JP2004072262A - Gamma correction method and apparatus for imaging means - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a gamma correction method and apparatus for an imaging means capable of automatically creating a lookup table with high accuracy for gamma correction for a CCD camera or the like in units of pixels without using an auxiliary tool. <P>SOLUTION: A flowchart indicating creating procedures of the lookup table for gamma correction includes: steps 1 to 7 of individually selecting all pixels P of a CCD camera in a prescribed sequence and deciding a base exposure time at which a unit output is obtained from the pixels selected sequentially when imaging under a prescribed illumination; steps 8 to 16 of setting an integer multiple of the decided base exposure time S[P] from the selected pixel to a real exposure time S and measuring an output value of the pixel through imaging under the illumination by each set real exposure time; a step 17 of creating the lookup table wherein each measured output value by each pixel is recorded in cross-reference with each input value comprising an integer used for calculation of the corresponding real exposure time; and applying the gamma correction to the CCD camera in units of pixels on the basis of the created lookup table. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention automatically creates a look-up table for gamma correction for each pixel for a gamma correction method and apparatus for an imaging unit, particularly for an imaging unit such as a CCD camera, and corrects the imaging unit based on the table. The present invention relates to a gamma correction method and apparatus for an imaging unit, which is preferably applied when performing a gamma correction.
[0002]
[Prior art]
In imaging means such as a CCD (Charge Coupled Device) camera, an output (voltage, current) having a magnitude corresponding to the intensity of incident light is generated, and image data (luminance value of each pixel) is created from the output. It has become. However, an output proportional to the intensity of the incident light is not always obtained. In particular, in the field of image measurement, when a CCD camera or the like is used to perform measurement or inspection based on obtained image data, it is required that the output is proportional to the input, that is, the γ value is 1. Is done. Therefore, in such a case, when the γ value is not 1 in the CCD camera or the like to be used, it is necessary to perform the correction (gamma correction). Further, depending on the target, correction to an arbitrary value other than 1 may be required. This gamma correction is one of the correction processes used in image processing, and is called in this way because the curve represented by the function (quadratic function) used at the time of correction is similar to the Greek character gamma. Have been.
[0003]
Conventionally, such gamma correction includes a method of performing correction using a look-up table including a conversion table for converting a luminance value of an input image into a luminance value of an output image. The following two methods are known as methods for creating a look-up table used for such gamma correction.
[0004]
(1) Using a test chart A calibration plate (test chart) composed of a pattern in which the density changes stepwise is imaged, and a lookup table is created from image data (brightness values) obtained at that time. I do.
[0005]
(2) Method of using the aperture of the camera lens The smaller the numerical value of the aperture, the higher the intensity of light incident on the camera. A uniform illumination light is imaged while changing the aperture by one scale, and a look-up table is created from image data of each aperture.
[0006]
[Problems to be solved by the invention]
However, the conventional methods of creating the look-up tables (1) and (2) have many problems such as (A) to (D) described below.
[0007]
(A) Correction cannot be performed in units of one pixel.
[0008]
(1) In the case of using a test chart In a CCD camera, since there is a difference in sensitivity for each pixel, a method of creating a look-up table for each pixel and performing gamma correction provides the highest accuracy. However, conventionally, i) a method of sharing one look-up table for the entire screen, or ii) a method of dividing each pixel into several regions and creating a look-up table for each region is used. And no gamma correction is performed for each pixel.
[0009]
(2) When a lens aperture is used Although it is possible to create a look-up table in pixel units, it is not generally performed.
[0010]
(B) Low accuracy due to input light.
[0011]
(1) When a test chart is used It is difficult to create a test chart in which a change (step) in density is accurate. Even if the change in the density of the test chart is accurately made, if the entire chart is not uniformly illuminated, the light according to the density of the chart is not input. Illuminating the entire chart uniformly is difficult in practice, and generally results in non-uniform illumination light. Therefore, the obtained data includes an error.
[0012]
(2) When a lens aperture is used Although the intensity of light is theoretically doubled at one scale of the aperture, it does not actually double exactly, so the obtained data contains errors. become.
[0013]
(C) In the lookup table to be created, since the interval of data that can be actually measured is wide, it is necessary to use an estimated value by interpolation for some data in the table, so that the accuracy is low. .
[0014]
(1) When a test chart is used If the change (step) in density is small, values in the middle of the lookup table must be interpolated as described above. In the case of an 8-bit data CCD camera, it is possible to reduce the data to be interpolated by reducing the change in density. However, in a camera having a large number of data bits such as 10 bits or more, a large amount of data is interpolated. Become.
