JP7503443B2 - Display MTF measuring device and program thereof - Google Patents

Description

The present invention relates to a display MTF measuring device and a program for measuring the MTF (Modulation Transfer Function), which indicates the spatial frequency characteristics of a display.

In recent years, displays have become increasingly high-resolution with the advent of UHDTV (Ultra High Definition Television), smartphones, tablet devices, and wobblation technology, which doubles the apparent number of pixels in a projector by moving a chip so that the pixel moves back and forth by half a pixel at high speed. The specifications of these resolutions are usually expressed in pixel resolution (number of pixels).
Meanwhile, various new technologies such as multi-chromatic displays having optical color channels other than the conventional RGB stripe, pentile, RGBW, and RGB primary colors, are becoming widespread as pixel structures of displays. In addition, while a conventional RGB stripe configures one pixel with a set of three optical color channels, a multi-chromatic display does not necessarily configure one pixel with RGB and other optical color channels in each pixel, and the pixel unit may be ambiguous, and even displays with the same number of pixels such as 4K and 8K may have different spatial resolution characteristics.
As a result, pixel resolution is no longer an appropriate proxy for the spatial resolution characteristics of a display.
Therefore, there is a need for an objective and accurate method of measuring the spatial resolution characteristics of a display, other than pixel resolution (number of pixels).

Conventionally, a method for photographing a display with a high-resolution camera and analyzing the spatial resolution characteristics of the display involves displaying a square wave (Non-Patent Document 1) or a sine wave (Non-Patent Document 2) on the display and measuring the modulation depth.
The conventional method for measuring the modulation depth of a display involves displaying a black and white square wave pattern P in the horizontal or vertical direction on the display and capturing an image with a camera, as shown in Fig. 22(a). Note that Fig. 22(a) shows an RGBW structure as an example of a pixel structure, and shows an example of displaying a horizontal square wave pattern P in which black and white are displayed for each pixel.
In the conventional method, the modulation depth is measured from an image captured by a camera by applying a one-pixel wide smoothing filter (dashed line in FIG. 22(b)) to the luminance at a sub-pixel position (solid line in FIG. 22(b)), with the horizontal axis representing the camera pixel position (horizontal or vertical direction) and the vertical axis representing the luminance (pixel value), as shown in FIG. 22(b) .

“IDMS(Information Display Measurements Standard)”, SID(Society of Information Display), ICDM(International Committee for Display Metrology), version1.03, pp.109-138, June 1, 2012. Triantaphillidou, S. and Jacobson, R.E., “Measurements of the modulation transfer function of image displays”, Journal of Imaging Science and Technology. 48 (1), pp.58-65, 2004.

A conventional method for measuring the modulation depth of a display involves displaying a horizontal or vertical square wave signal on the display and measuring the modulation depth.
In this case, the conventional method has a problem in that the modulation depth can be measured only in the horizontal or vertical direction.
However, images actually displayed on displays are not limited to horizontal or vertical directions, and diagonal lines are often used as well, and it is desirable to understand the characteristics of such images.

The present invention has been made in consideration of these problems and demands, and aims to provide a display MTF measuring device and a program for the same that can measure the MTF, which indicates the spatial frequency characteristics of a display, without being limited to the horizontal or vertical direction.

To solve the above problem, the display MTF measuring device of the present invention is a display MTF measuring device that measures the display MTF, which represents the spatial frequency characteristics of a display, from a measurement image, which is a drawing of a straight line displayed on a display, captured by a camera, and is configured to include an image input means, an ROI image extraction means, a line projection information generation means, a line spread function generation means, an MTF calculation means, and an MTF correction means.

In such a configuration, the display MTF measuring device inputs, by the image input means, a photographed image obtained by photographing an image for measurement from a camera.
Then, the display MTF measuring device extracts, by the ROI image extraction means, an ROI image of the set region of interest including the straight line from the captured image.
The display MTF measuring device then uses the line projection information generating means to generate line projection information by associating the position of each pixel of the ROI image along the slope of a straight line with a projection axis having sub-pixel intervals on which bins narrower than the pixel width based on the output signal from the camera are arranged horizontally.
The display MTF measurement device then uses the line spread function generating means to average the pixel values of the pixels in the ROI image for each bin of the projection axis associated with the line projection information. The averaged pixel values on the projection axis become a line spread function in the sub-pixel coordinate system of the projection axis, which indicates the spread characteristics of a straight line in the captured image.

The display MTF measuring device then calculates the measured MTF from the line spread function by the MTF calculation means as the MTF measured in the captured image. Note that this measured MTF is measured based on the spatial frequency of the camera, and is a superposition of the display MTF and the camera MTF.
Therefore, the display MTF measuring device corrects the measured MTF by the MTF correction means using the spatial frequency ratio between the display and the camera and the camera MTF, which is the MTF of the camera.
As a result, the MTF correction means converts the spatial frequency of the measured MTF into the spatial frequency of the display, and calculates the display MTF by raising the value attenuated by the camera MTF.

This display MTF measurement device can be operated with a display MTF measurement program that causes a computer to function as each of the above-mentioned means.

The present invention provides the following excellent effects.
According to the present invention, it is possible to measure the display MTF in the spatial frequency direction perpendicular to the lines displayed on the display.
Furthermore, according to the present invention, the display MTF can be measured regardless of the type of subpixel arrangement of the display.

1 is a block diagram showing the configuration of a display MTF measuring device according to a first embodiment of the present invention. 13 is a screen example showing an example of a straight line to be displayed on a display of a measurement target. FIG. 2 is an explanatory diagram for explaining an ROI in an image captured by a camera. FIG. 2 is an explanatory diagram for explaining an ROI image. FIG. 11 is an explanatory diagram for explaining the inclination of an edge of an ROI image. 10 is an explanatory diagram for explaining the correspondence relationship between pixel positions of an ROI image and bin positions of a projection axis. FIG. FIG. 11 is an explanatory diagram for explaining a method for generating a line spread function. FIG. 11 is a graph showing the measured MTF generated by the MTF calculation means, together with the display MTF and the camera MTF. FIG. 11 is a graph illustrating a technique for generating a display MTF by correcting a measured MTF. 5 is a flowchart showing the operation of the display MTF measuring device according to the first embodiment of the present invention. FIG. 11 is a block diagram showing the configuration of a display MTF measuring device according to a second embodiment of the present invention. FIG. 13 is an explanatory diagram for explaining the inclination of a second edge of an ROI image. 13A is a graph showing the relationship between the peak fitting function and the distribution of edge shapes, (a) is a graph of the peak fitting function, (b) is a graph obtained by applying the graph of the peak fitting function to the coordinates of a two-dimensional ROI image, and (c) is a graph of the vertex positions of the peak fitting function on the coordinates of the ROI image. FIG. 11 is an explanatory diagram for explaining the relationship between a peak fitting function and a line width. FIG. 1 is a block diagram showing a configuration of a spatial frequency ratio measuring device according to an embodiment of a reference example. 1 is a screen example showing an example of a point to be displayed on a display of a measurement target and a circle centered on that point; FIG. 2 is an explanatory diagram for explaining an image of a display captured by a camera. 1 is a graph showing the relationship between pixel values and pixel positions in the horizontal or vertical direction. 13 is an example of a screen showing an example in which values of the radius on the top, bottom, left and right sides are displayed based on a center point. 10 is a flowchart showing the operation of a spatial frequency ratio measurement device according to an embodiment of the reference example. FIG. 1 is a block diagram showing the configuration of a display MTF measurement device incorporating a spatial frequency ratio measurement device. FIG. 11 is an explanatory diagram for explaining a conventional method for measuring the display MTF, in which (a) is an example of displaying a square wave on a display with an RGBW pixel structure, and (b) is a graph showing an example of pixel values of a camera image obtained by capturing the square wave with a camera and smoothing it.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
[Configuration of display MTF measuring device]
First, with reference to FIG. 1, a configuration of a display MTF measuring device 1 according to a first embodiment of the present invention will be described.
The display MTF measuring device 1 measures an MTF that represents the spatial frequency characteristic of a display 2 (hereinafter, referred to as display MTF).

