub-integration intervals increases exponentially.
20. The method of claim 19 wherein an increase in dynamic range of said pixel output is proportional to a ratio of the total integration period to the smallest sub-integration interval of the plurality of discrete sub-integration intervals.
21. The method of claim 15 wherein said reference level is selected to be between 40% and 50% of a maximum of said analog output signal.
22. The method of claim 15 wherein said reference level is selected in relation to an anticipated constant instant radiation over said integration period.
23. The method of claim 22 wherein said reference level is selected such that if said analog output signal exceeds said reference level during a discrete sub-integration interval, said analog output signal will remain below a maximum limit during a next consecutive discrete sub-integration interval.