JP2012222739A - Flicker correction device, flicker correction method and flicker correction program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technique for detecting and correcting a frequency of a flicker at high speed even if a subject moves or a gap of phases between a field period and a flicker period is small.SOLUTION: The number n of fields in a line average value storing section before n fields 103 of an image signal imaged by using an imaging element whose exposure time is point sequential or line sequential, a cyclic coefficient in first circulation-type LPF 108 and second circulation-type LPF 109 for a frequency region of an extracted flicker component and a fade gain in a fade gain multiplication section 118 are controlled based on a movement detection result of a movement detector 110, and the flicker is corrected.

Description

  The present invention relates to a flicker correction technique for correcting a periodic change (hereinafter, referred to as flicker) in an output signal of an image sensor mainly due to illumination by an AC power source.

  In general, lighting (bulb, fluorescent lamp, mercury lamp, etc.) lit by an AC power source repeats light and dark with luminance change at a frequency twice the power frequency. When an object is imaged using an imaging device whose exposure time is point-sequential or line-sequential under such illumination, periodic horizontal stripes, so-called flickers, are generated in the image signal output from the imaging device. The period of the periodic horizontal stripes is obtained from the drive frequency and power supply frequency of the image sensor.

  For example, since a general fluorescent lamp that is lit by an AC power supply of 50 Hz repeats light emission at 100 Hz, in the NTSC system, the horizontal scanning (hereinafter referred to as a line) frequency is 15.75 kHz, so 1/100 sec = 157.5 lines. Repeat light and dark every time. Further, since the common multiple of the field period (1 / 59.94 sec) and the lighting blinking period (1/100 sec) is 1/20 sec, the same light / dark pattern is obtained every 1/20 sec, that is, almost every three fields.

  Further, since the fluorescent lamp that is turned on with an AC power supply of 60 Hz repeats light emission at 120 Hz, light / darkness is repeated every 1/120 sec = 131.3 lines. In addition, since the difference between the field period (1 / 59.94 sec) and the lighting blinking period (1/120 sec) is 16.67 μsec, the period until the field period and flicker period again match each other is every 500 fields. And appears on the screen as horizontal stripes that move very slowly. As described above, the image quality deterioration due to flicker is very disturbing.

  With respect to such a problem, in the technique disclosed in Patent Document 1, a line average value is stored in a memory for each 1/3 phase of a period in which the phase of the field period and the flicker period again coincide with each other. A reference value that eliminates the influence of flicker is created by averaging the average values, and flicker correction is performed. In particular, in the case of receiving 60 Hz flicker in the NTSC system, flicker correction is performed when the subject is stationary, but processing is performed so that the flicker correction is not performed when the subject is moving, so that it is difficult to be influenced by the motion of the subject. A flicker reduction device is realized.

JP 2009-81684 A

  However, in the conventional method, the memory for storing the line average value is updated every 1/3 phase of the cycle in which the phase of the field cycle and the flicker cycle again match, and in particular, the 60 Hz flicker in the NTSC system. In this case, three average values of an average value delayed for 167 field periods, an average value delayed for 334 field periods, and an average value delayed for 500 field periods are obtained to create a reference value that eliminates the influence of flicker. For this reason, if there is a movement in the 500 field period, a correct reference value cannot be obtained, so that a flicker cannot be detected, or at least the first flicker is determined even if it is a stationary subject. There was a problem that it took more than 500 field periods.

  The present invention has been made in order to solve the above-described problems of the conventional method. The object of the present invention is to perform flicker correction even when the subject moves, and further, the phase of the field period and flicker period. An object of the present invention is to provide a flicker correction technique that can start flicker correction at a high speed when the moving state changes to a stationary state even when the deviation is small.

  A flicker correction apparatus according to a first aspect of the present invention is a flicker correction apparatus that reduces flicker of an image picked up using an image pickup device whose exposure time is dot-sequential or line-sequential. Obtained by the line average value calculating unit, the storage unit for storing the average value obtained by the line average value calculating unit in one field period unit, and the line average value calculating unit. A motion detection unit that detects motion from an average value and an output change of the average value of the previous field stored in the storage unit, and the number of fields that remain stationary based on the motion detection result output from the motion detection unit A stationary field number measurement unit that measures the number of stationary fields, and a motion field number that is the number of fields that have continued to move based on the motion detection result output from the motion detection unit. The difference value between the average value obtained by the motion field number measuring unit to be measured, the line average value calculating unit, and the previous average value by n fields (n is an integer value) stored in the storage unit is calculated. The difference value calculation unit, the field number n calculation unit for calculating the field number n based on the result of the number of static fields output from the still field number measurement unit, and the average obtained by the line average value calculation unit A flicker component extraction unit that extracts a flicker frequency from a difference value obtained by the value and the difference value calculation unit, a flicker component extracted by the flicker component extraction unit, and a flicker frequency based on a result of the motion detection unit and a flicker frequency A flicker determination unit that calculates a flicker correction gain synchronized with the flicker frequency from the result of the flicker determination unit, A flicker correction gain multiplication unit that multiplies the input image signal by a correction gain from a flicker correction gain calculation unit, and the field number n calculation unit includes the number of fields according to the number of still fields and the number of motion fields. When the number of stationary fields is not zero, the number of fields n is reduced while the number of stationary fields is small, while the number of fields n is increased to a predetermined number as the number of stationary fields increases. The number of fields n is gradually increased, and when the number of motion fields is not zero, the number of fields n is set to zero.

  With this configuration, it is possible to appropriately change the flicker component extraction speed in accordance with the movement / still state of the subject, and the flicker component calculation result can be obtained quickly when switching from the movement to the stationary state. When the state continues for a long time, flicker component calculation can be obtained with high accuracy.

  A flicker correction apparatus according to a second aspect of the present invention is a flicker correction apparatus that reduces flicker of an image captured using an image sensor whose exposure time is point-sequential or line-sequential. Obtained by the line average value calculating unit, the storage unit for storing the average value obtained by the line average value calculating unit in one field period unit, and the line average value calculating unit. A motion detection unit that detects motion from an average value and an output change of the average value of the previous field stored in the storage unit, and the number of fields that remain stationary based on the motion detection result output from the motion detection unit A stationary field number measurement unit that measures the number of stationary fields, and a motion field number that is the number of fields that have continued to move based on the motion detection result output from the motion detection unit. The difference value between the average value obtained by the motion field number measuring unit to be measured, the line average value calculating unit, and the previous average value by n fields (n is an integer value) stored in the storage unit is calculated. A difference value calculation unit, a flicker component extraction unit that extracts a flicker frequency from the average value obtained by the line average value calculation unit and the difference value obtained by the difference value calculation unit, and the flicker component extraction unit A flicker determination unit that determines the presence / absence of flicker and a flicker frequency from the flicker component and the result of the motion detection unit; a flicker correction gain calculation unit that calculates a flicker correction gain synchronized with the flicker frequency from the result of the flicker determination unit; A fade gain calculation unit for calculating a fade gain from the measurement results of the stationary field number measurement unit and the motion field number measurement unit, and the fade A fade gain multiplication unit that multiplies the flicker correction gain by the fade gain calculated by the in-calculation unit, and a flicker correction gain multiplication unit that multiplies the input image signal by the correction gain input from the fade gain multiplication unit, The fade gain calculation unit controls the fade gain in accordance with the number of still fields and the number of motion fields. When the number of still fields is not zero, the fade gain is reduced while the number of still fields is small. On the other hand, as the number of static fields increases, the fade gain is gradually increased and increased to 1 time. When the number of motion fields is not zero, the fade gain is increased while the number of motion fields is small. While the number of motion fields increases Accordingly, the fade gain is gradually decreased and reduced to 0 times.

