Automatic flicker frequency detection device and method
View Patent ↗An image device, such as a digital camera, detects specific repeating patterns of signal variations by processing columnar information from the device's two-dimensional sensor array used to generate images. In one embodiment, the columnar information is derived from calculating row averages for two image frames, with each row average being a computed average of the multiple signal intensities generated from some or all of the sensors within a particular row. After the columnar information is determined for each of the two frames, a difference signal is generated as a sequence of the differences between the row averages for the first frame and the row averages for the second frame. This row averaging and frame differencing removes a large percentage of the signal energy that is not a result of the artifact of interest, such as the flicker generated by illumination having intensity fluctuations at 100 Hz or at 120 Hz. In some embodiments of the invention, the row averaging and/or the frame-to-frame comparison are deleted.
1. An imaging device comprising:
an imaging array of sensors arranged in a two-dimensional region having a columnar dimension and a perpendicular second dimension, said sensors having sensor outputs that are responsive to light intensities;
signal processing in communication with said array to receive said sensor outputs and to integrate said sensor outputs as image frames; and
a flicker detector coupled to receive columnar information of said array and configured to detect any one of a plurality of specific repeating patterns of intensity variations within said columnar information, said columnar information being a subset of said sensor outputs and being associated with both a plurality of said sensor outputs in a sequence consistent with said columnar dimension and a plurality of said sensor outputs in said second dimension, each said specific repeating pattern being indicative of a selected periodic fluctuation in source lighting, said flicker detector further includes a frame-to-frame comparator configured to output a columnar difference signal that is a function of differences between said columnar information of a first image frame and said columnar information of a second image frame, said flicker detector having an analyzer configured to detect any one of said specific repeating patterns within said columnar difference signal.
2. The imaging device of claim 1 wherein said sensors are positioned in rows and columns and wherein said flicker detector includes a row-by-row calculator configured to derive said columnar information as a sequence of row values in which each row value is a function of a plurality of said sensor outputs that are generated by sensors within a particular row corresponding to said row value.
3. The imaging device of claim 2 wherein said row-by-row calculator is configured to calculate an average of said sensor outputs in determining each said row value.
4. The imaging device of claim 1 wherein said analyzer is connected to receive said columnar difference signal and is configured to apply a plurality of sinusoidal basis functions to said difference signal in a Cross-correlation operation, each said sinusoidal basis function corresponding to one of said specific repeating patterns, said specific repeating patterns being indicative of 50 Hz and 60 Hz fluctuations in said source lighting.
5. The imaging device of claim 4 wherein said analyzer is configured to generate said sinusoidal basis functions in quadrature to detect a phase of artifacts introduced into said first and second image frames as a consequence of said fluctuations in said source lighting.
6. The imaging device of claim 1 wherein said analyzer is connected to receive said columnar difference signal and is configured to generate Discrete Fourier Transforms for said columnar difference signal.
7. The imaging device of claim 1 wherein said analyzer is connected to receive said columnar difference signal and is configured to detect said selected periodic fluctuations as a function of zero crossings by said columnar difference signal.
8. The imaging device of claim 1 wherein said analyzer is connected to receive said columnar difference signal and is configured to detect said periodic fluctuations by filtering said columnar difference signal at a plurality of frequencies, including 100 Hz and 120 Hz.
9. The imaging device of claim 1 wherein said signal processing is configured to implement variations in frame rate when in a flicker detection mode, said variations being selected to provide frame-to-frame shifts in locations of flicker artifacts from said first image frame and said second image frame.
10. The imaging device of claim 1 wherein said frame-to-frame comparator is positioned to derive a difference frame that is a function of differences between first and second frames of said image signals, said columnar information being determined from said difference frame.
11. A method of processing image information comprising the steps of:
exposing a two-dimensional array of sensors to a scene of interest in a row-by-row fashion, such that any periodic fluctuations in scene lighting introduce flicker artifacts in a columnar direction;
reading sensor outputs from said array to form at least two frames of image information; and
detecting said flicker artifacts in at least one said frame by defining a sequence of row values in which said sequence has a correspondence with said columnar direction of said array and by determining whether said sequence contains selected spatial frequencies that are indicative of said period fluctuations, said detecting further including forming a difference signal that is a function of differences between a first said sequence derived from a first said frame and a second said sequence derived from a second said frame.
12. The method of claim 11 wherein said step of detecting includes:
deriving said sequence as a function of calculating a representative row value for each of a plurality of rows in said array of sensors, each said representative row value being a function of a plurality of said sensor outputs read from sensors in a single said row; and
determining whether said sequence has spatial frequencies indicative of 50 Hz or 60 Hz lighting fluctuations.
13. The method of claim 12 wherein said step of detecting further includes applying cross-correlation between said difference signal and each one of a first sinusoidal basis function related to said 50 Hz lighting fluctuation and a second sinusoidal basis function related to said 60 Hz lighting fluctuation.
14. The method of claim 12 wherein said step of detecting further includes at least one of (a) performing Discrete Fourier Transformations of said difference signal. (b) detecting zero crossings by said difference signal and (c) performing bandpass filtering centered at 100 Hz and 120 Hz.
15. A method of processing image information comprising the steps of:
providing successive first and second frames of image information, each said frame having a row direction and a column direction;
calculating first averages of signal levels for each row of said first frame;
calculating second averages of signal levels for each row of said second frame;
forming a difference signal representative of differences between said first averages and said second averages, said difference signal being a data sequence that corresponds with a sequence of rows of data within said frames; and
detecting artifacts indicative of periodic lighting fluctuations during said forming of said first frame and during said forming of said second frame.
16. The method of claim 15 wherein said step of detecting artifacts includes determining whether said difference signal includes spatial frequencies of 100 Hz or 120 Hz.
17. The method of claim 16 wherein said step of detecting artifacts includes performing cross-correlations between said difference signal and sinusoidal basis functions that correspond to spatial frequencies.