[0015]
(2) When a lens aperture is used Since the intensity of light is doubled at one scale of the aperture, a lookup table cannot be created without interpolation. In particular, since the data interval on the bright side is wider than that on the dark side, the accuracy of interpolation decreases.
[0016]
(D) When a camera is installed at a factory automation (FA) production site or the like for image measurement for machine control or the like, a gamma correction lookup table is automatically installed with the camera attached. Updating is difficult to automate.
[0017]
(1) When a test chart is used A mechanism for automatically retreating and returning the test chart instead of the object is required separately. A mechanism or operation is required for managing and storing unused test charts in a state where their functions are not impaired due to contamination or the like. In practice, it is often difficult to provide such a mechanism due to limitations in installation space and the like.
[0018]
(2) When a lens aperture is used If a lens capable of automatically changing the aperture is used, automation can be performed. However, since the aperture is changed by a mechanical mechanism, it is difficult to perform gamma correction periodically (frequently) due to durability problems.
[0019]
SUMMARY OF THE INVENTION The present invention has been made to solve the above-described conventional problem, and automatically creates a high-precision lookup table for gamma correction in units of pixels without using an auxiliary jig. An object of the present invention is to provide a gamma correction method and apparatus for an imaging unit that can accurately perform gamma correction on an imaging unit such as a CCD camera as a result.
[0020]
[Means for Solving the Problems]
According to the present invention, all pixels constituting an image pickup unit included in an image pickup unit are individually selected in a predetermined order, and a base exposure time at which a unit output is obtained when an image is picked up under predetermined illumination for sequentially selected pixels. Determined, for the selected pixel, set an integer multiple of the determined base exposure time as the actual exposure time, image under the illumination for each set actual exposure time, and output each output value of the same pixel Measured, for each pixel, created a lookup table that recorded each measured output value, corresponding to each input value consisting of the integer value used to calculate the corresponding actual exposure time, created This object is achieved by performing gamma correction of the image pickup means on a pixel basis based on a look-up table.
[0021]
According to the present invention, there is further provided a selecting means for individually selecting all the pixels constituting the image pickup section included in the image pickup means in a predetermined order, and obtaining a unit output when the sequentially selected pixels are imaged under predetermined illumination. Determining means for determining a base exposure time to be determined, setting means for setting an integral multiple of the determined base exposure time for a selected pixel to an actual exposure time, and setting the actual exposure time under the illumination for each set actual exposure time. A measuring means for imaging each pixel and measuring each output value of the same pixel; and for each pixel, each measured output value corresponds to each input value consisting of an integer value used for calculating a corresponding actual exposure time. The above-described problem is similarly provided by providing a creating unit that creates a look-up table recorded with attached data, and a correcting unit that performs gamma correction of the imaging unit on a pixel-by-pixel basis based on the created lookup table. Settled Than it is.
[0022]
That is, in the present invention, the luminance value of the input image is set in integral multiples (actual exposure time) of the base exposure time at which a unit output is obtained, in units of sequentially selected pixels. The look-up table can be created automatically, and gamma correction can be accurately performed on a pixel-by-pixel basis.
[0023]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
[0024]
FIG. 1 is a block diagram showing an outline of a CCD line sensor camera 10 to which a gamma correction device of an imaging unit according to an embodiment of the present invention is applied.
[0025]
The CCD line sensor camera (imaging means) 10 includes an image sensor (imaging unit) 12 in which a large number of pixels are configured by CCD elements, and an analog / digital conversion unit 14 which converts an electric signal output from the sensor 12 during imaging. A measurement value recording RAM 15 for recording a signal after the A / D conversion value input through the CPU, a gamma correction unit 16 for gamma correction of an image signal read from the recording RAM 15, and a correction unit 16. The table recording RAM 18 for storing a lookup table to be referred to when performing the correction processing, a CPU (Central Processing Unit) 20 for executing various arithmetic processing, and the image sensor based on the processing result by the CPU 20. And a timing generation circuit 22 that outputs a signal for controlling the imaging timing to the imaging device 12.
[0026]
In the camera 10, the image signal corrected by the gamma correction unit 16 is output from the output unit as a video output (differential digital signal) via the RS422 standard serial interface 24. I have. The CPU 20 is connected to a ROM 26 and a RAM 28, respectively, and is connected to an external terminal (I / F) via an RS232C standard serial interface 30 so that various data can be input / output. .