The display 2 is a display device for which the display MTF is to be measured. The display 2 may be any device that displays images, such as a UHDTV, a smartphone, a tablet device, a projector, etc. The pixel structure of the display 2 is not limited to RGB stripes, and may be any pixel structure, such as pen tile, RGBW, etc.
The display MTF measuring device 1 operates by connecting a display 2, a camera 3 and a display device 4.

The camera (camera for measuring brightness) 3 is an imaging device for measuring brightness, which images the screen displayed on the display 2 .
The camera 3 captures the screen of the display 2 with a relative pixel magnification of at least 1. Here, the relative pixel magnification is the number of pixels of the camera output signal corresponding to one pixel width of the input signal format of the display 2 captured by the camera 3.
This relative pixel magnification is not particularly limited, but may be, for example, 3 times, 10 times, 30 times, etc. Note that this relative pixel magnification may be 1 time, that is, the spatial frequency of the display 2 and the spatial frequency of the camera 3 may be the same.
Furthermore, the camera 3 is positioned so as to face the display 2 directly, and so that the image captured by the camera 3 is tilted at a few degrees (for example, 3 degrees) with respect to the image displayed on the display 2 .

The display device 4 provides a user interface for operating the display MTF measurement device 1, and displays images captured by the camera 3, measurement results, etc. For example, the display device 4 is a liquid crystal display, an organic EL display, or the like.
The configuration of the display MTF measuring device 1 will now be described in detail.

As shown in FIG. 1, the display MTF measurement device 1 includes a measurement image display means 10, an image input means 11, an ROI setting means 12, an ROI image extraction means 13, a line projection information generation means 14, a line projection information storage means 15, a line spread function generation means 16, an MTF calculation means 17, a camera MTF storage means 18, and an MTF correction means 19.

The measurement image display means 10 displays an image (measurement image) for measuring the display MTF on the display 2. The measurement image display means 10 displays, as a measurement image, an image on the display 2 in which a straight line of a predetermined line width (for example, 1 pixel width) is drawn in any direction in which the MTF is measured. For example, the measurement image display means 10 displays, on the display 2, a measurement image in which a white vertical line is drawn on a background image of a single black pixel value. As a result, as shown in FIG. 2, a straight line L is displayed on the display 2. Note that the measurement image may be one in which the brightness of the pixel values is inverted (black and white inversion).

The image input means 11 inputs an image captured by the camera 3 (a captured image).
When the camera 3 is a video camera, the image input means 11 preferably averages the captured images (frames) that are sequentially input for every predetermined number of frames.
The image input means 11 stores the input photographed image in a memory (not shown).
Due to the difference in spatial frequency between the display 2 and the camera 3, as shown in FIG. 4, a straight line L appears in the captured image G with a line width of a plurality of pixels.
The captured image input by the image input means 11 and stored in the memory is output to the display device 4 via an image output means (not shown), and is also read out by the ROI image extraction means 13 .

The ROI setting means 12 sets a region of interest (ROI) including straight lines within the captured image captured by the camera 3. For example, the ROI setting means 12 sets information identifying the ROI (position and shape, mask data of the closed area, etc.) as ROI information by specifying a closed area (circle, ellipse, rectangle, arbitrary shape, etc.) with a pointing device (not shown) operated by the measurer within the captured image displayed on the display device 4.

For example, the ROI setting means 12 sets ROI information by having the measurer specify an ROI by a rectangle in the photographed image G as shown in Fig. 3. This ROI specifies an ROI image R including a straight line L as shown in Fig. 4. In Fig. 4, the ROI is made small in order to easily explain the pixel values.
Incidentally, if there is no need to change the ROI, the ROI setting means 12 may set the ROI only once.
The ROI setting means 12 outputs ROI information that identifies the set ROI to the ROI image extraction means 13 .

The ROI image extraction means 13 extracts, from the photographed image inputted by the image input means 11, an image of the ROI indicated by the ROI information set by the ROI setting means 12 as an ROI image.
The ROI image extraction means 13 outputs the extracted ROI image to the line projection information generation means 14 .
Furthermore, after the line projection information generating means 14 has completed generation of the line projection information, the ROI image extracting means 13 outputs the ROI image to the line spread function generating means 16 .
Here, the ROI image extraction means 13 outputs the ROI image to the line projection information generation means 14, and after receiving a notification of completion of the generation of the line projection information from the line projection information generation means 14, outputs the ROI image to the line spread function generation means 16.

The line projection information generating means 14 generates line projection information from the ROI image extracted by the ROI image extracting means 13. The line projection information is information that associates pixel positions of the ROI image with bin positions separated by sub-pixel widths of an axis (projection axis) perpendicular to a straight line in the ROI image.
The line projection information generating means 14 includes a line inclination detecting means 14a and a pixel position projecting means 14b.

The line inclination detection means 14a detects the inclination angle of a line on a predetermined reference axis (here, the y-axis) in a coordinate system (xy coordinate system) having two axes, the horizontal and vertical axes, in the ROI image.
The inclination angle of this line can be obtained by a general method. For example, the line inclination detection means 14a detects edges in the ROI image R by Sobel edge detection, and binarizes the ROI image R at the boundaries of the edges e to generate a binary image, as shown in Fig. 5. The line inclination detection means 14a then performs a Hough transform on the binary ROI image R, and detects the inclination angle θe of the edge e as the inclination angle of the line L. At this time, the line inclination detection means 14a obtains the inclination angle θe of the edge e at angle intervals of, for example, about 0.1 degrees.
The line inclination detection means 14a outputs the detected line inclination angle to the pixel position projection means 14b.

