Flicker detecting circuit and method in image sensor

The present invention relates to a semiconductor design technique, and more particularly, a circuit and method for detecting a flicker in a band form in an image sensor.

Recently, demand for digital cameras has explosively increased with the development of picture communications using the Internet. Moreover, demand for a compact camera module has increased as mobile communications terminals, such as Personal Digital Assistants (PDA), International Mobile Telecommunications-2000 (IMT-2000) terminals, and Code Division Multiple Access (CDMA) terminals, in which a camera is built, have become increasingly wide-spread.

The camera module essentially includes an image sensor. The image sensor typically is a device for converting an optical image into an electric signal. A Charge Coupled Device (CCD) and a Complementary Metal-Oxide-Semiconductor (CMOS) image sensor are widely used as such an image sensor.

The CCD disadvantageously has a complex driving method and a high power consumption. Further, the CCD has the disadvantages that the manufacturing process is complicated because the number of masking processes is large and a one-chip implementation is difficult because a signal processing circuit cannot be built in a chip. On the contrary, the CMOS image sensor is now more prevalent because it is possible to monolithically integrate a control circuit, a driving circuit and a signal processing circuit on a single chip. Moreover, the CMOS image sensor provides a lower voltage operation, a lower power consumption, and a potentially lower cost compared to the existing CCD owing to the compatibility with peripherals and the utility of the standard CMOS manufacturing process.

However, an analog signal introduced from a photo-receiving device, e.g., a photodiode, in the CMOS image sensor has diverse parasitic effects caused by a parasitic capacitance, a resistance, a dark current leakage, the inconsistency of features of a semiconductor device and so on. Such a parasitic effect essentially occurs in a semiconductor device and degrades a Signal to Noise Ratio (SNR) of image data. Thus the noise acts as a critical factor for limiting a performance of the image sensor.

Causes of noise in the CMOS image sensor may include kT/C noise related to a sampling of the image data, a 1/f noise related to a circuit used for amplifying an image signal, and a Fixed Pattern Noise (FPN) related to the inconsistent signal processing circuits in the image sensor. The FPN, among others, is visually very poor since it is seen in a vertical line or a strip form within the image and thus can easily be detected by human eyes.

Such a CMOS image sensor reads out data in a rolling shutter scheme, which sequentially reads out data in every row line from pixels aligned in a pixel array. The light intensity of one row line differs from that of another row line based on the standardized frequency 50 Hz or 60 Hz, inducing a flicker.

Hereinafter, the cause by which the flicker is induced in the image sensor employing the rolling shutter scheme is described.

As shown in FIG. 1 , the flicker results from a difference between the light intensities of every row line in an image pickup device using a rolling shutter. This is because every country defines a rated electric frequency as, e.g., 50 Hz or 60 Hz (the rated electric frequency is defined as both 50 Hz and 60 Hz in Japan), so that the luminance of a lighting device is varied with time variation. A representative example using the rated electric frequency is a fluorescent lamp universally utilized in ones real life. When an object illuminated by the fluorescent lamp is photographed by using the image pickup device employing the rolling shutter scheme, the flicker is induced in a lateral band on an image of the object.

In order to prevent the induction of such a flicker, the difference between the light intensities of each row line as the cause of inducing the flicker has to be uniformly maintained. FIG. 2 shows a method for maintaining the difference between the light intensities of each row line, in which an exposure time of the image pickup device is fixed to a time assigned to the rated electric frequency, e.g., 50 Hz or 60 Hz. That is, if integer multiples of time assigned to the rated electric frequency is given as the exposure time of the image pickup device, the light intensities provided from every row line become equal. Therefore, it is required to adjust the at least common multiple of the exposure time of the image pickup device in compliance with the rated electric frequency per country. Especially, under an environment that two frequencies 50 Hz and 60 Hz are mingled as in Japan, there needs a circuit for properly and automatically adjusting the at least common multiple for the exposure time.

As mentioned above, in order to automatically adjust the at least common multiple of the exposure time of the image pickup device in compliance with the rated electric frequency, a flicker detector is required for determining whether the flicker is present or not under an actual environment. Further, when the flicker induced due to a difference between the exposure time of the image pickup device and the time assigned to the rated electric frequency is detected by the flicker detector, an automatic exposure correcting circuit is required to correct the difference therebetween.

Whether the flicker is present or not can be determined by one method of using a separate particular pixel array, and another method of using a pure digital signal processing.