[0027]
The correction device according to the present embodiment includes a selection unit that individually selects all the pixels constituting the image sensor 12 in a predetermined order, and a unit output when the sequentially selected pixels are imaged under predetermined illumination (not shown). Determining means for determining the base exposure time that is obtained, setting means for setting an integral multiple of the determined base exposure time to the actual exposure time for the selected pixel, and Measuring means for imaging under illumination to measure the output value of the same pixel, and for each pixel, convert each measured output value to each input value consisting of the integer value used to calculate the corresponding actual exposure time Creating means for creating a look-up table recorded in association with the gamma correction unit (correcting means) 16 for performing gamma correction of the imaging means on a pixel basis based on the created lookup table; Eteiru.
[0028]
In the present embodiment, the respective units for executing the arithmetic processing including the image processing except for the correction unit 16 are realized by software or the like stored in the ROM 26 and the RAM 28 in the CPU 20. In particular, the measuring means is realized by the CPU 20 which reads a signal from the measured value recording RAM 15 in pixel units and recognizes the luminance value. The gamma correction unit 16 determines and outputs a corrected luminance value to be output from the look-up table recorded in the RAM 18 and the value of the image signal recorded in the RAM 15. I have. However, the correction means 16 may be realized by software in the CPU 20 in the same manner.
[0029]
Next, the operation of the present embodiment will be described with reference to the flowchart of FIG. 2 using a CCD line sensor camera in which the luminance data (gradation value) of each pixel is 8 bits and the number of effective pixels is 1024 as an example.
[0030]
The meaning of each symbol used in this flowchart is as follows.
[0031]
P: CCD pixel serial number (1 to 1024)
H: Lookup table input value (1-255)
S: Actual exposure time (time) at the time of imaging
・ In the case of a line sensor, the start pulse interval is used. ・ In the case of an area sensor, the shutter speed is used. S [P]: Base exposure time (time)
Smin, which means the actual exposure time when the output value of the pixel P becomes 1, the shortest lower limit exposure time (time) that can be set by the camera used
ΔS: minimum increase in exposure time (time) when searching for base exposure time (S [P])
V [P] [S]: luminance value of pixel P when imaged at actual exposure time S (0-255)
L [P] [H]: Output value for input value H of the lookup table (0-255)
[0032]
First, the correction device of this embodiment is initialized (step 1). In this initialization, the brightness of the illumination is adjusted to a level that satisfies the luminance value of the pixel P: V [P] [Smin] ≦ 1 when the image is captured with the shortest lower limit exposure time Smin that can be set in the camera to be used. This illumination level is determined experimentally in advance. If this level is known in advance, the operation here need not be performed. In the initialization in step 1, the value of the look-up table recording RAM 18 is cleared to zero. At this time, the value in the 0th column (for outputting 0 as a value after correction when the value before correction is 0) is also cleared to zero.
[0033]
After the setting of the illumination intensity used in step 1 is completed, the serial number P of the pixel is set to 1 (step 2). Here, as shown in the next step 3 and a step 17 described later, a gamma correction lookup table is created by sequentially selecting the first pixel to the 1024th pixel. . However, the order of selection may be arbitrary as long as it can be finally created for all pixels.
[0034]
Next, an initial actual exposure time S is set for the first pixel to obtain a base exposure time. As the actual exposure time S, the above-described lower limit exposure time Smin is set (step 4). However, if the exposure time when the output value (luminance value) is 1 (unit output) as the gradation value can be assumed in advance, a value slightly smaller than that value may be set. By doing so, the processing time can be reduced.
[0035]
Next, an image to be picked up at the lower limit of the actual exposure time S set in step 4 is input (step 5), and it is determined whether or not the luminance value (output) of the input P (= 1) th pixel is 1 or not. It is determined (step 6). If the luminance value of the P-th pixel has not reached 1, the actual exposure time S is increased by the smallest increment ΔS, which is the smallest changeable value (step 7), and again. The image input of step 5 is performed.
[0036]
In step 6, the actual exposure time S when the luminance value of the P-th pixel becomes 1 is recorded (set) as the base exposure time S [P] of the pixel P, and the input value H = 1 The unit output value 1 is set to the output value L [P] [1] of the gamma correction lookup table for the P-th pixel at this time (steps 8 and 9). However, since the value of the initial output value L [P] [1] is determined to be 1 in all the pixels, the initial output value L [P] [1] may be set collectively in the initialization in step 1 or may be particularly set as data. It is not necessary.
[0037]
After the process of setting the unit output value for the initial input value H = 1 is completed, the next input value 2 is set as the input value H for which the lookup table is to be created (step 10), and the lookup is performed in the next step 11 It is determined whether the measurement has been performed for all of the input values in the table. In this example, since the camera is an 8-bit camera, the maximum input value is 255 (= 2 ⁸ -1).