The pixel position projection means 14b associates each pixel of the ROI image with a position obtained by projecting the pixel along the inclination of the line onto an axis (projection axis) perpendicular to the line in the ROI image.
The projection axis is in sub-pixel units smaller than the pixel units of the ROI image, for example, 1/4 or 1/8 of a pixel.
Specifically, as shown in FIG. 6, the pixel position projection means 14b rotates each pixel position (x, y) of the ROI image R by an inclination angle (-θe) in the opposite direction to the rotation direction of the inclination angle of the line (the inclination angle θe of edge e) around a predetermined rotation center (e.g., the average value of the x and y coordinates of the pixels of the ROI image).
Then, the pixel position projection means 14b generates line projection information by associating each pixel position (x, y) of the ROI image R before rotation with the position of a bin separated by a sub-pixel width on the projection axis in sub-pixel units onto which the pixel is projected.
The pixel position projection means 14 b stores the generated line projection information in the line projection information storage means 15 .
After the pixel position projection means 14 b stores the line projection information in the line projection information storage means 15 , the line projection information generation means 14 outputs a completion notification of the generation of the line projection information to the ROI image extraction means 13 .

This method of generating projection information by binarizing the ROI image obtained by applying a low-pass filter to the captured image, calculating the tilt angle using a Hough transform, and projecting it onto the bins of the projection axis, as in the line projection information generating means 14, can also be used when generating projection information from conventional images of rectangular waves or sine waves in a fixed direction.

The line projection information storage means 15 stores the line projection information generated by the line projection information generation means 14. The line projection information storage means 15 is, for example, a general storage device such as a semiconductor memory.
The line projection information stored in the line projection information storage means 15 is referenced by the line spread function generating means 16 .

The line spread function generating means 16 generates a line spread function (LSF) from the ROI image extracted by the ROI image extracting means 13 .
Specifically, as shown in FIG. 7, the line spread function generating means 16 reads out the pixel values of pixels projected onto the same bin on the projection axis from the ROI image extracted by the ROI image extraction means 13 based on the line projection information stored in the line projection information storage means 15, and averages them for each bin.
This enables the line spread function generating means 16 to generate a line spread function LSF that indicates the characteristics of a line by associating sub-pixel positions on the projection axis with pixel values.
The line spread function generating means 16 outputs the generated line spread function to the MTF calculating means 17 .

The MTF calculation means 17 calculates the MTF from the line spread function generated by the line spread function generation means 16 .
Specifically, the MTF calculation means 17 performs a Fourier transform on the line spread function to obtain an absolute value, and normalizes the absolute value with a constant component (DC component) to obtain an MTF value for each spatial frequency.
The MTF (measured MTF) calculated by the MTF calculation means 17 is the MTF of the display 2 (display MTF) plus the MTF of the camera 3 (camera MTF).

Specifically, as shown in Figures 8 and 9, when the horizontal axis is the spatial frequency and the vertical axis is the MTF, the measured MTF (MTF _DISP +CAM ) is measured in which the camera MTF (MTF _CAM ) is superimposed on the display MTF (MTF _DISP ) that is originally to be measured. In Figures 8 and 9, f _SDISP and f _NDISP respectively indicate the sampling frequency and Nyquist frequency of the display 2 based on the pixels of the camera 3. Also, f _SCAM and f _NCAM respectively indicate the sampling frequency and Nyquist frequency of the camera 3 based on the pixels of the camera 3. Also, in Figure 8, the horizontal axis is the spatial frequency f _CAM of the camera 3, and in Figure 9, the horizontal axis is the spatial frequency f _DISP of the display 2.
The MTF calculation means 17 outputs the calculated measured MTF to the MTF correction means 19 .

The camera MTF storage means 18 stores the previously measured MTF (camera MTF) of the camera 3. The camera MTF storage means 18 is, for example, a general storage device such as a semiconductor memory.
The camera MTF is data in which an MTF value is associated with each spatial frequency.
The camera MTF previously measured by a known camera MTF measuring device can be used as the camera MTF previously stored in the camera MTF storage means 18. Known camera MTF measuring devices are disclosed in, for example, Japanese Patent Application Laid-Open No. 2018-136222, Japanese Patent Application Laid-Open No. 2018-013416, Japanese Patent Application Laid-Open No. 2015-094701, etc.

The MTF correction means 19 corrects the MTF (measured MTF) calculated by the MTF calculation means 17 to the MTF of the display 2 using the spatial frequency ratio of the display 2 and the camera 3 and the camera MTF stored in the camera MTF storage means 18.
The spatial frequency ratio (e.g., the ratio of sampling frequencies) between the display 2 and the camera 3 is assumed to be set in advance from the outside as known information. This spatial frequency ratio can be calculated from the pixel sizes of the display 2 and the camera 3, respectively, and the magnification of the camera 3. The magnification of the camera 3 can be calculated from the focal length of the lens of the camera 3 and the subject distance.

Now, the correction of the measured MTF in the MTF correcting means 19 will be described with reference to FIGS.
The measured MTF value at spatial frequency f is MTF _DISP+CAM (f), the camera MTF value is MTF _CAM (f), and the spatial frequency ratio which is the magnification of the spatial frequency of the display 2 to the spatial frequency of the camera 3 is m (m=f _SDISP /f _SCAM ), where f _SDISP is the sampling frequency of the display 2, and f _SCAM is the sampling frequency of the camera 3.

In the line spread function generating means 17, when averaging is performed for each bin as described in FIG. 13, the positions of the projected pixels are uniformly scattered across the width of each bin, and the values of pixels with different bin phases (binning phases) within the bin width are averaged, resulting in a low-pass filter being applied.
Therefore, MTF _DISP+CAM (f) is obtained by correcting MTF _{DISP+CAM+binning} (f), which is the actual measured MTF after applying a low-pass filter, according to the following equation (1).

Then, the MTF correction means 19 calculates MTF _DISP (f _DISP ), which is the value of the display MTF at the spatial frequency f _DISP , by the following equation (2).

Since the sampling points of the camera MTF and the display MTF are not aligned, the MTF correction means 19 calculates the respective MTFs with aligned sampling points by interpolation from the measured MTF MTF _DISP+CAM (f _DISP /m) and from the camera MTF MTF _CAM (f _DISP /m), and uses these in the above formula (2).
This enables the MTF correction means 19 to obtain the display MTF.
The MTF correction means 19 outputs the corrected display MTF as a measurement result to the outside (to the display device 4).

As described above, the display MTF measuring device 1 can measure the display MTF by using the vertical lines displayed on the display 2.
Furthermore, since the display MTF measuring device 1 determines the line spread function from the slope of a line displayed on the display 2, it is possible to measure the display MTF regardless of the pixel structure.
The display MTF measuring device 1 can be operated by a program (display MTF measuring program) for causing a computer to function as each of the above-mentioned means.