The former method using the particular pixel array, which is constructed separately from a main pixel array, determines whether the flicker is present or not under an actual illumination environment, thereby analogizing the existence of the flicker in a photographed image. This method can obtain an exact flicker detective result but has a disadvantage that a hardware resource is wasted by the addition of the particular pixel array, and time and financial cost for designing and controlling the particular pixel array increase.

The latter method using the pure digital signal processing processes a tendency of the flicker which appears in a single or a plurality of images into information to generate a flicker curve, and then determines a periodicity of the flicker curve from the flicker curve, thereby analogizing the existence of the flicker. This method is relatively inexpensive in a designing cost compared to the former method using the particular pixel array, because a separate hardware resource is not required and sufficient simulative experiences are executable in a Field Programmable Gate Array (FPGA).

However, this latter method using the pure digital signal processing to determine whether the flicker is present requires another separate hardware resource, i.e., a Discrete-Fourier Transform (DFT) operator, for calculating a complex DFT operation to determine the periodicity of the flicker curve.

Embodiments of the present invention are directed to provide a flicker detecting circuit and method in an image sensor, for determining whether a flicker is present or not, while minimizing a semiconductor-implemented area and the hardware resources.

In accordance with an aspect of the present invention, there is provided A circuit for photographing an image to obtain a first frame and a second frame and detecting a flicker by using the first and the second frames, the circuit comprising: a luminance average value output unit for extracting luminance values from pixel data of the first and the second frames to generate first luminance average values for pixel lines of the first frame and second luminance average values for pixel lines of the second frame; a flicker curve generating unit for subtracting the second luminance average values from the first luminance average values, thereby generating a flicker curve; and a flicker detecting unit for extracting a plurality of local minimum points from the flicker curve, calculating a distance between each two neighboring local minimum points of the extracted local minimum points, and determining whether the flicker is present based on the distances and the frequency numbers of the distances.

In accordance with another aspect of the present invention, there is provided a circuit for photographing an image to obtain M frames and detecting a flicker by using the M frames, M being an integer larger than 2, the circuit comprising: a luminance average value output unit for extracting luminance values from pixel data of M−1 neighboring frame pairs among the M frames to generate first luminance average values for pixel lines of a first frame of each frame pair and second luminance average values for pixel lines of a second frame of said each frame pair; a flicker curve generation unit for subtracting the second luminance average values from the first luminance average values, thereby generating a flicker curve; a flicker image detecting unit for extracting a plurality of local minimum points from the flicker curve, calculating a distance between each two neighboring local minimum points of the extracted local minimum points, and detecting a flicker image based on the distances and the frequency numbers of the distances; and a flicker state determining unit for determining whether a flicker is present based on the detecting number of the flicker images when the detection of the flicker image is completed with respect to the M frames.

In accordance with still another aspect of the present invention, there is provided a circuit for photographing an image to obtain M frames and detecting a flicker by using the M frames, M being an integer larger than 2, the circuit comprising: a luminance average value output unit for extracting luminance values from pixel data of the M frames to generate luminance average values for pixel lines of each of the M frames; a flicker curve generating unit for calculating luminance average value differences between two frames constituting a frame pair based on the luminance average values for the pixel lines of the frame pair, the M frames being classified into (M−1) frame, thereby generating (M−1) flicker curves by using the differences; a flicker image detecting unit for extracting a plurality of local minimum points from each of the (M−1) flicker curves, calculating a distance between each two neighboring local minimum points of the extracted local minimum points, and detecting a flicker image from each of the flicker curves based on the distances and the frequency number of the distances; and a flicker state determining unit for determining whether the flicker is present based on the detecting number of the flicker images when the detection of the flicker images is completed with respect to all of the M frames.

In accordance with still another aspect of the present invention, there is provided a method for detecting a flicker in an image sensor, the method comprising: extracting luminance values from pixel data of a first frame obtained by photographing an image to generate first luminance average values for pixel lines of the first frame; extracting luminance values from pixel data of a second frame obtained by consecutively photographing the image to generate second luminance average values for pixel lines of the second frame; subtracting the second luminance average values from the first luminance average values to generate a flicker curve; and extracting a plurality of local minimum points from the flicker curve to calculate a distance between each two neighboring local minimum points of the extracted local minimum points and detecting a flicker image based on the distances and the frequency numbers of the distances.