  With this configuration, it is possible to appropriately change the magnitude of the fade gain according to the movement / still state of the subject. When the stationary state continues for a long time and flicker correction accuracy is reliable, a normal correction gain is output. If flicker correction accuracy is not very reliable, such as when the motion continues for a long time or immediately after the motion has stopped, the correction gain is reduced to make the negative effect of the correction error noticeable. I can not.

  A flicker correction apparatus according to a third aspect of the present invention is a flicker correction apparatus that reduces flicker of an image captured using an image sensor whose exposure time is point-sequential or line-sequential, and is an effective period of an input image signal Obtained by the line average value calculating unit, the storage unit for storing the average value obtained by the line average value calculating unit in one field period unit, and the line average value calculating unit. A motion detection unit for detecting motion from an average value and an output change of the average value of the previous field stored in the storage unit; an average value obtained by the line average value calculation unit; A difference value calculation unit that calculates a difference value from the previous average value by n fields (n is an integer value), an average value obtained by the line average value calculation unit, and a difference obtained by the difference value calculation unit Flicker frequency from value A flicker component extracting unit that extracts a cyclic coefficient based on the motion detection result output from the motion detecting unit, and calculating the cyclic coefficient to the flicker component extracted by the flicker component extracting unit. A cyclic low-pass filter unit that applies a low-pass filter based on the cyclic coefficient calculated by the unit, and a flicker that determines the presence / absence of flicker and the flicker frequency based on the output from the cyclic low-pass filter unit and the detection result of the motion detection unit A flicker correction gain calculation unit that calculates a flicker correction gain synchronized with a flicker frequency based on a determination result of the determination unit, the flicker determination unit, and the correction gain calculated by the flicker correction gain calculation unit; A flicker correction gain multiplication unit for multiplying by the cyclic coefficient calculation unit In one field immediately after the state is changed to the stationary state, the cyclic coefficient is once set to zero, and when the stationary state continues, the cyclic coefficient is set to a predetermined value, and when the moving state is entered, the cyclic coefficient is set to 1. It is characterized by that.

  With this configuration, it is possible to appropriately change the magnitude of the cyclic coefficient of the cyclic low-pass filter according to the movement / still state of the subject, and by initializing the calculation results once when the movement has stopped. The cyclic low-pass filter output can be converged at high speed to obtain the desired flicker correction gain, and the cyclic low-pass filter output is maintained even when the motion state continues for a while, so that the flicker correction gain fluctuates. This makes it possible to perform flicker correction stably.

  A flicker correction method according to a first aspect of the present invention is a flicker correction method for reducing flicker of an image picked up using an image pickup device whose exposure time is dot-sequential or line-sequential. A line average value calculating step for calculating an average value for each line, a storing step for storing the average value obtained in the line average value calculating step in one field period unit, and a line average value calculating step. A motion detection step for detecting motion from the average value and an output change of the average value of the previous field stored in the storage step, and a stillness that is the number of fields in which stillness has continued based on the motion detection result in the motion detection step A stationary field number measuring step for measuring the number of fields, and a motion that continues to move based on the motion detection result in the motion detection step. A motion field number measurement step for measuring the number of motion fields, which is the number of motion fields, an average value obtained in the line average value calculation step, and an average value previous by n fields (n is an integer value) stored in the storage step A difference value calculation step for calculating a difference value between the field number n, a field number n calculation step for calculating the field number n based on the result of the number of still fields measured in the number of still fields measurement step, and the line average value The flicker component extraction step for extracting the flicker frequency from the average value obtained in the calculation step and the difference value obtained in the difference value calculation step, the flicker component extracted in the flicker component extraction step, and the motion detection step A flicker determination step for determining presence / absence of flicker and a flicker frequency based on a motion detection result; From the determination result in the flicker determination step, a flicker correction gain calculation step for calculating a flicker correction gain synchronized with a flicker frequency, and a flicker correction gain for multiplying the input image signal by the correction gain calculated in the flicker correction gain calculation step. A step of calculating the number of fields n, wherein the number of fields n is controlled according to the number of still fields and the number of motion fields, and when the number of still fields is not zero, While the number of fields is small, the number of fields n is reduced, while the number of fields n is gradually increased to a predetermined number as the number of static fields increases, and when the number of motion fields is not zero, the field number Set the number n to zero Features.

  With this configuration, it is possible to appropriately change the flicker component extraction speed in accordance with the movement / still state of the subject, and the flicker component calculation result can be obtained quickly when switching from the movement to the stationary state. When the state continues for a long time, flicker component calculation can be obtained with high accuracy.

  A flicker correction method according to a second aspect of the present invention is a flicker correction method for reducing flicker of an image picked up using an image pickup device whose exposure time is point-sequential or line-sequential. A line average value calculating step for calculating an average value for each line, a storing step for storing the average value obtained in the line average value calculating step in one field period unit, and a line average value calculating step. A motion detection step for detecting motion from the average value and an output change of the average value of the previous field stored in the storage step, and a stillness that is the number of fields in which stillness has continued based on the motion detection result in the motion detection step A stationary field number measuring step for measuring the number of fields, and a motion that continues to move based on the motion detection result in the motion detection step. A motion field number measurement step for measuring the number of motion fields, which is the number of motion fields, an average value obtained in the line average value calculation step, and an average value previous by n fields (n is an integer value) stored in the storage step A difference value calculating step for calculating a difference value between, and a flicker component extracting step for extracting a flicker frequency from the average value obtained in the line average value calculating step and the difference value obtained in the difference value calculating step; Based on the flicker component extracted in the flicker component extraction step and the motion detection result in the motion detection step, the flicker determination step for determining the presence / absence of flicker and the flicker frequency, and on the basis of the determination result in the flicker determination step A flicker correction gain calculating step for calculating a flicker correction gain synchronized with the frequency; A fade gain calculating step for calculating a fade gain from the measurement results of the stationary field number measuring step and the motion field number measuring step; and a fade gain multiplying step for multiplying the flicker correction gain by the fade gain calculated in the fade gain calculating step. A flicker correction gain multiplication step for multiplying the input image signal by the correction gain from the fade gain multiplication step, and in the fade gain calculation step, the fade gain is calculated according to the number of still fields and the number of motion fields. When the number of stationary fields is not zero, the fade gain is decreased while the number of stationary fields is small, while the fade gain is gradually increased as the number of stationary fields increases. When the number of motion fields is not zero, the fade gain is increased while the number of motion fields is small, while the fade gain is gradually decreased as the number of motion fields increases. It is reduced to 0 times.

  With this configuration, it is possible to appropriately change the magnitude of the fade gain according to the movement / still state of the subject. When the stationary state continues for a long time and flicker correction accuracy is reliable, a normal correction gain is output. If flicker correction accuracy is not very reliable, such as when the motion continues for a long time or immediately after the motion has stopped, the correction gain is reduced to make the negative effect of the correction error noticeable. I can not.