[Operation of Display MTF Measuring Device]
Next, the operation of the display MTF measuring device 1 will be described with reference to FIG. 10 (for the configuration, see FIG. 1).
It is assumed that the MTF of the camera 3 measured in advance (camera MTF) is stored in the camera MTF storage means 18. Also, it is assumed that the spatial frequency ratio between the display 2 and the camera 3 is known and is input from the outside.

In step S1, the measurement image display means 10 displays on the display 2 an image in which a straight line having a width of 1 pixel is drawn as an image for measuring the display MTF (measurement image).
In step S2, the image input means 11 inputs an image (photographed image) captured by the camera 3. At this time, the camera 3 captures the image while being tilted at an angle of several degrees with respect to the vertical line displayed on the display 2.

In step S3, the ROI setting means 12 sets a region of interest (ROI) including a straight line within the photographed image input in step S2 (see FIG. 3).
In step S4, the ROI image extraction means 13 extracts an image of the ROI set in step S3 (ROI image) from the photographed image input in step S2.

In step S5, the line inclination detection means 14a of the line projection information generation means 14 detects the inclination angle of the line from the ROI image extracted in step S4 (see FIG. 5). For example, the line inclination detection means 14a generates a binary image by binarizing the boundaries of edges detected by Sobel edge detection in the ROI image, and detects the inclination angle of the line by performing a Hough transform on the binary image.

In step S6, the pixel position projection means 14b of the line projection information generation means 14 generates line projection information by associating each pixel of the ROI image extracted in step S4 with the position where the pixel is projected along the inclination angle of the line detected in step S5 on an axis (projection axis) perpendicular to the line in the ROI image (see FIG. 6). At this time, the projection axis is in sub-pixel units smaller than the pixel unit of the ROI image.

In step S7, the pixel position projection means 14b stores the line projection information generated in step S6 in the line projection information storage means 15.
In step S8, the line spread function generating means 16 generates a line spread function (LSF) from the ROI image extracted in step S4. Here, the line spread function generating means 16 reads out pixel values of pixels projected onto the same sub-pixel bin of the projection axis from the ROI image based on the line projection information stored in the line projection information storage means 15, and averages the pixel values for each bin to generate a line spread function (see FIG. 7).

In step S9, the MTF calculation means 17 calculates the MTF from the line spread function generated in step S8. Here, the MTF calculation means 17 performs a Fourier transform on the line spread function to obtain the absolute value for each spatial frequency, and normalizes it by a constant component (DC component) to calculate the MTF (measured MTF).

In step S10, the MTF correction means 19 calculates the MTF of the display 2 (display MTF) by correcting the difference in spatial frequency and the attenuation of the MTF through the camera 3 using the MTF calculated in step S9 (measured MTF) based on the spatial frequency ratio between the display 2 and the camera 3 and the camera MTF stored in the camera MTF storage means 18.
By the above operation, the display MTF measuring device 1 can measure the display MTF by using the vertical lines displayed on the display 2.

[Another configuration of the display MTF measuring device]
Next, the configuration of a display MTF measuring apparatus 1B according to a second embodiment of the present invention will be described with reference to FIG.
The display MTF measuring device 1B measures the MTF of a display 2 (display MTF) in the same manner as the display MTF measuring device 1 (FIG. 1).
The display MTF measuring device 1 (FIG. 1) detects the line inclination angle by Hough transform at angle intervals of about 0.1 degree in the line inclination detection means 14a of the line projection information generating means 14, whereas the display MTF measuring device 1B is further configured to detect the line inclination angle with high precision.

As shown in FIG. 11, the display MTF measuring device 1B comprises a measurement image display means 10, an image input means 11, an ROI setting means 12, an ROI image extraction means 13, a line projection information generation means 14B, a line projection information storage means 15, a line spread function generation means 16, an MTF calculation means 17, a camera MTF storage means 18, and an MTF correction means 19.
The configuration other than the line projection information generating means 14B is the same as that of the display MTF measuring device 1 (FIG. 1), so the same reference numerals are used and the description will be omitted.

The line projection information generating means 14 B generates line projection information from the ROI image extracted by the ROI image extracting means 13 .
The line projection information generating means 14B includes a line inclination detecting means 14a, a second line inclination detecting means 14a2, and a pixel position projecting means 14b2.

The line inclination detection means (first line inclination detection means) 14a has the same configuration as the display MTF measurement device 1 (FIG. 1). Here, the line inclination detection means 14a detects, by a Hough transform, an inclination angle θe of an edge e with respect to a predetermined reference axis (y-axis) at angle intervals of about 0.1 degrees as the inclination angle of a line L, as shown in FIG.
The line inclination detection means 14a outputs the detected line inclination angle to a second line inclination detection means 14a2.

The second line inclination detection means 14a2 detects the inclination of the most approximate peak position of the distribution of pixel values of the rotated ROI image, which has been corrected by rotating the ROI image with the inclination angle of the line detected by the line inclination detection means 14a, and the distribution of function values obtained by applying a peak fitting function on an axis perpendicular to the reference axis in the direction of the reference axis, as the deviation amount of the inclination angle of the line.

The second line inclination detection means 14a2 rotates each pixel position (x, y) of the ROI image R shown in FIG. 5, which is arranged in a coordinate system (xy coordinate system) with two axes, horizontal and vertical, around a predetermined rotation center (e.g., the average value of the x and y coordinates of the pixels of the ROI image) by an inclination angle (-θe) in the opposite direction to the rotation direction of the line inclination angle (inclination angle θe of edge e), and further detects the inclination angle θe' of the line with respect to the reference axis (y axis) as shown in FIG. 12.

The second line inclination detection means 14a2 calculates, as the line inclination angle θe', the inclination of the peak position with respect to the reference axis that is most similar between the distribution of pixel values of the ROI image (hereinafter referred to as the rotated ROI image) rotated by the inclination angle θe shown in FIG. 12 and the distribution of function values obtained by applying a peak fitting function on an axis (x-axis) perpendicular to the reference axis in the direction of the reference axis (y-axis).

Now, a method for calculating the inclination angle θe′ in the second line inclination detection means 14a2 will be described with reference to FIG.
13A is a graph showing an example of a peak function used in the peak fitting function. The horizontal axis indicates the x-coordinate, and the vertical axis indicates the function value v. The peak fitting function is a peak function that has a convex shape at the peak position (x=p ₁ ). For example, this peak function is a Gaussian function, a Lorentz function, or the like. Alternatively, a function may be used in which a cosine curve, a quadratic curve, a triangular wave, or the like is used for the convex portion, and a constant such as "0" is used for the other portions.
FIG. 13B shows the distribution of function values v(x, y) calculated by applying the peak fitting function of FIG. 13A in the reference axis (y-axis) direction.