In accordance with further still another aspect of the present invention, there is provided a method for detecting a flicker in an image sensor, the method comprising: extracting luminance and chrominance values from pixel data of a first frame obtained by photographing an image to generate first luminance average values and first chrominance average values for pixel lines of the first frame; extracting luminance and chrominance values from pixel data of a second frame obtained by consecutively photographing the image to generate second luminance average values and second chrominance average values for pixel lines of the second frame; estimating a motion of the second frame by using the first chrominance average values and the second chrominance average values; correcting the second luminance average values based on the estimated motion of the second frame; subtracting the corrected second luminance average values from the first luminance average values to generate a flicker curve; and extracting a plurality of local minimum points from the flicker curve to calculate a distance between each two neighboring local minimum points of the extracted local minimum points and detecting a flicker image based on the distances and the frequency numbers of the distances.

FIG. 1 is a view for illustrating a difference between the light intensities of every row line appearing in a typical image pickup device employing a rolling shutter method;

FIG. 2 is a view of illustrating a method for adjusting an exposure time, to uniformly maintain the difference between the light intensities of every row line appearing in a typical image pickup device employing a rolling shutter scheme;

FIG. 3 is a conceptive view of illustrating a mechanism for detecting a flicker in an image sensor in accordance with the present invention;

FIG. 4 is a block diagram of a flicker detecting circuit in accordance with a preferred embodiment of the present invention;

FIG. 5 is a flow chart of describing an operational characteristic of a motion estimator shown in FIG. 4 ;

FIGS. 6 and 7 are views of illustrating a motion estimating method;

FIG. 8 is a view of illustrating an operational characteristic of a flicker curve detector shown in FIG. 4 ;

FIG. 9 is a flow chart of providing a flicker detecting method of the image sensor in accordance with the present invention;

FIG. 10 is a timing diagram of illustrating a timing of calculating and storing averages of luminance and chrominance values shown in FIG. 9 ;

FIG. 11 is a view of showing a flicker curve extracted at the flicker curve detector shown in FIG. 4 ;

FIG. 12 is a view of showing a flicker curve extracted through a motion correcting process;

FIG. 13 is a view of showing a flicker curve filtered at a filtering unit shown in FIG. 4 ;

FIG. 14 is a view of calculating a distance between a local minimum point and a next local minimum point in a filtered flicker curve;

FIG. 15 is a histogram for illustrating coefficients exhibiting the frequency numbers of distance values; and

FIG. 16 is a block diagram of a flicker detecting circuit in accordance with another preferred embodiment of the present invention.

Hereinafter, the preferred embodiments of the present invention will be described with reference to the accompanying drawings. In the description and the drawings, the identical reference numerals refer to as the identical elements performing the same functions.

FIG. 3 is a conceptive view of illustrating a mechanism of detecting a flicker in an image sensor in accordance with a preferred embodiment of the present invention, and FIG. 4 is a block diagram of a flicker detecting circuit in the image sensor in accordance with the preferred embodiment of the present invention.

In FIG. 3 , the mechanism for detecting the flicker in the image senor in accordance with the preferred embodiment of the present invention consecutively photographs one picture several times, e.g., twice, and detects the flicker by subtracting luminance values of a second photographed image (second frame FRAME 2 ) of two photographed images from luminance values of a first photographed image (first frame FRAME 1 ) of the two photographed images. That is, when luminance values for each pixel line detected from the second frame FRAME 2 are subtracted from luminance values for each pixel line detected from the first frame FRAME 1 , information of an object in the first and the second frames FRAME 1 and FRAME 2 is removed and as a result there remains only a difference between the two frames, i.e., a lag of a flicker flow.

The flicker detecting circuit 100 shown in FIG. 4 includes a first luminance and chrominance average value generation unit 10 for detecting luminance values and chrominance values of pixel data RGB( 1 ) to calculate a luminance average value and a chrominance average value for each row pixel line in the first frame FRAME 1 , a first storage unit 20 for storing the luminance average value and the chrominance average value for each row pixel line in the first frame FRAME 1 output from the first luminance and chrominance average value generation unit 10 , a second luminance and chrominance average generation unit 30 for detecting luminance values and chrominance values of pixel data RGB( 2 ) to calculate a luminance average value and a chrominance average value for each row pixel line in the second frame FRAME 2 , a second storage unit 40 for storing the luminance average value and the chrominance average value for each row pixel line in the second frame FRAME 2 output from the second luminance and chrominance average value generation unit 30 , a motion estimating unit 50 for comparing the chrominance average value for each row pixel line in the first frame FRAME 1 and that of the second frame FRAME 2 respectively stored in the first and the second storage units 20 and 40 to estimate a motion of the second frame FRAME 2 , a flicker curve generating unit 60 for correcting the luminance average values of the second frame FRAME 2 stored in the second storage unit 40 based on the motion estimated result from the motion estimating unit 50 and subtracting the corrected luminance average values of the second frame FRAME 2 from the luminance average values of the first frame FRAME 1 , thereby generating a flicker curve, and a flicker image detecting unit 70 for determining by using the flicker curve whether a flicker image corresponding to the compared two frames is present or not.