[0038]
At this stage, since the input value P is of course equal to or less than 255, the process proceeds in the Yes direction in step 11, and the actual exposure time S corresponding to the input value H = 2 at that time is H times the base exposure time S [P]. Is set (step 12), an image is actually taken at the actual exposure time S and an image is input (step 13), and the luminance value V [P] [S] at that time is measured, and the measured luminance value (Gradation value) is set to the output value L [P] [V [P] [S]] of the gamma correction look-up table of the Pth pixel, and the output value L [ [P] [V [P] [S]] is stored in the memory (RAM) (step 14). Thereafter, when the measured luminance value V [P] [S] at the input value H has not reached 255, the input value H is incremented by 1 (steps 15 and 16), and a look-up table for gamma correction is created. The target input value is changed to the next input value whose output has not yet been measured, and the above steps 11 to 16 are repeated until the last input value H = 255.
[0039]
When the measurement of the output luminance values for all the input values H of the first P = 1 pixel is completed (No in Step 11), or when the measured luminance value V [P] [S] reaches 255 (Step 15). No), the pixel number P is incremented by 1, and the pixel for which the look-up table for gamma correction is created is changed to the next pixel that has not yet been measured (step 17). In addition to performing each of the sixteen processes, each process is repeated until there are no more target pixels in step 3. After the above processes are completed for all pixels (No in step 3), a table interpolation process is performed (step 18), and the creation of the lookup table is completed.
[0040]
In the table interpolation process performed in the last step 18, when the corrected luminance value L is recorded in accordance with the step 14, when the value is not fixed in the lookup table (the initial value is 0), Therefore, the entire table is scanned L, and an undetermined value is interpolated and determined.
[0041]
This will be described as a specific example of the table in the nth row (n pixels, P pixels in the flowchart) in the lookup table recording RAM 18. The state in which the unit output value = 1 is set with respect to the initial input value H = 1 in step 9 can be represented as shown in FIG. Next, assuming that the output value 2 is obtained for the input value H = 2 in the step 14, the result is as shown in FIG. 11B. Similarly, in the step 14, the next input value H = 3, H Assuming that the output values are 4, 7 for = 4, an undetermined portion (column) where the value remains at the initial value 0 occurs as shown in FIGS. become.
[0042]
In the case where such a column whose value is undetermined occurs, the interpolation in the case where the m-th column is determined to be W and the (m + k) -th column is determined to be Z (Z−W = 1) in FIG. As shown in the image of the method, interpolation is performed using the final value that is closest to the table. If this interpolation method is applied to FIG. 3D, the result is as shown in FIG.
[0043]
If H = 256, and the values of the table up to 255 columns have not been determined in the state where the determination processing of step 11 has been exited, as shown in FIG. Processing (termination processing) is performed.
[0044]
In the present embodiment, FIG. 7 shows an image of data stored in the measured value recording RAM 15 and the look-up table recording RAM 18 corresponding to each pixel of the image sensor 12 of the CCD line sensor camera. With the processing described in detail above, it is possible to create a look-up table in which output values based on actually measured values are recorded.
[0045]
The lookup table recording RAM used here uses a CCD line sensor camera of 1024 pixels and 8 bits (0 to 255 gradations) as a camera, and therefore 262144 [bytes] (1024 [pixels]). X 256 [gray scale (input)] x 8 [bits (output)] = 1024 [rows] x 256 [columns] x 1 [bytes].
[0046]
In this lookup table, the gamma correction is performed in pixel units based on the actual measurement data created according to the above-described flowchart, so that the row number corresponds to the serial number P of the pixel of the CCD and the column number corresponds to the input value H of the lookup table. The corrected output value L based on the actually measured value is recorded in the RAM at the address corresponding to the cell defined by each row and each column. That is, the correction in the case where “the value L is recorded in the P-th row and the H-th column” in the recording RAM is “correction of the P-th pixel when the luminance value input to the P-th pixel is the H gradation. Let the subsequent output value be L gradation. "
[0047]
FIG. 8 shows the relationship between the input value and the output value of one pixel before and after correction. The output value before correction is a luminance value actually measured according to the flowchart. (A) and (B) show images when the luminance value actually measured at an arbitrary γ value is corrected to γ = 1, and FIGS. (C) and (D) show images in the reverse case. Is shown.
[0048]
Generally, image data input / output devices such as scanners and color printers have their own gamma values. In order to faithfully reproduce an image, it is necessary to make the overall gamma from the image input to the final output equal to 1. When gamma correction is performed when processing image data between input and output devices, the gamma value is the product of the values of each gamma conversion. For example, if the gamma of the scanner is 0.56 and the gamma of the display is 1.8, the overall gamma is 0.56 × 1.8, which is almost 1.