When there is a line inclination with respect to the reference axis (y-axis) in the rotated ROI image shown in FIG. 12, if the distribution of pixel values in the rotated ROI image and the distribution of function values in FIG. 13(b) are most closely approximated, the peak position _p1 of the peak fitting function will be inclined by θe′ with respect to the reference axis (y-axis), as shown in FIG. 13(c).
Therefore, the second line inclination detection unit 14a2 calculates a more accurate line inclination angle θe′ by determining the inclination _p2 of the peak position _p1 with respect to the reference axis (y-axis). Note that in FIG. 13C, μ represents ( _p1 + _p2y ).
Hereinafter, the peak fitting function is the function shown in the following formula (3).

Furthermore, σ is a constant that determines the spread of the function.
The second line inclination detection unit 14a2 searches for parameters p ₁ and p ₂ that maximize the correlation between the pixel values of the rotated ROI image shown in FIG. 12 and the function value v(x, y) of the above equation (3).

In the above formula (3), _p1 is the peak position of the peak fitting function, and since the ROI is set in the ROI setting means 12 with the line as the approximate center, it is preferable to set the average value of the x coordinate of the rotated ROI image or "0" as the initial value. Also, _p2 is the inclination of the peak position with respect to the reference axis, and since it is approximately "0" in the rotated ROI image, it is preferable to set the initial value to "0". This can increase the search speed of the parameters _p1 and _p2 . This can also prevent a search for an erroneous local solution and stabilize the optimization.

It should be noted that the lines of the ROI image have different thicknesses depending on the ratio of the spatial frequencies of the display 2 and the camera 3 .
14, when the function of the above-mentioned formula (3) is fitted to a line L of a rotated ROI image, in order to prevent variation in fitting, it is preferable that the full width at half maximum (FWHM) of the function corresponds to the number of pixels of the width of the line L. Note that the full width at half maximum (FWHM) does not necessarily have to coincide with the number of pixels of the width of the line L, and it is preferable that the value be close to the number of pixels of the width of the line L.
Therefore, σ in the above formula (3) is, for example, m/3, where m is the magnification of the spatial frequency of the camera 3 relative to the spatial frequency of the display 2, that is, the relative pixel magnification which is the number of pixels of the camera 3 corresponding to the width of one pixel of the display 2. In addition, in Fig. 14, μ represents ( _p1 + _p2y ) in the above formula 2.
This allows the second line inclination detection unit 14a2 to search for p ₁ and p ₂ with high accuracy.

The second line inclination detection unit 14a2 calculates the arctangent of the following equation (4) for the found parameter _p2 to find the line inclination angle θe′.

The second line inclination detection means 14a2 outputs the rotated ROI image and the line inclination angle θe' calculated by the peak fitting function to the pixel position projection means 14b2.

The pixel position projection means 14b2 rotates the rotated ROI image so that the line is perpendicular to the projection axis, and associates the pixel positions (bins) projected onto the projection axis perpendicular to the line with the pixel positions of the original ROI image. Note that the projection axis is in sub-pixel units, which are smaller than the pixel units of the ROI image.

Specifically, the pixel position projection means 14b2 generates a corrected rotated ROI image in which the tilt is corrected by rotating the rotated ROI image generated by the line tilt detection means 14a by an inclination angle (-θe') in the opposite direction to the rotation direction of the line inclination angle θe' detected by the second line inclination detection means 14a2.
Then, the pixel position projection means 14b2 generates line projection information by associating each pixel position (x, y) of the ROI image R before rotation with the position of a bin separated by a sub-pixel width of the projection axis in sub-pixel units onto which the pixels of the corrected rotated ROI image are projected in the vertical direction.
In addition, the pixel position projection means 14b2 may rotate the ROI image extracted by the ROI image extraction means 13 by a tilt angle (-(θe+θe')) in the opposite direction to the rotation direction of the tilt angle (θe+θe') at once to generate a corrected rotated image, and project it onto the vertical projection axis.

As a result, in addition to the effects of the display MTF measuring device 1, the display MTF measuring device 1B can accurately determine the slope of the line and improve the calculation accuracy of the display MTF.
The display MTF measuring apparatus 1B can be operated by a program (display MTF measuring program) for causing a computer to function as each of the above-mentioned means.

The operation of the display MTF measurement device 1B is basically the same as that of the display MTF measurement device 1, except that in step S5 of FIG. 10, the line inclination angle is detected in two stages by the line inclination detection means 14a and the second line inclination detection means 14a2. Therefore, a description thereof will be omitted.

Although the display MTF measuring devices 1 and 1B have been described above, the configuration and operation of the display MTF measuring devices 1 and 1B are not limited to this embodiment.
For example, although a vertical line is drawn on the display 2 here, a horizontal line, a line at a 45 degree angle, or a line in any other direction may also be drawn.
Also, here, the display MTF measurement devices 1, 1B are connected to a display 2, and a straight line is displayed on the screen of the display 2 by the measurement image display means 10. However, if the display 2 is capable of displaying an image depicting a straight line independently of the display MTF measurement devices 1, 1B, the display MTF measurement devices 1, 1B do not need to be connected to the display 2. In that case, the display MTF measurement devices 1, 1B may omit the measurement image display means 10 from their configuration.

Further, here, the line projection information generating means 14, 14B generate the line projection information when the ROI image extracting means 13 extracts the ROI image.
However, if the camera 3 is a video camera and the inclination of the straight line displayed on the display 2 can be considered to be invariant for the captured images sequentially input, it is not necessary to generate the line projection information sequentially. In this case, the line projection information generating means 14, 14B need only generate the line projection information the first time the ROI is set by the ROI setting means 12, and store it in the line projection information storage means 15.

In addition, here, the second line inclination detection means 14a2 detects the amount of deviation of the line inclination angle from the reference axis in a rotated ROI image obtained by rotating the ROI image by the line inclination angle detected by the line inclination detection means 14a, using the above-mentioned equations (3) and (4).
However, the second line inclination detection means 14a2 may not rotate the original ROI image, but may instead set the initial value of parameter _p2 in equation (3) to the non-zero inclination angle of the line detected by the line inclination detection means 14a, and detect the corrected inclination angle by correcting the inclination angle of the line with respect to the reference axis of the original ROI image.
This corresponds to the second line tilt detection means 14a2 detecting, as a corrected tilt angle, the tilt of the peak position that is most similar between the distribution of pixel values of the original ROI image extracted by the ROI image extraction means 13 and the distribution of function values obtained by applying a peak fitting function on an axis perpendicular to the reference axis in the direction of the reference axis.

In this case, the second line inclination detection unit 14a2 outputs the detected corrected inclination angle to the pixel position projection unit 14b2.
Then, the pixel position projection means 14b2 rotates the original ROI image extracted by the ROI image extraction means 13 by the corrected tilt angle detected by the second line tilt detection means 14a2 so that the line is perpendicular to the projection axis. This makes it possible to generate a corrected rotated ROI image by rotating it by one degree.