Another preferred embodiment of the present invention may further comprise a flicker state determining unit 80 , which obtains a plurality of (in particular, three or more) frames from the same picture, determines whether a flicker image is present or not in the plurality of frames, and as a result, determines whether a flicker is present or not in the photographed image.

The first luminance and chrominance average value generation unit 10 has a first detector 11 for detecting a luminance value for each pixel in the first frame FRAME 1 , a second detector 14 for detecting a chrominance value Cb/Cr for each pixel in the first frame FRAME 1 , a first accumulator 12 for accumulating luminance values detected and output from the first detector 11 , a second accumulator 15 for accumulating chrominance values detected and output from the second detector 14 , a first average value calculator 13 for calculating luminance average values of the accumulated luminance values from the first accumulator 12 in a manner that the luminance average values are calculated for row pixel lines in the first frame FRAME 1 , and a second average value calculator 16 for calculating chrominance average values of the accumulated chrominance values from the second accumulator 15 in a manner that the chrominance average values are calculated for the row pixel lines in the first frame FRAME 1 .

The first storage unit 20 has a first storage 21 for storing the luminance average values output from the first average value calculator 13 , and a second storage 22 for storing the chrominance average values output from the second average value calculator 16 . The first and second storages 21 and 22 subsequently store the luminance and chrominance average values for all row pixel lines of the first frame FRAME 1 , respectively.

The second luminance and chrominance average value generation unit 30 has a third detector 31 for detecting a luminance value for each pixel in the second frame FRAME 2 , a fourth detector 34 for detecting a chrominance value Cb/Cr for each pixel in the second frame FRAME 2 , a third accumulator 32 for accumulating luminance values detected and output from the third detector 31 , a fourth accumulator 35 for accumulating chrominance values detected and output from the fourth detector 34 , a third average value calculator 33 for calculating luminance average values of the accumulated luminance values from the third accumulator 32 in a manner that the luminance average values are calculated for row pixel lines in the second frame FRAME 2 , and a fourth average value calculator 36 for calculating chrominance average values of the accumulated chrominance values from the fourth accumulator 35 in a manner that the chrominance average values are calculated for the row pixel lines in the second frame FRAME 2 .

The second storage unit 40 has a third storage 41 for storing the luminance average values output from the third average value calculator 33 , and a fourth storage 42 for storing the chrominance average values output from the fourth average value calculator 36 . The third and fourth storages 41 and 42 subsequently store the luminance and chrominance average values for all row pixel lines of the second frame FRAME 2 , respectively.

In this case, it is possible to use a storage device for storing codes digitalized by an analog-to-digital converter in an image sensor, e.g., a Static Random Access Memory (SRAM), as the first and second storage units 20 and 40 as mentioned above.

When an object represented in the first frame FRAME 1 and that represented in the second frame FRAME 2 are unmatched by a shake, the motion estimating unit 50 estimates an offset of the second frame FRAME 2 in order to correct the offset incurred in the second frame FRAME 2 .

That is, when an image pickup device photographs the first frame FRAME 1 and subsequently the second frame FRAME 2 , if the object appears to be moved by the shake, the object represented in the first frame FRAME 1 is unmatched to that represented in the second frame FRAME 2 , as a result of which the offset is incurred in the second frame FRAME 2 . The correction of such an offset is practicable by approaching the second frame FRAME 2 to the first frame FRAME 1 .

As described above, in obtaining a difference between the luminance values for each row pixel line in the first and the second frames FRAME 1 and FRAME 2 , when the object is moved, a flicker curve includes a lag with respect to the motion of the object as well as a flicker flow. Therefore, in the flicker curve, there is information of the object's motion in addition to information of the flicker. Such a phenomenon in a real life is usually caused by a shake when an image pickup device is supported by hands. Thus the flicker curve can be exactly extracted only after the phenomenon is compensated.

The motion estimating unit 50 estimates the object's motion by using the chrominance average values of the first and the second frames FRAME 1 and FRAME 2 respectively stored in the second and the fourth storages 22 and 42 . Since the object's motion caused by the shake is usually an up and down motion, the motion estimation is achieved by moving in an up and down direction addresses of the fourth storage 42 that stores the chrominance average values of the second frame FRAME 2 , in up and down direction.

An example of the motion (hereinafter, referred to as an offset) estimating method using the motion estimating unit 50 is shown in FIG. 5 .