[0049]
According to the present embodiment, as shown in FIGS. 4C and 4D, since the gamma value can be corrected to an arbitrary value other than 1, such gamma correction between input / output devices can be performed.
[0050]
Here, a specific operation at the time of correction will be briefly described. In the measured value recording RAM 15, a signal from the sensor (A / D converted) is recorded as a luminance value before gamma correction. The look-up table also includes a corrected luminance value L [P] [V [P] [S] corresponding to the luminance value V [P] [S] of the pixel P recorded in the measured value recording RAM 15. ] Is recorded.
[0051]
In a specific example, in order to obtain the corrected luminance value L ′ that satisfies γ = 1, first, the corrected luminance value L that satisfies γ = 1 is obtained, and then the following gamma correction equation L ′ = 255 × (L / 255) ^{1 /} γ
L ′ is calculated from L according to This L ′ is obtained for all values of the lookup table. This conversion to an arbitrary γ value is omitted in the flowchart, but is performed after the CPU 20 completes the “table interpolation process” in step 18.
[0052]
According to this embodiment described in detail above, a gamma correction look-up table for a CCD camera or the like can be created more accurately than in the conventional method, and, for each pixel, correction for gray scale or the like can be performed. Since it can be created automatically (periodically) without using an auxiliary jig, highly accurate gamma correction can be performed in pixel units by using the look-up table. .
[0053]
As described above, the present invention has been specifically described. However, the present invention is not limited to the above-described embodiment, and can be variously modified without departing from the gist thereof.
[0054]
For example, the specific configuration of the correction device according to the present invention is not limited to the configuration shown in the embodiment.
[0055]
In the above-described embodiment, the case where the imaging unit is a CCD camera has been described. However, the present invention is not limited to this, and a CMOS (Complementary Metal Oxide Semiconductor) or the like may be used.
[0056]
Further, the unit output for determining the base exposure time may be a predetermined current value or voltage value instead of the gradation value.
[0057]
Further, the setting of the input value H does not necessarily have to be performed for all gradation values, and may be, for example, every other or every other gradation value.
[0058]
【The invention's effect】
As described above, according to the present invention, a high-precision lookup table used for gamma correction of an imaging unit such as a CCD camera is automatically created in units of pixels without using an auxiliary jig. As a result, the gamma correction of the imaging means can be performed accurately.
[Brief description of the drawings]
FIG. 1 is a block diagram showing an outline of a correction apparatus according to an embodiment of the present invention; FIG. 2 is a flowchart showing a procedure for creating a look-up table for gamma correction; FIG. FIG. 4 is an explanatory diagram showing a method of interpolating an undetermined value of a lookup table. FIG. 5 is an explanatory diagram showing a result of interpolating an undetermined value of a lookup table. FIG. 7 is an explanatory diagram showing a method of interpolating a terminal undetermined value. FIG. 7 is an explanatory diagram showing an entire image of a lookup table to be created. FIG. 8 is a line for explaining the relationship between input and output before and after gamma correction. Figure [Explanation of symbols]
10 CCD line sensor camera

Claims (4)

All the pixels constituting the imaging unit of the imaging means are individually selected in a predetermined order, and for the sequentially selected pixels, a base exposure time at which a unit output is obtained when imaging is performed under predetermined illumination is determined.
For the selected pixel, set an integral multiple of the determined base exposure time to the actual exposure time, measure each output value of the same pixel by imaging under the illumination for each set actual exposure time,
For each pixel, create a lookup table in which each measured output value is recorded in association with each input value consisting of an integer value used to calculate the corresponding actual exposure time,
A gamma correction method for an imaging unit, wherein the gamma correction of the imaging unit is performed on a pixel-by-pixel basis based on the created lookup table. 2. The gamma correction method according to claim 1, wherein the unit output is a settable unit gradation value. A selection unit for individually selecting all pixels constituting the imaging unit of the imaging unit in a predetermined order,
A determination unit that determines a base exposure time at which a unit output is obtained when an image is captured under predetermined illumination for sequentially selected pixels,
Setting means for setting the actual exposure time to an integer multiple of the determined base exposure time for the selected pixel;
Measuring means for imaging under the illumination for each set actual exposure time and measuring each output value of the same pixel,
Creation means for creating a look-up table in which each output value measured for each pixel is recorded in association with each input value consisting of an integer value used for calculating the corresponding actual exposure time,
A gamma correction unit for performing gamma correction of the imaging unit on a pixel-by-pixel basis based on the created look-up table. 4. The gamma correction device according to claim 3, wherein the unit output is a settable unit gradation value.

Images