  A flicker correction method according to a third aspect of the present invention is a flicker correction method for reducing flicker of an image picked up using an image pickup device whose exposure time is dot-sequential or line-sequential. A line average value calculating step for calculating an average value for each line, a storing step for storing the average value obtained in the line average value calculating step in one field period unit, and a line average value calculating step. The motion detection step for detecting motion from the average value and the output change of the average value one field before stored in the storage step, the average value obtained in the line average value calculation step, and the n stored in the storage step Difference value calculation step for calculating a difference value from the previous average value by the field (n is an integer value), and the average obtained in the line average value calculation step A flicker component extracting step for extracting a flicker frequency from the difference value obtained in the difference value calculating step, a cyclic coefficient calculating step for calculating a cyclic coefficient based on a motion detection result in the motion detecting step, and the flicker component A cyclic low-pass filter step for applying a low-pass filter to the flicker component extracted in the extraction step based on the cyclic coefficient calculated in the cyclic coefficient calculation step; an output from the cyclic low-pass filter step; and detection in the motion detection step A flicker determination step for determining presence / absence of flicker and a flicker frequency based on the result, a flicker correction gain calculation step for calculating a flicker correction gain synchronized with the flicker frequency based on a determination result in the flicker determination step, and the flicker Correction gay A flicker correction gain multiplication step for multiplying the input image signal by the correction gain calculated in the calculation step. In the cyclic coefficient calculation step, the cyclic coefficient is temporarily set to zero once in one field immediately after the moving state is changed to the stationary state. The cyclic coefficient is set to a predetermined value when the stationary state continues, and the cyclic coefficient is set to 1 when the moving state is reached.

  With this configuration, it is possible to appropriately change the magnitude of the cyclic coefficient of the cyclic low-pass filter according to the movement / still state of the subject, and by initializing the calculation results once when the movement has stopped. The cyclic low-pass filter output can be converged at high speed to obtain the desired flicker correction gain, and the cyclic low-pass filter output is maintained even when the motion state continues for a while, so that the flicker correction gain fluctuates. This makes it possible to perform flicker correction stably.

A flicker correction program according to a first aspect of the present invention is a flicker correction program for reducing flicker of an image picked up using an image pickup device whose exposure time is point-sequential or line-sequential. This is a flicker correction program for execution.
A line average value calculating step for calculating an average value for each line during the effective period of the input image signal:
A storage step of storing the average value obtained in the line average value calculation step in units of one field period:
A motion detection step of detecting motion from the output change of the average value obtained in the line average value calculation step and the average value of the previous field stored in the storage step:
A stationary field number measuring step of measuring the number of stationary fields, which is the number of fields that have been stationary based on the motion detection result in the motion detecting step:
A motion field number measurement step of measuring the number of motion fields, which is the number of fields in which motion has continued, based on the motion detection result in the motion detection step:
A difference value calculating step of calculating a difference value between the average value obtained in the line average value calculating step and the average value previous by n fields (n is an integer value) stored in the storing step:
Calculating the number of fields n based on the result of the number of stationary fields measured in the stationary field number measuring step, and controlling the number of fields n according to the number of stationary fields and the number of motion fields. When the number of stationary fields is not zero, the number of fields n is reduced while the number of stationary fields is small, while the number of fields n is gradually increased to a predetermined number as the number of stationary fields increases. The field number n calculating step of setting the field number n to zero when the motion field number is not zero.
Flicker component extraction step for extracting a flicker frequency from the average value obtained in the line average value calculation step and the difference value obtained in the difference value calculation step:
Flicker determination step for determining presence / absence of flicker and flicker frequency based on the flicker component extracted in the flicker component extraction step and the motion detection result in the motion detection step:
Flicker correction gain calculation step for calculating a flicker correction gain synchronized with the flicker frequency from the determination result in the flicker determination step:
A flicker correction gain multiplication step for multiplying the input image signal by the correction gain calculated in the flicker correction gain calculation step:

  With this configuration, it is possible to appropriately change the flicker component extraction speed in accordance with the movement / still state of the subject, and the flicker component calculation result can be obtained quickly when switching from the movement to the stationary state. When the state continues for a long time, flicker component calculation can be obtained with high accuracy.

A flicker correction program according to a second aspect of the present invention is a flicker correction program for reducing flicker of an image captured using an image sensor whose exposure time is point-sequential or line-sequential. This is a flicker correction program for execution.
A line average value calculating step for calculating an average value for each line during the effective period of the input image signal:
A storage step of storing the average value obtained in the line average value calculation step in units of one field period:
A motion detection step of detecting motion from the output change of the average value obtained in the line average value calculation step and the average value of the previous field stored in the storage step:
A stationary field number measuring step of measuring the number of stationary fields, which is the number of fields that have been stationary based on the motion detection result in the motion detecting step:
A motion field number measurement step of measuring the number of motion fields, which is the number of fields in which motion has continued, based on the motion detection result in the motion detection step:
A difference value calculating step of calculating a difference value between the average value obtained in the line average value calculating step and the average value previous by n fields (n is an integer value) stored in the storing step:
Flicker component extraction step for extracting a flicker frequency from the average value obtained in the line average value calculation step and the difference value obtained in the difference value calculation step:
Flicker determination step for determining presence / absence of flicker and flicker frequency based on the flicker component extracted in the flicker component extraction step and the motion detection result in the motion detection step:
A flicker correction gain calculating step for calculating a flicker correction gain synchronized with the flicker frequency based on the determination result in the flicker determination step:
A step of calculating a fade gain from the measurement results of the stationary field number measuring step and the motion field number measuring step, wherein the fade gain is controlled in accordance with the number of stationary fields and the number of motion fields, and the number of stationary fields When the stationary field is not zero, the fade gain is decreased while the number of stationary fields is small, while the fade gain is gradually increased and increased to 1 time as the number of stationary fields increases. When the motion is not zero, the fade gain is increased while the number of motion fields is small, while the fade gain is gradually decreased and reduced to 0 times as the number of motion fields increases.
A fade gain multiplication step of multiplying the flicker correction gain by the fade gain calculated in the fade gain calculation step:
Flicker correction gain multiplication step of multiplying the input image signal by the correction gain from the fade gain multiplication step:

  With this configuration, it is possible to appropriately change the magnitude of the fade gain according to the movement / still state of the subject. When the stationary state continues for a long time and flicker correction accuracy is reliable, a normal correction gain is output. If flicker correction accuracy is not very reliable, such as when the motion continues for a long time or immediately after the motion has stopped, the correction gain is reduced to make the negative effect of the correction error noticeable. I can not.