Referring to FIG. 5 , in a certain region of the two frames FRAME 1 and FRAME 2 , a value of sum of absolute difference (SAD) is repetitively obtained with fixing one frame FRAME 1 and changing a reference value of the other frame FRAME 2 . At this time, as the reference value of the other frame FRAME 2 is changed, a point having a minimum SAD is found at a particular reference value and the particular reference value is stored in a storage. Then after searching a range of all reference values, a reference value having a minimum value in the range of all reference values is determined and applied as a motion reference value. This is described by an example shown in FIG. 6 .

In FIG. 6 , for the convenience of the explanation, it is assumed that the chrominance average values for row pixel lines LINE 0 to LINEN of the first frame FRAME 1 are sequentially stored in the second storage 22 and the chrominance average values for row pixel lines LINE 0 to LINEN of the second frame FRAME 2 are sequentially stored in the fourth storage 42 , N being a positive integer.

Rather than the chrominance average values stored in the storages 22 and 42 are not compared to each other according to the corresponding addresses in each of the storages 22 and 42 , the chrominance average values of the first frame FRAME 1 stored in the second storage 22 are fixed and the chrominance average values of the second frame FRAME 2 stored in the fourth storage 42 are varied by changing addresses of the fourth storage 42 , and then the chrominance average values of the second frame FRAME 2 are subjected to the comparison with and the subtraction from those of the first frame FRAME 1 , thereby estimating the offset in the second frame FRAME 2 . At this time, a variably changed address range is set to a certain value by a user. For example, the variably changed address range may be set within an address range ADD 0 to ADD 3 In a concrete way, as shown in FIG. 7 , while the second storage 22 sequentially increases a comparative address from ‘ADD 0 ’ to ‘ADDN’, and the fourth storage 42 sequentially increases a comparative address from ‘ADD 0 ’ to ‘ADDN’, the subtraction of the chrominance average values of the second frame FRAME 2 from those of the first frame FRAME 1 is executed and an SDA of the subtracted values is calculated.

First, a chrominance average value of the second frame FRAME 2 stored in the address ‘ADD 0 ’ of the fourth storage 42 is subtracted from the chrominance average value of the first frame FRAME 1 stored in the address ‘ADD 0 ’ of the second storage 22 , a chrominance average value of the second frame FRAME 2 stored in an address ‘ADD 1 ’ of the fourth storage 42 is subtracted from a chrominance average value of the first frame FRAME 1 stored in an address ‘ADD 1 ’ of the second storage 22 , a chrominance average value of the second frame FRAME 2 stored in an address ‘ADD 2 ’ of the fourth storage 42 is subtracted from the chrominance average value of the first frame FRAME 1 stored in an address ‘ADD 2 ’ of the second storage 22 , . . . and a chrominance average value of the second frame FRAME 2 stored in the address ‘ADDN’ of the fourth storage 42 is subtracted from a chrominance average value of the first frame FRAME 1 stored in the address ‘ADDN’ of the second storage 22 . Thus obtained subtracted values are summed as an SAD 0 .

Then, after increasing an address value of the fourth storage 42 by one, the subtraction with respect to the chrominance average values in the first and second frames FRAME 1 and FRAME 2 of each address of the second and fourth storages 22 and 42 is carried out. That is to say, the chrominance average value of the second frame FRAME 2 stored in the address ‘ADD 1 ’ of the fourth storage 42 is subtracted from the chrominance average value of the first frame FRAME 1 stored in the address ‘ADD 0 ’ of the second storage 22 , the chrominance average value of the second frame FRAME 2 stored in the address ‘ADD 2 ’ of the fourth storage 42 is subtracted from the chrominance average value of the first frame FRAME 1 stored in the address ‘ADD 1 ’ of the second storage 22 , the chrominance average value of the second frame FRAME 2 stored in an address ‘ADD 3 ’ of the fourth storage 42 is subtracted from the chrominance average value of the first frame FRAME 1 stored in the address ‘ADD 2 ’ of the second storage 22 , . . . and the chrominance average value of the second frame FRAME 2 stored in the address ‘ADDN’ of the fourth storage 42 is subtracted from a chrominance average value of the first frame FRAME 1 stored in an address ‘ADDNl ’ of the second storage 22 . Thus obtained subtraction values are summed as an SAD 1 .

The above process is carried out within a given address range, i.e., an address range of ADD 0 to ADD 3 . In this case, the address range may be suitably chosen in a range where a distortion in an image quality does not occur.