[Configuration of spatial frequency ratio measuring device (relative pixel magnification measuring device)]
Next, the configuration of a spatial frequency ratio measuring apparatus 5 according to a reference embodiment will be described with reference to FIG.
The spatial frequency ratio (relative pixel magnification) input to the display MTF measuring device 1 (FIG. 1) is information that can be calculated from the pixel sizes of the display 2 and the camera 3, respectively, and the magnification of the camera 3. However, here, a spatial frequency ratio measuring device 5 that determines the spatial frequency ratio between the display 2 and the camera 3 by measurement will be described.

The spatial frequency ratio measuring device 5 operates by connecting the display 2, the camera 3 and the display device 4.
The display 2 is a display device for measuring the spatial frequency ratio. The display 2 may be the same as the display 2 described with reference to FIG.
The camera 3 is an imaging device for measuring the spatial frequency ratio. The camera 3 may be the same as the camera 3 described in FIG. 1. The camera 3 captures the screen of the display 2 at a spatial frequency higher than the spatial frequency of the display 2.
Although the camera 3 is disposed facing the display 2 , it may be rotated in a plane parallel to the screen of the display 2 .
The display device 4 displays images captured by the camera 3, measurement results, etc. For example, the display device 4 is a liquid crystal display, an organic EL display, or the like.
The configuration of the spatial frequency ratio measuring device 5 will now be described in detail.

As shown in FIG. 15, the spatial frequency ratio measuring device 5 includes a measurement image display means 50, an image input means 51, a circle center search means 52, a local peak search means 53, a peak-to-peak pixel number calculation means 54, and a pixel number ratio calculation means 55.

The measurement image display means 50 displays an image (measurement image) for measuring the spatial frequency ratio on the display 2. The measurement image is an image in which a circle and a center point are drawn on the single pixel value image such that the brightness differs in the order of the single pixel value image, the circle, and the center point of the circle.
The measurement image display means 50 displays on the display 2 a measurement image in which a point and a circle of a predetermined radius (e.g., 10 pixels) with the point as its center point are drawn at an arbitrary position on the display 2. For example, as shown in Fig. 16, the measurement image display means 50 displays on the display 2 a measurement image in which a white point (center point O) with a higher (brighter) pixel value than the background image and a circle C are drawn on a black background image. The center point O has a higher (brighter) pixel value than the circle C. Note that the measurement image may have the brightness of the pixel values inverted (black and white inversion).

The center point O and the circle C are approximately one pixel wide, but to improve distinguishability in the camera 3, it is preferable that the measurement image display means 50 performs anti-aliasing processing on the measurement image to make the center point O and the circle C several pixels wide.

The image input means 51 inputs an image captured by the camera 3 (a captured image).
17, the captured image G input by the image input means 51 is captured in a state in which the center point O and the circle C have higher pixel values than the background image. Also, the captured image G is captured in a state in which the center point O has a higher pixel value than the circle C.
It is preferable that the image input means 51 applies a low-pass filter such as a Gaussian filter to the captured image, thereby enabling the image input means 51 to blur the pixel structure (sub-pixel structure) of the display 2 captured in the captured image and suppress noise of the camera 3. In addition, when the camera 3 is a video camera, it is preferable that the image input means 51 averages the captured images (frames) input sequentially for every predetermined number of frames.
The image input means 51 outputs the input photographed image to the circle center search means 52 and the local peak search means 53 .

The circle center search means 52 searches for the position of the center point of a circle in a photographed image. Note that in the photographed image G shown in Fig. 17, the center point O has a higher pixel value than the circle C, so the circle center search means 52 determines the pixel position with the highest pixel value in the photographed image G as the position of the center point of the circle.
The circle center search means 52 outputs the position of the center point of the circle that has been searched for to the local peak search means 53 .

The local peak search means 53 searches for the position of a local peak of pixel values on horizontal and vertical lines in the photographed image that pass through the center point searched for by the circle center search means 52 .
On a horizontal line passing through the center point, there are three local peaks: the center point and two points where the horizontal line and the circle intersect.
Similarly, on a vertical line passing through the center point, there are three local peaks: the center point and two points where the vertical line and the circle intersect.

For example, when the position of the center point in the captured image is pixel position "150" on a horizontal line and the radius of the circle is "100", as shown in Fig. 18, on a graph with pixel position (here, pixel position on the horizontal line) on the horizontal axis and pixel value on the vertical axis, pixel positions "50", "150", and "250" are positions of local peaks where the pixel values are locally higher than other pixels. The same applies to vertical lines.
The local peak search means 53 outputs the pixel positions of the three local peaks found on the horizontal line and the vertical line to the peak-to-peak pixel number calculation means 54 .

The peak-to-peak pixel number calculation means 54 calculates the number of pixels between the local peaks found by the local peak search means 53 for each of the horizontal and vertical directions.
The peak-to-peak pixel number calculation means 54 calculates the difference between the pixel position of the right local peak and the pixel position of the central local peak from the pixel positions of the three local peaks on the horizontal line. This difference corresponds to the number of pixels in the radius on the right side from the center of the circle. The peak-to-peak pixel number calculation means 54 also calculates the difference between the pixel position of the central local peak and the pixel position of the left local peak. This difference corresponds to the number of pixels in the radius on the left side from the center of the circle.

The peak-to-peak pixel number calculation means 54 also calculates the difference between the pixel position of the lower local peak and the pixel position of the central local peak from the pixel positions of the three local peaks on the vertical line. This difference corresponds to the number of pixels in the radius below the center of the circle. The peak-to-peak pixel number calculation means 54 also calculates the difference between the pixel position of the central local peak and the pixel position of the upper local peak. This difference corresponds to the number of pixels in the radius above the center of the circle.
In this manner, the peak-to-peak pixel number calculation means 54 calculates the radii in four directions from the center point of the circle.
The peak-to-peak pixel number calculation means 54 then calculates the average value of the radii in the four directions, and sets this as the number of pixels in the radius of the photographed circle.

If the difference between the radii in the four directions is greater than a predetermined error (for example, one pixel), it means that the camera 3 is not facing the display 2. Therefore, it is preferable that the peak-to-peak pixel number calculation means 54 displays the radii in the four directions on the display device 4, for example, as shown in Fig. 19. Fig. 19 shows an example in which a center point and a circle are drawn on the screen, and the number of pixels of the radius is displayed near a point on the circumference that has been found to be a local peak.
This makes it easier for the person taking the measurements to position the camera 3 so that it faces the display 2 directly.
The peak-to-peak pixel number calculation means 54 outputs the number of pixels of the radius of the circle, which is the calculated number of pixels between the local peaks, to the pixel number ratio calculation means 55 .