A flicker correction program according to a third aspect of the present invention is a flicker correction program for reducing flicker of an image picked up using an image pickup device whose exposure time is point-sequential or line-sequential. This is a flicker correction program for execution.
A line average value calculating step for calculating an average value for each line during the effective period of the input image signal:
A storage step of storing the average value obtained in the line average value calculation step in units of one field period:
A motion detection step of detecting motion from the output change of the average value obtained in the line average value calculation step and the average value of the previous field stored in the storage step:
A difference value calculating step of calculating a difference value between the average value obtained in the line average value calculating step and the average value previous by n fields (n is an integer value) stored in the storing step:
Flicker component extraction step for extracting a flicker frequency from the average value obtained in the line average value calculation step and the difference value obtained in the difference value calculation step:
In the step of calculating a cyclic coefficient based on the motion detection result in the motion detection step, the cyclic coefficient is temporarily set to zero for one field immediately after the moving state is changed to the stationary state, and when the stationary state continues. While the cyclic coefficient is set to a predetermined value, the cyclic coefficient is set to 1 when a motion state is reached:
A cyclic low-pass filter step of applying a low-pass filter to the flicker component extracted in the flicker component extraction step based on the cyclic coefficient calculated in the cyclic coefficient calculation step:
Flicker determination step of determining the presence / absence of flicker and the flicker frequency based on the output from the cyclic low-pass filter step and the detection result in the motion detection step:
A flicker correction gain calculating step for calculating a flicker correction gain synchronized with the flicker frequency based on the determination result in the flicker determination step:
A flicker correction gain multiplication step for multiplying the input image signal by the correction gain calculated in the flicker correction gain calculation step:

  With this configuration, the cyclic coefficient of the cyclic low-pass filter can be appropriately changed according to the movement / still state of the subject, and the calculation result can be initialized once when the movement is stopped. The cyclic low-pass filter output can be converged at high speed to obtain a desired flicker correction gain, and the cyclic low-pass filter output is retained even when the motion state continues for a while, so that the flicker correction gain fluctuates. This makes it possible to perform flicker correction stably.

  According to the present invention, flicker correction can be performed even when the subject moves, and even when there is little phase shift between the field period and the flicker period, flicker is performed at high speed when the moving state changes to the stationary state. An excellent effect that correction can be started is obtained.

The block diagram which shows the structure of the flicker correction apparatus in one Example of this invention. The figure which shows an example of the image which the flicker has generate | occur | produced The figure which shows the line average value of the image of FIG. The figure which shows the line average value of 3 fields Figure showing the result of dividing the line average value of the current field by the average value of three fields The figure which shows the difference value of the line average value of the present field and n field previous Figure showing the result of dividing the difference value by the line average value of the current field A diagram showing state control according to a motion determination result example (A) A diagram showing a measurement result of the number of still fields (B) A diagram showing a measurement result of the number of motion fields (C) A diagram showing a calculation result of the number of fields n (D) A cyclic coefficient Figure showing the calculation result (E) Figure showing the fade gain calculation result

  Hereinafter, a flicker correction apparatus that reduces flicker of an image captured using an image sensor having an exposure time that is dot-sequential or line-sequential according to an embodiment of the present invention will be described with reference to the drawings.

  FIG. 1 is a diagram showing an embodiment of a flicker correction apparatus according to the present invention. In FIG. 1, reference numeral 100 denotes a line average value calculation unit, reference numeral 101 denotes a line average value before one field storage unit, reference numeral 102 denotes a line average value before two fields storage unit, and reference numeral 103 denotes a line average value storage unit before n fields. The line average value storage unit 101 before 1 field, the line average value storage unit 102 before field, and the line average value storage unit 103 before n field store the average value obtained by the line average value calculation unit 100 in units of one field period. The storage unit for storing is configured.

  Reference numeral 104 denotes a three-field average value calculation unit, 105 denotes a difference value calculation unit, 106 denotes a 50 Hz flicker component extraction unit, 107 denotes a 60 Hz flicker component extraction unit, 108 denotes a first cyclic LPF unit, and 109 denotes a second value. Cyclic LPF unit, 110 a motion detection unit, 111 a motion field number measurement unit, 112 a still field number measurement unit, 113 a field number n calculation unit, 114 a cyclic coefficient calculation unit, 115 a flicker determination unit, and 116 A flicker correction gain calculation unit, 117 is a fade gain calculation unit, 118 is a fade gain multiplication unit, 119 is a flicker correction gain multiplication unit, 120 is a delay unit, and 121 is a microcomputer.

  The line average value calculation unit 100 calculates an average value for each line of the effective period of the input image signal.

  The one-field previous line average value storage unit 101 and the two-field previous line average value storage unit 102 can hold a line average value for one field, and outputs each line with a delay of one field period. Similarly, the line average value storage unit 103 before n fields can hold the line average value for one field, and outputs it after delaying an arbitrary n field period.

  In the 3-field average value calculation unit 104, the output y0 of the line average value calculation unit 100, the output y-1 of the previous line average value storage unit 101, the average value of the output y-2 of the previous line average value storage unit 102 Is calculated. That is, the average of the line average values at the same line position for every three fields is calculated, and a reference value that eliminates the influence of flicker is created.

  The difference value calculation unit 105 subtracts the n field previous line average value output from the n field previous line average value storage unit 103 from the line average value output from the line average value calculation unit 100.

  The 50 Hz flicker component extraction unit 106 divides the line average value output from the line average value calculation unit 100 by the average value output from the three-field average value calculation unit 104, thereby eliminating the flicker signal from which the density of the subject has been eliminated. Then, the frequency is converted, and only the frequency region of the 50 Hz flicker component is extracted.

  The 60 Hz flicker component extraction unit 107 divides the difference value output from the difference value calculation unit 105 by the average value output from the line average value calculation unit 100 to obtain a flicker signal from which the density of the subject is excluded. After that, frequency conversion is performed, and only the frequency region of the 60 Hz flicker component is extracted.

  The first cyclic LPF unit 108 performs low pass with the cyclic coefficient output from the cyclic coefficient calculation unit 114 on the 50 Hz flicker component output from the 50 Hz flicker component extraction unit 106 and the flicker component cyclic LPF result of the previous field. Apply a filter.

  The second cyclic LPF unit 109 low-passes the cyclic coefficient output from the cyclic coefficient calculation unit 114 with respect to the 60 Hz flicker component output from the 60 Hz flicker component extraction unit 107 and the flicker component cyclic LPF result of n fields before. Apply a filter.

  The motion detection unit 110 compares the difference between the line average value output from the line average value calculation unit 100 and the average value output from the one-field previous line average value storage unit 101 with a certain threshold value (threshold value). Detecting the movement of

  The motion field number measuring unit 111 measures the number of fields in which motion has continued (number of motion fields) based on the motion detection result output from the motion detection unit 110.

  The stationary field number measuring unit 112 measures the number of fields (still field number) that have been stationary based on the motion detection result output from the motion detecting unit 110.

  The field number n calculation unit 113 calculates (determines) the field number n based on the result of the number of still fields output from the still field number measurement unit 112.

  The cyclic coefficient calculation unit 114 determines a cyclic coefficient based on the motion detection result output from the motion detection unit 110.

  The flicker determination unit 115 adjusts flicker components based on the presence or absence of motion and the presence or absence of flicker based on the detection result of the motion detection unit 110 and the outputs from the first cyclic LPF unit 108 and the second cyclic LPF unit 109. And frequency determination (50 Hz or 60 Hz).

  Based on the determination result of the flicker determination unit 115, the flicker correction gain calculation unit 116 calculates (generates) a sine wave synchronized with the flicker frequency by superimposing trigonometric functions.

  The fade gain calculation unit 117 calculates a fade gain based on the measurement results of the motion field number measurement unit 111 and the stationary field number measurement unit 112.

  The fade gain multiplication unit 118 multiplies the flicker correction gain output from the flicker correction gain calculation unit 116 by the fade gain output from the fade gain calculation unit 117.