When one of the chrominance average values of the second frame FRAME 2 stored in the fourth storage 42 is closest to one of the chrominance average values of the first frame FRAME 1 stored in the second storage 22 through the above process, the luminance average value corresponding to that point is set to a final reference value.

For example, if a minimum value is SAD 1 among SAD 0 to SAD 3 in case where the address range is set to ADD 0 to ADD 3 , the minimum value SAD 1 is set to a final reference value. Therefore, the flicker curve generating unit 60 sets SAD 1 as a reference value in subtracting the luminance average values of the second frame FRAME 2 stored in the third storage 41 from the luminance average values of the first frame FRAME 1 stored in the first storage 21 .

The flicker curve generating unit 60 has a flicker curve extractor 61 for extracting a flicker curve based on the final reference value of the second frame FRAME 2 that is corrected based on the offset information estimated from the motion estimating unit 50 , and a filter 62 for filtering the flicker curve extracted at the flicker curve extractor 61 .

The flicker curve extractor 61 extracts the flicker curve by subtracting the luminance average values of the second frame FRAME 2 stored in the third storage 41 from the luminance average values of the first frame FRAME 1 stored in the first storage 21 . At this time, the luminance average values of the second frame FRAME 2 subtracted from the luminance average values of the first frame FRAME 1 are luminance average values corrected based on the offset estimated at the motion estimator 50 .

As an example, a case where the SAD 1 described above is the final reference value is described. As shown in FIG. 8 , the first storage 21 considers ‘ADD 0 ’ as a beginning point of a subtraction address and the third storage 41 considers ‘ADD 1 ’ as a beginning point of the subtraction address. That is to say, a luminance average value of the second frame FRAME 2 stored in the address ‘ADD 1 ’ of the third storage 41 is subtracted from a luminance average value of the first frame FRAME 1 stored in the address ‘ADD 0 ’ of the first storage 21 , a luminance average value of the second frame FRAME 2 stored in an address ‘ADD 2 ’ of the third storage 41 is subtracted from a luminance average value stored in an address ‘ADD 1 ’ of the first frame FRAME 1 of the first storage 21 , a luminance average value of the second frame FRAME 2 stored in an address ‘ADD 3 ’ of the third storage 41 is subtracted from a luminance average value of the first frame FRAME 1 stored in an address ‘ADD 2 ’ of the first storage 21 , . . . and a luminance average value of the second frame FRAME 2 stored in an address ‘ADD 0 ’ of the third storage 41 is subtracted from the luminance average value of the first frame FRAME 1 stored in an address ‘ADDN’ of the first storage 21 .

The filter 62 performs a low-pass filtering for the flicker curve from the flicker curve extractor 61 . The flicker curve from the flicker curve extractor 61 does not include only flicker information because of a light reception sensitivity and non-uniformity of the image pickup device and a circuit noise. Therefore, a low-pass filter is used to extract pure flicker curve information. Then information of the low-pass filter flicker curve from the flicker curve generating unit 60 is input to the flicker image detecting unit 70 .

The flicker image detector 70 has a local minimum point extractor 71 for extracting points exhibiting local minimum values (valley parts) of the low-pass filtered flicker curve in order to determine a periodicity of the low-pass filtered flicker curve, and a flicker image determiner 72 for using the extracted local minimum points to determine whether a flicker is present or not.

The flicker image determiner 72 obtains a distance between each two neighboring local minimum points among the extracted local minimum points, classifies the obtained distances into several distance ranges according to their values, calculates the frequency number showing the number of distances included in each distance range, and compares the frequency number for each distance range with a first threshold. Then, if the frequency number is larger than the first threshold, the flicker image determiner 72 multiplies the frequency number by an average distance value of a corresponding distance range, e.g., 5 if the distance range is from 0 to 10, and then determines that the flicker is present if the multiplied value is larger than a second threshold, i.e., when the periodicity of flicker curve is out of a certain range.

A flicker detecting operation described above may be repetitively executed with respect to a predetermined number of frames. Namely, the units 10 , 20 , 30 , 40 , 50 , 60 and 70 for detecting the flicker with respect to the first and second frames FRAME 1 and FRAME 2 repetitively execute the flicker detecting operation with respect to a predetermined number of consecutive frames, e.g., second and third frames, third and fourth frames, etc., and transmit flicker detecting information to the flicker state determining unit 80 .

The flicker state determining unit 80 counts the detecting number of flicker image in the flicker image detecting information transmitted from the flicker image detecting unit 70 . If the detecting number of flicker image with respect to the predetermined number of frames is equal to or larger than a third threshold, the flicker state determining unit 80 finally determines that a flicker is present in an entire photographed picture under current state and outputs a flicker state detecting signal. Thereafter, an automatic exposure device of the image sensor corrects an exposure time in response to the flicker state detecting signal.