The pixel ratio calculation means 55 calculates the ratio between the number of pixels of the radius of the circle in the measurement image displayed on the display 2 (the number of pixels of the display 2) and the number of pixels of the radius of the circle in the captured image (the number of pixels of the camera 3).
The pixel ratio calculation means 55 may set the radius of the circle of the measurement image displayed on the display 2 in advance via an external input device (not shown), or may input it from the measurement image display means 50.
Furthermore, the pixel number ratio calculation means 55 receives the number of pixels of the radius of the circle on the photographed image from the peak-to-peak pixel number calculation means 54 .
The ratio between the number of pixels of the radius of the circle in the measurement image (the number of pixels of display 2) and the number of pixels of the radius of the circle in the captured image (the number of pixels of camera 3) corresponds to the ratio between the spatial frequency of display 2 and the spatial frequency of camera 3 based on the pixels of camera 3.
Here, if the number of pixels of the radius of the circle in the measurement image is N _D (pixel _DISP ), the number of pixels of the radius of the circle on the captured image is N _C (pixel _CAM ), the spatial frequency of the display 2 based on the pixels of the camera 3 is F _D (cycles/pixel _CAM ), and the spatial frequency of the camera 3 is F _C (cycles/pixel _CAM ), then N _C /N _D =F _C /F _D. Note that the pixel size (pixel _DISP ) of the display 2 is usually made larger than the pixel size (pixel _CAM ) of the camera 3.
The pixel number ratio calculation means 55, for example, divides the number of pixels N _C of the radius of the circle on the captured image by the number of pixels N _D of the radius of the circle in the measurement image to obtain a spatial frequency ratio (relative pixel magnification) between the display 2 and the camera 3 based on the pixels of the camera 3.

As described above, the spatial frequency ratio measuring device 5 can measure the spatial frequency ratio between the display 2 and the camera 3 even if the pixel sizes of the display 2 and the camera 3 and the magnification of the camera 3 are unknown.
The spatial frequency ratio measuring device 5 can be operated by a program (spatial frequency ratio measuring program) for causing a computer to function as each of the above-mentioned means.

[Operation of spatial frequency ratio measuring device]
Next, the operation of the spatial frequency ratio measuring device 5 will be described with reference to FIG. 20 (for the configuration, see FIG. 15).

In step S21, the measurement image display means 50 displays on the display 2 an image (measurement image) for measuring the spatial frequency ratio, which is an image depicting a point and a circle of a predetermined radius with the point as its center point. Note that it is preferable that the measurement image display means 50 applies anti-aliasing to the measurement image in advance, so that the center point and the circle have a width of several pixels.

In step S22, the image input unit 51 inputs an image (photographed image) captured by the camera 3. Note that it is preferable that the image input unit 51 applies a low-pass filter to the photographed image.
In step S23, the circle center search means 52 searches for the pixel position having the highest pixel value in the photographed image as the position of the center point of the circle.

In step S24, the local peak search means 53 searches for the position of a local peak on the horizontal line and vertical line in the captured image that pass through the center point searched for in step S23. As a result, the center point and the point where the horizontal line intersects with the circle on the horizontal line, and the center point and the point where the vertical line intersects with the circle on the vertical line are searched for as the position of a local peak.

In step S25, the peak-to-peak pixel number calculation means 54 calculates the number of pixels between the local peaks found in step S24 for each of the horizontal and vertical directions. Here, the peak-to-peak pixel number calculation means 54 calculates the number of pixels between the center point and each of the local peak points on the upper, lower, left, and right sides.
At this time, although not shown as a step, the peak-to-peak pixel number calculation means 54 displays the number of pixels between the peaks in each of the four directions on the display device 4 (see FIG. 19). If the measurer determines from the number of pixels between the peaks in each of the four directions that the camera 3 is not directly facing the display 2, the measurer can simply reposition the camera 3 and perform the measurement from the beginning.

In step S26, the peak-to-peak pixel number calculation means 54 calculates the average number of pixels from the center point calculated in step S25 to four points above, below, left and right, as the radius of the circle.
In step S27, the pixel number ratio calculation means 55 calculates the pixel number ratio between the radius of the circle in the measurement image displayed on the display 2 in step S21 and the radius of the circle on the captured image taken by the camera 3 calculated in step S26 as a spatial frequency ratio.

By the above operations, the spatial frequency ratio measuring device 5 can measure the spatial frequency ratio between the display 2 and the camera 3.
This allows the spatial frequency ratio measuring apparatus 5 to determine by measurement the spatial frequency ratio described in the display MTF measuring apparatus 1 shown in FIG.

The spatial frequency ratio measured by the spatial frequency ratio measuring device 5 can be used not only in the display MTF measuring device 1, but also, for example, when measuring the display MTF using conventional methods.

Although the spatial frequency ratio measurement device 5 has been described above, the configuration and operation of the spatial frequency ratio measurement device 5 are not limited to this embodiment.
For example, here, the peak-to-peak pixel number calculation means 54 calculates the number of pixels of the radius of the circle by the number of pixels between the local peak at the center and the local peaks above, below, left, and right. However, the size of the circle can also be specified by the diameter instead of the radius.
In this case, the local peak search means 53 searches for the positions of two local peaks, excluding the center point, on a horizontal line and a vertical line passing through the center point of the circle in the captured image searched for by the circle center search means 52. The peak-to-peak pixel number calculation means 54 then calculates a diameter based on the number of pixels between the upper and lower local peaks and between the left and right local peaks. The pixel number ratio calculation means 55 then calculates a spatial frequency ratio based on the ratio of the number of pixels between the diameter of the circle in the measurement image displayed on the display 2 and the diameter of the circle calculated by the peak-to-peak pixel number calculation means 54.

Here, the spatial frequency ratio measurement device 5 is connected to a display 2, and displays a point and a circle with the point as its center on the screen of the display 2 using a measurement image display means 50. However, if the display 2 is capable of displaying an image of a point and a circle independently of the spatial frequency ratio measurement device 5, the spatial frequency ratio measurement device 5 does not need to be connected to a display 2. In that case, the spatial frequency ratio measurement device 5 may omit the measurement image display means 50 from its configuration.

The spatial frequency ratio measurement device 5 may be provided inside the display MTF measurement device 1 described in Fig. 1. For example, as shown in Fig. 21, the display MTF measurement device 1 may be provided with the components of the spatial frequency ratio measurement device 5, that is, the measurement image display means 50, the image input means 51, the circle center search means 52, the local peak search means 53, the peak-to-peak pixel number calculation means 54, and the pixel number ratio calculation means 55, as a display MTF measurement device 1B. In this case, the measurement image display means 10, 50 and the image input means 11, 51 may be switched between functioning as the display MTF measurement device 1 and functioning as the spatial frequency ratio measurement device 5, depending on an external instruction.
Of course, the spatial frequency ratio measurement device 5 may be configured to be provided inside the display MTF measurement device 1B described in Fig. 11. In that case, the line projection information generation means 14 in Fig. 21 may be replaced with the line projection information generation means 14B described in Fig. 11.