  The flicker correction multiplication unit 119 multiplies the image signal subjected to the delay adjustment by the delay unit 120 by the flicker correction gain output by the fade gain multiplication unit 118 and multiplied by the fade gain, thereby flickering the input image. Reduce.

  The delay unit 120 adjusts a delay corresponding to generation of a correction gain for an input image signal.

  The operation of the flicker correction apparatus of the present embodiment configured as described above will be described below. The content of the operation is a flicker correction method for reducing flicker of an image captured using an image sensor whose exposure time is dot-sequential or line-sequential, and each operation (step) is executed using a computer. Obviously it can be done. The program is a flicker correction program for reducing flicker of an image captured using an image sensor whose exposure time is point-sequential or line-sequential, and causes the computer to execute the following operations (steps). This is a flicker correction program.

  As shown in FIG. 1, an image signal captured using an image sensor whose exposure time is point-sequential or line-sequential is input to the line average value calculation unit 100 and the delay unit 120 of the flicker correction apparatus.

  The line average value calculation unit 100 calculates an average value y0 for each line during the effective period of the input image signal.

  The calculated average value y0 is the one-field previous line average value storage unit 101, the n-field previous line average value storage unit 103, the three-field average value calculation unit 104, the difference value calculation unit 105, the 50 Hz flicker component extraction unit 106, 60 Hz. This is input to the flicker component extraction unit 107 and the motion detection unit 110.

  The one-field previous line average value storage unit 101 stores the input line average value y0 and outputs an average value y-1 delayed by one field period.

  The average value y−1 output from the one-field previous line average value storage unit 101 is input to the two-field previous line average value storage unit 102, the three field average value calculation unit 104, and the motion detection unit 110.

  The two-field previous line average value storage unit 102 stores the input line average value y−1 and outputs the average value y−2 delayed by one field to the three-field average value calculation unit 104.

  As a result, y0, y-1, and y-2 are input to the three-field average value calculation unit 104.

  The three-field average value calculation unit 104 calculates the three-field average value by calculating the average of these three line average values based on the input y0, y-1, and y-2, and influences flicker. Create the excluded reference value.

  The average value (three field average value) calculated by the three field average value calculation unit 104 is input to the 50 Hz flicker component extraction unit 106.

  The 50 Hz flicker component extraction unit 106 divides the line average value output from the line average value calculation unit 100 by the average value output from the three-field average value calculation unit 104 to thereby remove the flicker signal from which the density of the subject has been eliminated. Is calculated. Then, the frequency of the flicker signal is converted, and only the frequency region of the 50 Hz flicker component is extracted.

  The difference value calculation unit 105 subtracts the n-field previous line average value output from the n-field previous line average value storage unit 103 from the line average value output from the line average value calculation unit 100 to obtain a 60 Hz flicker component extraction unit 107. To enter.

  The 60 Hz flicker component extraction unit 107 divides the difference value output from the difference value calculation unit 105 by the average value output from the line average value calculation unit 100, thereby calculating a flicker signal from which the density of the subject is eliminated. . Then, the frequency of the flicker signal is converted, and only the frequency region of the 60 Hz flicker component is extracted.

  Hereinafter, the operations described so far will be described with reference to the drawings.

  FIG. 2 shows an image pickup apparatus driven by the NTSC system having a vertical scanning frequency (field frequency) of 59.94 Hz and a horizontal scanning frequency of 15.75 kHz, and is a general fluorescent lamp that is lit by an alternating current power source of fp = 50 Hz. It is a figure which shows the example of an image at the time of image | photographing a white chart under illumination.

  FIG. 3 is a plot of line average values of the image. As shown in this figure, the line average value is a signal in which the flicker component is modulated by the density of the subject.

  From the relationship between the power supply frequency and the drive frequency of the image sensor, the average value Ys (Y, n) of the y line in n frames is expressed by the following equation.

here,
Is (y): DC component of the y line (corresponding to the contrast of the subject)
A: Flicker size f _h : Horizontal scanning frequency f _v : Vertical scanning frequency f _p : Power supply frequency.

As a result, horizontal stripes appear at the f _h / (2 × f _p ) line period for the y line, and the phase shifts at (2π × 2 × f _p ) / f _v for the n frame.

That is, the flicker when f _p = 50 Hz is
f _h / (2 × f _p ) = 15.75 kHz / 100 = 157.5 A horizontal stripe appears at a line period, and (2π × 2 × f _p ) / f _v = 2π × 100 / 59.94 Hz≈2π per field × (1 + 2/3)
From the fact that the phase is shifted only by this, it can be seen that the circuit takes three rounds due to the nature of the trigonometric function.

Similarly, when f _p = 60 Hz, the flicker is
f _h / (2 × f _p ) = 15.75 kHz / 120 = 131.25 A horizontal stripe appears at a line period, and per field,
(2π × 2 × f _p ) / f _v = 2π × 120 / 59.94 Hz≈2π × (2 + 1/500)
From the fact that the phase is shifted only by this, it can be seen that it is a cycle of 500 fields due to the nature of the trigonometric function.

FIG. 4 is a plot of the line average value of the current field when f _p = 50 Hz, the line average value of the previous field, the line average value of the previous field, and the average value of these three fields. . As described above, since the line average value circulates in three fields, it is possible to extract the shade of the subject from which the influence of flicker is eliminated by calculating the average value for three fields.

  FIG. 5 shows a result of dividing the line average value of the current field by the average value of the three fields obtained previously, and a flicker signal from which the contrast of the subject is excluded is extracted.

FIG. 6 is a plot of the line average value of the current field when f _p = 60 Hz, the line average value n fields before, and the difference value of these two fields.

  FIG. 7 shows the result of dividing the previously obtained difference value of the two fields by the line average value of the current field, and a flicker signal excluding the shade of the subject is extracted.

  The extraction of the flicker signal excluding the contrast of the subject will be described using mathematical expressions.

The line average value in the n ₁ and n ₂ frames is expressed by the following equation.

  When the above equation is simplified, the following [1] and [2] are obtained, and ([2]-[1]) / [2] obtained from these [1] and [2] is as follows.

  Since flicker by a general fluorescent lamp is A << 1, the above ([2]-[1]) / [2] can be approximated by the following equation, and a flicker signal that is not affected by the density of the subject is obtained. Can be extracted.

  The frequency region of the flicker component extracted by the 50 Hz flicker component extraction unit 106 includes the cyclic low-pass filter result of the flicker component of the previous field by the first cyclic LPF 108 and the cyclic coefficient output from the cyclic coefficient calculation unit 114. Based on the low-pass filter. The frequency region of the flicker component after the cyclic low-pass filter is input to the flicker determination unit 115.

  Similarly, the frequency region of the flicker component extracted by the 60 Hz flicker component extraction unit 106 is output from the cyclic low-pass filter result of the flicker component n fields before and the cyclic coefficient calculation unit 114 by the second cyclic LPF 109. A low-pass filter is applied based on the cyclic coefficient. The frequency region of the flicker component after the cyclic low-pass filter is input to the flicker determination unit 115.

  The motion detection unit 110 compares the difference between the line average value output from the line average value calculation unit 100 and the average value output from the one-field previous line average value storage unit 101 with a certain threshold value to determine whether or not the subject has moved. Judging. The determination result of the presence / absence of motion is input to the motion field number measurement unit 111, the static field number measurement unit 112, the cyclic coefficient calculation unit 114, and the flicker determination unit 115.