FIG. 9 is a flow chart for illustrating a flicker detecting operation in accordance with a preferred embodiment of the present invention. This flicker detecting operation in the preferred embodiment of the present invention will be described hereinafter based on the flicker detecting circuit and operational characteristics thereof as shown in FIG. 4 .

Referring to FIG. 9 , at step S 91 , luminance average values and chrominance average values of pixel data RGB( 1 ) of the first FRAME 1 shown in FIG. 3 are calculated by the first luminance and chrominance average value generation unit 10 shown in FIG. 4 and then are respectively stored in the first and second storages 21 and 22 in a manner that the luminance average values and the chrominance average values are respectively classified with each row pixel line in the first frame FRAME 1 . In the rolling shutter scheme, data values of pixels in an image are sequentially output from respective row pixel lines. Therefore, as shown in FIG. 10 , luminance values and chrominance values are calculated and accumulated from the data values for pixels output from the respective row pixel lines, and then the luminance average values and the chrominance average values for the respective row pixel lines are respectively calculated and stored in the first storage 21 and the second storage 22 during a blank period between a previous row pixel line LINE 0 and a current row pixel line LINE 1 , i.e., when image values are not output, in response to a row pixel line data synchronization signal, i.e., a horizontal synchronization signal HSYC. That is, during the blank period between two row pixel lines, the luminance and chrominance average values of the previous row pixel line are calculated and stored in the first and the second storages 21 and 22 , respectively. Subsequently, in the same manner described above, luminance average values and chrominance average values of pixel data RGB( 2 ) of the second FRAME 2 are calculated by the second luminance and chrominance average value generation unit 30 and then are respectively stored in the third and the fourth storages 41 and 42 in a manner that the luminance average values and the chrominance average values are respectively classified with row pixel lines in the second frame FRAME 2 .

Then, at step S 92 , a motion correction due to the shake is carried out. The motion correction is for correcting a mismatching between the first frame FRAME 1 and the second frame FRAME 2 caused by the shake when the first frame FRAME 1 and subsequently the second frame FRAME 2 are consecutively photographed. If a difference between a luminance value of a row pixel line extracted from the first frame FRAME 1 and that of a row pixel line extracted from the second frame FRAME 2 is obtained under a state that the motion correction is not performed, a flicker curve having an undesired offset is extracted.

The motion correction is achieved in the same manner as described with reference to FIGS. 5 to 7 .

At step S 93 , a flicker curve is extracted by the flicker curve extractor 61 . In case where the second frame FRAME 2 does not have an undesired offset, i.e., there is no motion of the object, the flicker curve extractor 61 extracts a flicker curve by subtracting the luminance average values of the second frame FRAME 2 stored in the third storage 41 from those of the first frame FRAME 1 stored in the first storage 21 without the motion correction. Thus extracted flicker curve is shown in FIG. 11 . That is, by subtracting the luminance average values of the second frame FRAME 2 from those of the first frame FRAME 1 , detailed information for the object is cleared and, therefore, there is only a difference between the two frames, i.e., a lag of a flicker flow. In other words, the lag originated from the difference between the two frames results in the flicker curve shown in FIG. 11 .

FIG. 12 is a block diagram for comparing a flicker curve without the motion correction and a flicker curve with the motion correction. That is, an upper waveform represents the flicker curve without the motion correction and a lower waveform represents the flicker curve with the motion correction.

As seen from FIG. 12 , since the motion caused by the hand shake in up and down directions should be corrected to obtain an exactly estimated result, a flicker curve can be more exactly extracted with the motion correction than that extracted without the motion correction.

Subsequently, at step S 94 , the extracted flicker curve is filtered. The flicker curve extracted by the flicker curve extractor 61 does not include exact flicker information due to a light reception sensitivity and non-uniformity of the image pickup device and a circuit noise. Therefore, at step S 94 , in order to solve this problem, the extracted flicker curve is low-pass filtered in the filer 62 such that only the exact flicker information is extracted.

FIG. 13 is a view for comparing a flicker curve with a filtered flicker extracted after the low-pass filtering is carried out. That is, an upper waveform represents the flicker curve and a lower waveform represents the filtered flicker curve.

At step S 95 , a periodicity of the filtered flicker curve is determined and then it is determined whether the periodicity is originated from the flicker band.

The method of determining the periodicity thereof will be described hereinafter.