REFERENCE SIGNS LIST 1, 1B, 1C Display MTF measuring device 10 Measurement image display device 11 Image input means 12 ROI setting means 13 ROI image extraction means 14, 14B Line projection information generating means 14a Line inclination detecting means (first line inclination detecting means)
14a2 Second line inclination detection means 14b, 14b2 Pixel position projection means 15 Line projection information storage means 16 Line spread function generation means 17 MTF calculation means 18 Camera MTF storage means 2 Display 3 Camera 4 Display device 5 Spatial frequency ratio measurement device 50 Measurement image display means 51 Image input means 52 Circle center search means 53 Local peak search means 54 Peak-to-peak pixel number calculation means 55 Pixel number ratio calculation means

Claims (9)

1. A display MTF measurement device for measuring a display MTF representing a spatial frequency characteristic of a display from a measurement image, the measurement image being a straight line drawn on the display, captured by a camera, the display MTF measurement device comprising:
an image input means for inputting the captured image from the camera;
ROI image extraction means for extracting an ROI image of a set region of interest including the line from the captured image;
a line projection information generating means for generating line projection information by associating a position of each pixel of the ROI image with a projection axis of a bin at sub-pixel intervals along a gradient of the line;
a line spread function generating means for averaging pixel values of each pixel of the ROI image for each bin of the projection axis associated with the line projection information to generate a line spread function;
an MTF calculation means for calculating a measured MTF as an MTF measured in the captured image from the line spread function;
an MTF correction unit that calculates the display MTF by correcting the measured MTF using a spatial frequency ratio between the display and the camera and a camera MTF that is an MTF of the camera ;
When the value of the measured MTF at a certain spatial frequency f is MTF _{DISP + CAM + binning} (f), the value of the camera MTF is MTF _CAM (f), and the spatial frequency ratio which is the magnification of the spatial frequency of the camera to the spatial frequency of the camera is m,
A display MTF measuring device characterized in that the MTF correction means corrects MTF _DISP+CAM (f _CAM ) at the spatial frequency f _CAM of the camera by MTF _{DISP+CAM+binning} (f _CAM )/sinc(f _CAM /n _bin ), and then calculates MTF _DISP (f _DISP ), which is the display MTF value at the spatial frequency f _DISP of the display, by MTF _DISP (f _DISP ) = MTF _DISP+CAM (f _DISP /m)/MTF _CAM (f _DISP /m) . 1. A display MTF measurement device for measuring a display MTF representing a spatial frequency characteristic of a display from a measurement image, the measurement image being a straight line drawn on the display, captured by a camera, the display MTF measurement device comprising:
an image input means for inputting the captured image from the camera;
ROI image extraction means for extracting an ROI image of a set region of interest including the line from the captured image;
a line projection information generating means for generating line projection information by associating a position of each pixel of the ROI image with a projection axis of a bin at sub-pixel intervals along a gradient of the line;
a line spread function generating means for averaging pixel values of each pixel of the ROI image for each bin of the projection axis associated with the line projection information to generate a line spread function;
an MTF calculation means for calculating a measured MTF as an MTF measured in the captured image from the line spread function;
an MTF correction unit that calculates the display MTF by correcting the measured MTF using a spatial frequency ratio between the display and the camera and a camera MTF that is an MTF of the camera ;
The line projection information generating means
a first line inclination detection means for detecting an inclination angle of the line with respect to a predetermined reference axis by a Hough transform;
a second line inclination detection means for detecting, as a deviation amount of the inclination angle of the line, an inclination of a peak position that is most approximated between a distribution of pixel values of a rotated ROI image obtained by rotating the ROI image by the inclination angle of the line detected by the first line inclination detection means and a distribution of function values obtained by applying a peak fitting function on an axis perpendicular to the reference axis in the direction of the reference axis;
a pixel position projection means for projecting pixels of a corrected rotated ROI image obtained by rotating the rotated ROI image by the deviation amount so that the line is perpendicular to the projection axis onto the projection axis, and for generating the line projection information by associating positions of each pixel of the original ROI image before rotation with bins on the projection axis;
A display MTF measuring device comprising : When the reference axis is the y-axis, the axis perpendicular to the reference axis is the x-axis, a constant determining the spread of the peak fitting function is σ, and parameters are p ₁ and p ₂ , the peak fitting function is expressed as follows:

A function expressed as:
The display MTF measuring device according to claim 2, characterized in that the second line inclination detection means searches for _p1 and _p2 that most closely approximate the pixel distribution of the rotated ROI image on the xy coordinate system to the distribution of the function values of the function, and calculates arctan ( _p2 ) to determine the amount of deviation in the inclination angle of the line. 4. The display MTF measuring device according to claim 3 , wherein the constant σ is a value such that a full width at half maximum of the function is approximate to the number of pixels of a line width in the rotated ROI image. 1. A display MTF measurement device for measuring a display MTF representing a spatial frequency characteristic of a display from a measurement image, the measurement image being a straight line drawn on the display, captured by a camera, the display MTF measurement device comprising:
an image input means for inputting the captured image from the camera;
ROI image extraction means for extracting an ROI image of a set region of interest including the line from the captured image;
a line projection information generating means for generating line projection information by associating a position of each pixel of the ROI image with a projection axis of a bin at sub-pixel intervals along a gradient of the line;
a line spread function generating means for averaging pixel values of each pixel of the ROI image for each bin of the projection axis associated with the line projection information to generate a line spread function;
an MTF calculation means for calculating a measured MTF as an MTF measured in the captured image from the line spread function;
an MTF correction unit that calculates the display MTF by correcting the measured MTF using a spatial frequency ratio between the display and the camera and a camera MTF that is an MTF of the camera ;
The line projection information generating means
a first line inclination detection means for detecting an inclination angle of the line with respect to a predetermined reference axis by a Hough transform;
a second line inclination detection means for detecting, as a corrected inclination angle of the line, an inclination with respect to the reference axis of a peak position that is most similar between a distribution of pixel values of the ROI image and a distribution of function values obtained by applying a peak fitting function on an axis perpendicular to the reference axis in the direction of the reference axis, using an inclination angle of the line detected by the first line inclination detection means as an initial value;
a pixel position projection means for projecting pixels of an image obtained by rotating the ROI image at the corrected tilt angle so that the line is perpendicular to the projection axis onto the projection axis, and for generating the line projection information by associating the positions of each pixel of the original ROI image before rotation with a bin on the projection axis;
A display MTF measuring device comprising : 6. The display MTF measuring device according to claim 1, further comprising a measurement image display means for displaying the measurement image on the display. When a video camera that sequentially inputs images is used as the camera,
7. The display MTF measuring device according to claim 1, wherein the line projection information generating means generates the line projection information only once, initially, in the ROI image in which the region of interest is set. 8. The display MTF measuring device according to claim 1, wherein the MTF calculation means calculates the measured MTF by performing a Fourier transform on the line spread function, taking an absolute value for each spatial frequency, and normalizing the value by a constant component. A display MTF measurement program for causing a computer to function as the display MTF measurement device according to any one of claims 1 to 8 .

Images