  The motion field number measurement unit 111 measures the number of fields in which motion has continued (number of motion fields) based on the motion detection result output from the motion detection unit 110. The result is input to the fade gain calculation unit 117 (see FIG. 8A).

  The stationary field number measuring unit 112 measures the number of fields (still field number) that has been stationary based on the motion detection result output from the motion detecting unit 110. The result is input to the field number n calculation unit 113 and the fade gain calculation unit 117 (see FIG. 8B).

  The field number n calculation unit 113 calculates (determines) the field number n based on the result of the number of still fields output from the still field number measurement unit 112. When the number of still fields is not zero (at rest), a multiple of 3 is selected as the number of fields n in order to avoid the influence of 50 Hz flicker, while the number of fields n is reduced while the number of stationary fields is small, As the number of static fields increases, the field number n is gradually increased to a predetermined number.

  Specifically, when the number of static fields is small, the number of fields n is reduced to 3, 6, etc. By doing so, it is possible to extract the flicker signal in about four fields at the shortest. Also, as the number of static fields increases, the field number n is gradually increased to a predetermined number. By doing so, the value of the field difference signal is increased, and as a result, the flicker signal extraction accuracy can be increased (see FIG. 8C).

  On the other hand, when the number of static fields is zero (when moving), the number of fields n is set to zero.

  The field number n calculated in this way is input to the n-field previous line average value storage unit 103.

  The cyclic coefficient calculation unit 114 determines a cyclic coefficient based on the motion detection result output from the motion detection unit 110. Specifically, the cyclic coefficient calculation unit 114 temporarily sets the cyclic coefficient to zero for one field immediately after the moving state is changed to the stationary state, and sets the cyclic coefficient to a predetermined value when the stationary state continues. When the movement state is reached, the cyclic coefficient is set to 1. In other words, the stationary state is moderately cycled (for example, 0.75), and when moving, the cyclic coefficient is increased to hold the calculation result (for example, 0.95). When the moving state is changed to the stationary state, an operation is performed such that the cyclic coefficient is once set to zero and the calculation result is initialized (see FIG. 8D).

  The calculated cyclic coefficient is input to the first cyclic LPF unit 108 and the second cyclic LPF unit 109.

  The flicker determination unit 115 adjusts flicker components based on the presence or absence of motion and the presence or absence of flicker based on the detection result of the motion detection unit 110 and the outputs from the first cyclic LPF unit 108 and the second cyclic LPF unit 109. And frequency determination (50 Hz or 60 Hz). The flicker component adjusted based on the determination result is input to the flicker correction gain generation unit 116.

  Based on the determination result of the flicker determination unit 115, the flicker correction gain calculation unit 116 generates a sine wave synchronized with the flicker frequency by superimposing trigonometric functions, and outputs the sine wave to the fade gain multiplication unit 118.

  The fade gain calculation unit 117 calculates (controls) the fade gain based on the measurement results of the motion field number measurement unit 111 and the still field number measurement unit 112, that is, according to the number of still fields and the number of motion fields. Specifically, when the number of stationary fields is not zero, the fade gain is decreased while the number of stationary fields is small, while the fade gain is gradually increased and increased to 1 time as the number of stationary fields increases. To go. On the contrary, when the number of motion fields is not zero, the fade gain is increased while the number of motion fields is small, while the fade gain is gradually decreased and reduced to 0 times as the number of motion fields increases. That is, in the case of a stationary state, the fade gain is gradually increased according to the length of the stationary state and increased to 1. Conversely, in the case of motion, the fade gain remains at 1 for a predetermined period immediately after the transition from stationary to motion, and if the motion continues after a predetermined time, the fade gain is gradually reduced to zero. (See FIG. 8E). The calculated fade gain is input to the fade gain multiplier 118.

  The fade gain multiplication unit 118 multiplies the flicker correction gain output from the flicker correction gain calculation unit 116 by the fade gain output from the fade gain calculation unit 117.

  The flicker correction multiplication unit 119 multiplies the image signal subjected to delay adjustment by the delay unit 120 by a flicker correction gain. The flicker correction multiplication unit 119 outputs a signal with corrected flicker.

  As described above, according to the present invention, flicker correction can be performed even when the subject moves, and even if there is little phase shift between the field period and flicker period, the moving state changes to the stationary state. An excellent effect is obtained that the flicker correction can be started at a high speed.

  The present invention is useful in the field of camera systems where shooting locations such as movie cameras, in-vehicle cameras, and portable cameras are not fixed and flicker correction of different frequencies is required.

100 line average value calculation unit 101 1 field previous line average value storage unit 102 2 field previous line average value storage unit 103 n field previous line average value storage unit 104 3 field average value calculation unit 105 difference value calculation unit 106 50 Hz flicker component extraction Unit 107 60 Hz flicker component extraction unit 108 first cyclic LPF
109 Second cyclic LPF
110 Motion detection unit 111 Motion field number measurement unit 112 Still field number measurement unit 113 Field number n calculation unit 114 Cyclic coefficient calculation unit 115 Flicker determination unit 116 Flicker correction gain calculation unit 117 Fade gain calculation unit 118 Fade gain multiplication unit 119 Flicker correction Gain multiplication unit 120 Delay unit 121 Microcomputer

Claims (9)