First, as shown in FIG. 14 , a plurality of local minimum points of the flicker curve are obtained by the local minimum point extractor 71 .

Then, at step S 96 , the flicker image determiner 72 figures out the distance Dist between two neighboring local minimum points of the obtained local minimum points and determines the flicker curve as a curve formed by the flicker band based on the distances of the local minimum points as described above with reference to FIG. 4 .

As exemplarily illustrated in FIG. 15 , in order to obtain the frequency numbers of the distances between local minimum points in the flicker curve, the distances are counted and recorded in a storage medium, such as a SRAM, in the form of histogram having a plurality of distance ranges classified according to the distance values. A coefficient recorded in the form of histogram represents the number of distances accumulated in a distance range corresponding to the coefficient, i.e., the frequency number. Then, the coefficients for the plurality of distance ranges are respectively compared with a threshold, e.g., a first threshold and, if there is at least one coefficient larger than the first threshold, the coefficient is multiplied with an average distance value of a corresponding distance range and the multiplied value becomes a representative value exhibiting a degree of flicker displayed on a display, and a flicker is detected when the representative value is larger than a threshold, e.g., a second threshold.

The extraction of the flicker curve and the determination of the flicker as described above can be repetitively executed for a predetermined number of frames, and lead to determine whether each frame includes a flicker, thereby counting the detecting number of the flicker.

That is, at step S 97 , it is determined whether the flicker detecting operation is repetitively executed with respect to the predetermined number of frames. If the detecting operation is not completed for the predetermined number of frames, then the process is returned back to the step S 91 to repeat the flicker detecting operation for a subsequent frame; and if otherwise, then the process proceeds to step S 98 .

At step S 98 , the flicker state determining unit 80 compares the detecting number of flicker with a threshold, e.g., a third threshold, to determine that a flicker is present in the photographed image if the detecting number is equal to or larger than the third threshold and reports this result to an automatic exposure device. Then the automatic exposure device recalculates an exposure time based on the flicker detection result.

FIG. 16 shows a flicker detecting circuit 200 in accordance with another preferred embodiment of the present invention. While the flicker detecting circuit 100 in accordance with the preferred embodiment shown in FIG. 4 detects the flicker curve with respect to the two neighboring frames to detect a flicker, the flicker detecting circuit 200 shown in FIG. 16 is implemented such that flicker curves are extracted in parallel with respect to a plurality of, e.g., M, frames, M being an integer larger than 2.

The flicker detecting circuit 200 includes an average value generation unit 110 ( 110 - 1 to 110 -M) for detecting luminance and chrominance average values of respective row pixel lines for the M frames, a storage unit 120 ( 120 - 1 to 120 -M) for storing the average values in a manner that the luminance average values and the chrominance average values are classified with the respective row pixel lines with respect to the M frames, a motion estimating unit 150 for estimating motions of a second to Mth frames, a flicker curve generating unit 160 for compensating a motion of a corresponding frame based on the estimated result of the motion estimating unit 150 and extracting a plurality of flicker curves (e.g., M−1 flicker curves) caused by differences between respective two neighboring frames, a flicker image detecting unit 170 for determining whether each of the flicker curves has a flicker image by using the M−1 flicker curves, and a flicker state determining unit 180 for counting the detecting number of flicker, determining that there exists a flicker in the photographed image if the detecting number of flicker is equal to or larger than a threshold, and transmitting the determined result to an automatic exposure device in the image sensor. These components in the flicker detecting circuit 200 similarly operate as described referring to FIG. 4 and thus details of the construction and operation thereof are not described herein.

As described above, the present invention is advantageously independent on an object by using differences of luminance values of at least two frames to extract flicker curves, and can avoid efficiency degradation in a semiconductor-implemented area because it is constructed of purely digital processing circuits.

Further, the present invention does not degrade efficiency in a semiconductor-implemented area because it does not employ a separate operating circuit by using a histogram in determining a periodicity of a flicker curve to reuse a storage medium, e.g., SRAM, built in an image sensor.

The present application contains subject matters related to the Korean patent application No. KR 2005-0127857, file in the Korean Patent Office on Dec. 22, 2005, the entire contents of which being incorporated herein by reference.

While the scope of the present invention has been described with respect to the preferred embodiments, these embodiments are for illustrating the present invention and are not to be construed to limit the present invention. In other words, in accordance with the present invention, it is possible to determine whether a flicker is present or not in all system which photograph an image by using a rolling shutter scheme. Further, those skilled in the art will appreciate that various modifications, changes and variations can be made to the preferred embodiments of the present invention as fall within the true spirit and scope of this present invention.