A flicker correction apparatus that reduces flicker in an image captured using an image sensor whose exposure time is point-sequential or line-sequential,
A line average value calculation unit for calculating an average value for each line of the effective period of the input image signal;
A storage unit for storing the average value obtained by the line average value calculation unit in units of one field period;
A motion detection unit that detects motion from the average value obtained by the line average value calculation unit and the output change of the average value of the previous field stored in the storage unit;
A stationary field number measuring unit that measures the number of stationary fields, which is the number of fields that have been stationary based on the motion detection result output from the motion detecting unit;
A motion field number measurement unit that measures the number of motion fields, which is the number of fields that have continued to move based on the motion detection result output from the motion detection unit;
A difference value calculation unit that calculates a difference value between the average value obtained by the line average value calculation unit and the previous average value by n fields (n is an integer value) stored in the storage unit;
A field number n calculating unit for calculating the field number n based on the result of the number of static fields output from the stationary field number measuring unit;
A flicker component extraction unit that extracts a flicker frequency from the average value obtained by the line average value calculation unit and the difference value obtained by the difference value calculation unit;
A flicker determination unit that determines the presence or absence of flicker and a flicker frequency from the flicker component extracted by the flicker component extraction unit and the result of the motion detection unit;
From the result of the flicker determination unit, a flicker correction gain calculation unit that calculates a flicker correction gain synchronized with the flicker frequency;
A flicker correction gain multiplication unit that multiplies the input image signal by the correction gain from the flicker correction gain calculation unit,
The field number n calculation unit
The number of fields n is controlled according to the number of stationary fields and the number of motion fields,
When the number of stationary fields is not zero, the number of fields n is reduced while the number of stationary fields is small, while the number of fields n is gradually increased to a predetermined number as the number of stationary fields increases. ,
When the motion field number is not zero, the field number n is set to zero.
A flicker correction apparatus characterized by that. A flicker correction apparatus that reduces flicker in an image captured using an image sensor whose exposure time is point-sequential or line-sequential,
A line average value calculation unit for calculating an average value for each line of the effective period of the input image signal;
A storage unit for storing the average value obtained by the line average value calculation unit in units of one field period;
A motion detection unit that detects motion from the average value obtained by the line average value calculation unit and the output change of the average value of the previous field stored in the storage unit;
A stationary field number measuring unit that measures the number of stationary fields, which is the number of fields that have been stationary based on the motion detection result output from the motion detecting unit;
A motion field number measurement unit that measures the number of motion fields, which is the number of fields that have continued to move based on the motion detection result output from the motion detection unit;
A difference value calculation unit that calculates a difference value between the average value obtained by the line average value calculation unit and the previous average value by n fields (n is an integer value) stored in the storage unit;
A flicker component extraction unit that extracts a flicker frequency from the average value obtained by the line average value calculation unit and the difference value obtained by the difference value calculation unit;
A flicker determination unit that determines the presence or absence of flicker and a flicker frequency from the flicker component extracted by the flicker component extraction unit and the result of the motion detection unit;
From the result of the flicker determination unit, a flicker correction gain calculation unit that calculates a flicker correction gain synchronized with the flicker frequency;
A fade gain calculating unit that calculates a fade gain from the measurement results of the stationary field number measuring unit and the motion field number measuring unit;
A fade gain multiplier for multiplying the flicker correction gain by the fade gain calculated by the fade gain calculator;
A flicker correction gain multiplier that multiplies the input image signal by the correction gain input from the fade gain multiplier;
The fade gain calculation unit controls the fade gain according to the number of still fields and the number of motion fields,
When the number of stationary fields is not zero, the fade gain is decreased while the number of stationary fields is small, while the fade gain is gradually increased and increased to 1 time as the number of stationary fields increases.
When the number of motion fields is not zero, the fade gain is increased while the number of motion fields is small, while the fade gain is gradually decreased and decreased to 0 times as the number of motion fields increases.
A flicker correction apparatus characterized by that. A flicker correction apparatus that reduces flicker in an image captured using an image sensor whose exposure time is point-sequential or line-sequential,
A line average value calculation unit for calculating an average value for each line of the effective period of the input image signal;
A storage unit for storing the average value obtained by the line average value calculation unit in units of one field period;
A motion detection unit that detects motion from the average value obtained by the line average value calculation unit and the output change of the average value of the previous field stored in the storage unit;
A difference value calculation unit that calculates a difference value between the average value obtained by the line average value calculation unit and the previous average value by n fields (n is an integer value) stored in the storage unit;
A flicker component extraction unit that extracts a flicker frequency from the average value obtained by the line average value calculation unit and the difference value obtained by the difference value calculation unit;
A cyclic coefficient calculation unit that calculates a cyclic coefficient based on a motion detection result output from the motion detection unit;
A cyclic low-pass filter unit that applies a low-pass filter to the flicker component extracted by the flicker component extraction unit based on the cyclic coefficient calculated by the cyclic coefficient calculation unit;
A flicker determination unit that determines presence / absence of flicker and a flicker frequency based on an output from the cyclic low-pass filter unit and a detection result of the motion detection unit;
A flicker correction gain calculation unit that calculates a flicker correction gain synchronized with a flicker frequency based on a determination result of the flicker determination unit;
A flicker correction gain multiplication unit that multiplies the input image signal by the correction gain calculated by the flicker correction gain calculation unit,
The cyclic coefficient calculation unit temporarily sets the cyclic coefficient to zero for one field immediately after the moving state is changed to the stationary state, and sets the cyclic coefficient to a predetermined value when the stationary state continues, while the moving coefficient enters the moving state. When the cyclic coefficient is set to 1,
A flicker correction apparatus characterized by that. A flicker correction method for reducing flicker of an image captured using an image sensor whose exposure time is point-sequential or line-sequential,
A line average value calculating step for calculating an average value for each line of the effective period of the input image signal;
A storage step of storing the average value obtained in the line average value calculating step in units of one field period;
A motion detection step of detecting motion from the average value obtained in the line average value calculation step and the output change of the average value of the previous field stored in the storage step;
A stationary field number measuring step of measuring the number of stationary fields, which is the number of fields that have been stationary based on the motion detection result in the motion detecting step;
A motion field number measuring step for measuring the number of motion fields, which is the number of fields in which motion has continued based on the motion detection result in the motion detection step;
A difference value calculating step of calculating a difference value between the average value obtained in the line average value calculating step and the previous average value by n fields (n is an integer value) stored in the storing step;
A field number n calculating step of calculating the field number n based on the result of the number of static fields measured in the stationary field number measuring step;
A flicker component extracting step of extracting a flicker frequency from the average value obtained in the line average value calculating step and the difference value obtained in the difference value calculating step;
A flicker determination step for determining presence / absence of flicker and a flicker frequency based on the flicker component extracted in the flicker component extraction step and the motion detection result in the motion detection step;
A flicker correction gain calculating step for calculating a flicker correction gain synchronized with the flicker frequency from the determination result in the flicker determination step;
A flicker correction gain multiplication step of multiplying the input image signal by the correction gain calculated in the flicker correction gain calculation step,
In the step of calculating the number of fields n,
The number of fields n is controlled according to the number of stationary fields and the number of motion fields,
When the number of stationary fields is not zero, the number of fields n is reduced while the number of stationary fields is small, while the number of fields n is gradually increased to a predetermined number as the number of stationary fields increases. ,
When the motion field number is not zero, the field number n is set to zero.
A flicker correction method characterized by that. A flicker correction method for reducing flicker of an image captured using an image sensor whose exposure time is point-sequential or line-sequential,
A line average value calculating step for calculating an average value for each line of the effective period of the input image signal;
A storage step of storing the average value obtained in the line average value calculating step in units of one field period;
A motion detection step of detecting motion from the average value obtained in the line average value calculation step and the output change of the average value of the previous field stored in the storage step;
A stationary field number measuring step of measuring the number of stationary fields, which is the number of fields that have been stationary based on the motion detection result in the motion detecting step;
A motion field number measuring step for measuring the number of motion fields, which is the number of fields in which motion has continued based on the motion detection result in the motion detection step;
A difference value calculating step of calculating a difference value between the average value obtained in the line average value calculating step and the previous average value by n fields (n is an integer value) stored in the storing step;
A flicker component extracting step of extracting a flicker frequency from the average value obtained in the line average value calculating step and the difference value obtained in the difference value calculating step;
A flicker determination step for determining presence / absence of flicker and a flicker frequency based on the flicker component extracted in the flicker component extraction step and the motion detection result in the motion detection step;
A flicker correction gain calculating step for calculating a flicker correction gain synchronized with the flicker frequency based on the determination result in the flicker determination step;
A fade gain calculating step of calculating a fade gain from the measurement results of the stationary field number measuring step and the motion field number measuring step;
A fade gain multiplication step of multiplying the flicker correction gain by the fade gain calculated in the fade gain calculation step;
A flicker correction gain multiplication step for multiplying the input image signal by the correction gain from the fade gain multiplication step,
In the fade gain calculation step, the fade gain is controlled according to the number of still fields and the number of motion fields,
When the number of stationary fields is not zero, the fade gain is decreased while the number of stationary fields is small, while the fade gain is gradually increased and increased to 1 time as the number of stationary fields increases.
When the number of motion fields is not zero, the fade gain is increased while the number of motion fields is small, while the fade gain is gradually decreased and decreased to 0 times as the number of motion fields increases.
A flicker correcting method characterized by the above. A flicker correction method for reducing flicker of an image captured using an image sensor whose exposure time is point-sequential or line-sequential,
A line average value calculating step for calculating an average value for each line of the effective period of the input image signal;
A storage step of storing the average value obtained in the line average value calculating step in units of one field period;
A motion detection step of detecting motion from the average value obtained in the line average value calculation step and the output change of the average value of the previous field stored in the storage step;
A difference value calculating step of calculating a difference value between the average value obtained in the line average value calculating step and the previous average value by n fields (n is an integer value) stored in the storing step;
A flicker component extracting step of extracting a flicker frequency from the average value obtained in the line average value calculating step and the difference value obtained in the difference value calculating step;
A cyclic coefficient calculation step for calculating a cyclic coefficient based on the motion detection result in the motion detection step;
A cyclic low-pass filter step for applying a low-pass filter to the flicker component extracted in the flicker component extraction step based on the cyclic coefficient calculated in the cyclic coefficient calculation step;
A flicker determination step for determining presence / absence of flicker and a flicker frequency based on an output from the cyclic low-pass filter step and a detection result in the motion detection step;
A flicker correction gain calculating step for calculating a flicker correction gain synchronized with the flicker frequency based on the determination result in the flicker determination step;
A flicker correction gain multiplication step of multiplying the input image signal by the correction gain calculated in the flicker correction gain calculation step,
In the cyclic coefficient calculation step, the cyclic coefficient is temporarily set to zero for one field immediately after the moving state is changed to the stationary state, and when the stationary state continues, the cyclic coefficient is set to a predetermined value, and the moving state is set. When the cyclic coefficient is set to 1,
A flicker correction method characterized by that. A flicker correction program for reducing flicker of an image captured using an image sensor whose exposure time is point-sequential or line-sequential, and causing a computer to execute the following steps.
A line average value calculating step for calculating an average value for each line during the effective period of the input image signal:
A storage step of storing the average value obtained in the line average value calculation step in units of one field period:
A motion detection step of detecting motion from the output change of the average value obtained in the line average value calculation step and the average value of the previous field stored in the storage step:
A stationary field number measuring step of measuring the number of stationary fields, which is the number of fields that have been stationary based on the motion detection result in the motion detecting step:
A motion field number measurement step of measuring the number of motion fields, which is the number of fields in which motion has continued, based on the motion detection result in the motion detection step:
A difference value calculating step of calculating a difference value between the average value obtained in the line average value calculating step and the average value previous by n fields (n is an integer value) stored in the storing step:
Calculating the number of fields n based on the result of the number of stationary fields measured in the stationary field number measuring step, and controlling the number of fields n according to the number of stationary fields and the number of motion fields. When the number of stationary fields is not zero, the number of fields n is reduced while the number of stationary fields is small, while the number of fields n is gradually increased to a predetermined number as the number of stationary fields increases. The field number n calculating step of setting the field number n to zero when the motion field number is not zero.
Flicker component extraction step for extracting a flicker frequency from the average value obtained in the line average value calculation step and the difference value obtained in the difference value calculation step:
Flicker determination step for determining presence / absence of flicker and flicker frequency based on the flicker component extracted in the flicker component extraction step and the motion detection result in the motion detection step:
Flicker correction gain calculation step for calculating a flicker correction gain synchronized with the flicker frequency from the determination result in the flicker determination step:
A flicker correction gain multiplication step for multiplying the input image signal by the correction gain calculated in the flicker correction gain calculation step: A flicker correction program for reducing flicker of an image captured using an image sensor whose exposure time is point-sequential or line-sequential, and causing a computer to execute the following steps.
A line average value calculating step for calculating an average value for each line during the effective period of the input image signal:
A storage step of storing the average value obtained in the line average value calculation step in units of one field period:
A motion detection step of detecting motion from the output change of the average value obtained in the line average value calculation step and the average value of the previous field stored in the storage step:
A stationary field number measuring step of measuring the number of stationary fields, which is the number of fields that have been stationary based on the motion detection result in the motion detecting step:
A motion field number measurement step of measuring the number of motion fields, which is the number of fields in which motion has continued, based on the motion detection result in the motion detection step:
A difference value calculating step of calculating a difference value between the average value obtained in the line average value calculating step and the average value previous by n fields (n is an integer value) stored in the storing step:
Flicker component extraction step for extracting a flicker frequency from the average value obtained in the line average value calculation step and the difference value obtained in the difference value calculation step:
Flicker determination step for determining presence / absence of flicker and flicker frequency based on the flicker component extracted in the flicker component extraction step and the motion detection result in the motion detection step:
A flicker correction gain calculating step for calculating a flicker correction gain synchronized with the flicker frequency based on the determination result in the flicker determination step:
A step of calculating a fade gain from the measurement results of the stationary field number measuring step and the motion field number measuring step, wherein the fade gain is controlled in accordance with the number of stationary fields and the number of motion fields, and the number of stationary fields When the stationary field is not zero, the fade gain is decreased while the number of stationary fields is small, while the fade gain is gradually increased and increased to 1 time as the number of stationary fields increases. When the motion is not zero, the fade gain is increased while the number of motion fields is small, while the fade gain is gradually decreased and reduced to 0 times as the number of motion fields increases.
A fade gain multiplication step of multiplying the flicker correction gain by the fade gain calculated in the fade gain calculation step:
Flicker correction gain multiplication step of multiplying the input image signal by the correction gain from the fade gain multiplication step: A flicker correction program for reducing flicker of an image captured using an image sensor whose exposure time is point-sequential or line-sequential, and causing a computer to execute the following steps.
A line average value calculating step for calculating an average value for each line during the effective period of the input image signal:
A storage step of storing the average value obtained in the line average value calculation step in units of one field period:
A motion detection step of detecting motion from the output change of the average value obtained in the line average value calculation step and the average value of the previous field stored in the storage step:
A difference value calculating step of calculating a difference value between the average value obtained in the line average value calculating step and the average value previous by n fields (n is an integer value) stored in the storing step:
Flicker component extraction step for extracting a flicker frequency from the average value obtained in the line average value calculation step and the difference value obtained in the difference value calculation step:
In the step of calculating a cyclic coefficient based on the motion detection result in the motion detection step, the cyclic coefficient is temporarily set to zero for one field immediately after the moving state is changed to the stationary state, and when the stationary state continues. While the cyclic coefficient is set to a predetermined value, the cyclic coefficient is set to 1 when a motion state is reached:
A cyclic low-pass filter step of applying a low-pass filter to the flicker component extracted in the flicker component extraction step based on the cyclic coefficient calculated in the cyclic coefficient calculation step:
Flicker determination step of determining the presence / absence of flicker and the flicker frequency based on the output from the cyclic low-pass filter step and the detection result in the motion detection step:
A flicker correction gain calculating step for calculating a flicker correction gain synchronized with the flicker frequency based on the determination result in the flicker determination step:
A flicker correction gain multiplication step for multiplying the input image signal by the correction gain calculated in the flicker correction gain calculation step: