CN101651845B - A method for testing the clarity of moving images of a display device - Google Patents

Abstract

The invention provides a method for testing definition of moving images of display devices, comprising the following steps: A: using a high speed camera to continuously shoot and record moving test strips of display elements or the display devices; B: carrying out moving tracking pretreatment, aligning the images of a plurality of continuous frames in a mode of unchanging relative position or alignment moving centre according to the moving tracking effect of human vision, and taking out brightness and chroma data of pixel coordinates with unchanging relative positions; C: carrying out time-domain filtering on the taken brightness and chroma data of the pixel coordinates with unchanging relative positions; and D: selecting a test image after time-domain filtering, taking out the brightness and chroma data of the pixel of the domain to be analyzed, and carrying out spatial domain filtering. The test method considers time-domain characteristics and spatial domain characteristics of human eyes simultaneously, the test result is more accurate, and the problem that evaluation parameters of definition for the moving images and static images are inconsistent is solved.

Description

A method for testing the clarity of moving images of a display device

technical field

The invention relates to the technical field of moving image definition testing, in particular to a method for testing the moving image definition of a display device.

Background technique

It has become an inevitable trend for the TV industry to transition from traditional analog TV to high-definition digital TV. DLP), organic photoelectric display (OLED), etc. Due to the difference in physical realization mechanism and driving mode, there are great differences in the quality of moving images. Consumers not only require excellent static clarity, but also require display devices to have good motion image display capabilities. Display related industries need moving image sharpness testing technology. At present, there are some testing technologies at home and abroad:

(1) Subjective test method

The test signal generator sends out the moving image definition stripe test signal, or the tailing response time test signal, and the threshold value is judged subjectively by the human eye, and the parameters such as the moving image definition TV line, or the tailing response time are read out. However, the problems of this method are: relying on subjective testing, not an objective testing method, the testing efficiency is not high, and the accuracy is limited, so it is difficult to apply in factory production lines and research institutes.

(2) Tracking camera method

The design idea of this method is based on the motion tracking effect of the eyeball, that is, when the human eye observes a low-speed moving object, the eyeball will track the moving object and keep focusing to observe clearly. Therefore, if you want to measure dynamic sharpness, you must simulate the two behaviors of human eye tracking and focusing. The technical scheme adopts a feedback control device to manipulate the camera to track the image position, and at the same time automatically adjusts the focus of the camera. High-speed camera is not required, and the exposure time should be slightly longer than the field refresh time. There are two methods for specific realization: rotary tracking and linear tracking. In the algorithm, the human eye contrast sensitivity function (CSF) is used to filter the spatial distribution of the test image, and finally the perceptible boundary blur width (PEBW) is obtained, and the perceptible boundary blur width (PEBW) is used as the evaluation parameter index of moving image clarity.

However, this method has the following problems:

(2.1) Perceptible boundary blur width (PEBW) is used as an evaluation parameter, which lacks connection with the "television line" of static definition, and it is difficult to compare static definition and dynamic definition;

(2.2) The realization of the system relies on high-precision tracking and automatic focusing system, which is very difficult and expensive;

(2.3) The human visual system simulation algorithm of this test method adopts the human eye contrast sensitivity function (CSF) model, which only considers the spatial characteristics of human vision and does not consider the temporal characteristics, so the human visual system The simulation is not comprehensive enough. The contrast sensitivity function (CSF) model of the human eye refers to the relationship between the contrast of the stripes subjectively felt by the human eye and the frequency of the spatial distribution of the stripes when the human eye watches the black and white stripes in the spatial distribution. Subjective experiments have shown that the human eye is very sensitive to black and white stripes of a certain frequency. As the frequency of black and white stripes is higher or lower than the most sensitive frequency, the contrast perceived by the human eye will gradually decrease. Therefore, this model belongs to human vision. Spatial properties do not take temporal properties into account.

(3) High Speed Camera

This method does not need to track the camera, but uses a high-speed camera to continuously record the image changes in a field, for example, record 8-10 high-speed camera frames of the moving image of the display in a field, and then use the "field integration" algorithm for processing : Align the frame position according to the principle of constant relative position, and then according to the principle of persistence of human vision, integrate and average the luminance data at the position where the relative position of the high-speed camera frame in a field is constant, and finally obtain the perceivable boundary blur Width (PEBW), and the perceivable boundary blur width (PEBW) is used as the evaluation parameter index of moving image sharpness.

However, this method has the following problems:

(3.1) Perceptible boundary blur width (PEBW) is used as an evaluation parameter, which lacks connection with the "television line" of static definition, and it is difficult to compare static definition and dynamic definition;

(3.2) The human visual system simulation algorithm of this test method uses the "field integral" algorithm. This model only considers the temporal characteristics of human vision, but does not consider the spatial characteristics, so the simulation of the human visual system is not comprehensive enough. Moreover, the "field integral" algorithm is not accurate enough for the simulation of the time characteristics of human vision, and its theoretical basis is the phenomenon of "persistence of vision of the human eye". But the essence of this phenomenon is the time-domain frequency response of the human eye, not the integral effect, so this algorithm also has certain problems.

(4) Motion Picture Response Time

The method takes the moving picture response time (MPRT) of the display as a parameter, and can test the moving picture response time (MPRT) when different gray levels change. The technical implementation uses the flickering squares between different gray levels within a certain period as the test pattern, and uses the transient photometer to measure the brightness response curve of the TV, thereby calculating the transient step response time, and further calculating the possible Perceived Border Blur Width (PEBW).

However, this method has the following problems:

(4.1) Evaluation parameter Motion Picture Response Time (MPRT) is the amount of time, although it can be converted into Perceivable Border Blur Width (PEBW) through some algorithm, but the conversion process will inevitably introduce errors, and it is different from the "TV line" of static definition. "There is still a lack of connection, and it is difficult to compare static sharpness with dynamic sharpness;

(4.2) The transient photometer tests the average brightness of an area, which cannot simulate or compensate the "movement tracking" characteristic of the human eye, so the test results lack the connection with the subjective evaluation results.

In short, the existing various testing methods have the following two problems:

(1) The human visual system model is not perfect enough. The human eye vision system simulation algorithm of the existing tracking camera test method adopts the human eye contrast sensitivity function (CSF) model, which only considers the spatial characteristics of human vision, and does not consider the temporal characteristics. However, the human vision system simulation algorithm of the high-speed camera test method adopts the "field integral" algorithm. This model only considers the temporal characteristics of human vision, and does not consider the spatial characteristics. Moreover, the "field integral" algorithm is not accurate enough for the simulation of the time characteristics of human vision, and its theoretical basis is the phenomenon of "persistence of vision of the human eye". However, the essence of this phenomenon is the time-domain frequency response of the human eye, not the integral effect, so there are certain problems in the algorithm's simulation of the time-domain characteristics of the human eye.

(2) The evaluation parameters of moving image sharpness are inconsistent with the evaluation parameters of static image sharpness. The existing tracking camera test methods and high-speed camera test methods both use the perceptible boundary blur width (PEBW) as the evaluation parameter index of moving image sharpness, which is inconsistent with the static image sharpness parameters, and it is difficult to distinguish between moving image and static image sharpness. compare between.

Contents of the invention

The object of the present invention is to provide a method for testing the clarity of moving images of display equipment, which can simultaneously consider the characteristics of time domain and space domain of the human eye, and the test results are more accurate; moreover, it solves the problem of evaluating the clarity of moving images and static images Inconsistent parameters.

The method for testing the moving image clarity of a display device of the present invention comprises the following steps:

Step A: The high-speed camera continuously shoots and records the motion test stripes of the display device or display device;

Step B: Carry out motion tracking preprocessing, according to the motion tracking effect of human vision, align the images of multiple consecutive frames according to the relative position unchanged or in the way of aligning the motion center, and take out the brightness of the pixel coordinates with the relative position unchanged and colorimetric data;

Step C: performing time-domain filtering on the extracted luminance and chrominance data of pixel coordinates whose relative positions are unchanged;

Step D: Select a test picture after time-domain filtering, extract the luminance and chrominance data of the pixels in the area to be analyzed, and perform spatial domain filtering;

Step E: Judging whether a TV line can be subjectively identified according to the threshold value of the modulation degree; wherein, the modulation degree refers to the ratio of the brightness difference between the black and white stripes to the sum of the brightness of the black and white stripes; a certain TV line of the filtered image corresponds to If the modulation degree is greater than 5%, the TV line is considered to be identifiable, otherwise it is not identifiable, and the boundary position between identifiable and unrecognizable is the TV line of the moving image definition at this speed.

Wherein, the following steps are further included before said step A:

The video signal generator generates black and white test stripes with a certain moving speed, stripe spacing and thickness, and drives the display device under test to correctly display the moving image.

Wherein, in the step C, the time-domain filtering is a processing method using a low-pass filter; in the step D, the spatial domain filtering is a processing method using a band-pass filter.

Wherein, said step E includes the following steps:

Step E1: taking out the luminance and chrominance data of a certain TV line in the image processed in step D to form an array R2(0, j);

Step E2: Find all the maximum and minimum points of the brightness distribution of the TV line to be analyzed with the position. If the test pattern is composed of m black stripes and n white stripes alternately, there should be m minimum and minimum values n maximum values;

Step E3: According to the brightness data of these maximum and minimum points, the modulation degree is calculated by the formula, there should be m to n modulation degrees, and the worst value of these modulation degrees is taken as the modulation degree of the TV line;

Step E4: Repeat steps E1 to E3 to calculate the modulation degrees of all the sampling points that need to be tested. Sampling according to a certain analysis interval;

Step E5: Find the TV line whose modulation degree is closest to 5%, which is the calculation result.

运动图像的调制度为：其中，V₀ ⁰为静止测试图中的黑条纹对应的亮度值；V₁ ⁰为静止测试图中的白条纹对应的亮度值；V₀ ¹为运动测试图中的黑条纹对应的亮度值；V₁ ¹为运动测试图中的白条纹对应的亮度值。Among them, the modulation degree of the still image is:

The modulation degree of the moving image is: Wherein, V ₀ ⁰ is the brightness value corresponding to the black stripes in the static test chart; V ₁ ⁰ is the brightness value corresponding to the white stripes in the static test chart; V ₀ ¹ is the brightness value corresponding to the black stripes in the motion test chart; V ₁ ¹ is the brightness value corresponding to the white stripes in the motion test image.

Wherein, the evaluation parameter of the moving image sharpness measurement result is an evaluation parameter that can be directly compared with the static image sharpness, including at least one of "line", "line pair", "line pair/screen size" or "television line" .

Among them, when the test stripes are stationary, the brightness changes of the pixels in the area to be analyzed in the moving direction are distributed according to a sine wave:

$\mathrm{<span\; class="notranslate">\; G</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; cos</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}<span\; class="notranslate">\; \&CenterDot;</span>\frac{{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="0"\; label="W\; L\; W"\; filenames="DEST\_PATH\_HA20182933200910091377101E00011.png,DEST\_PATH\_HA20182933200910091377101E00012.png"\; state="\{\{state\}\}">W</figure-callout></span><figure-callout\; id="0"\; label="W\; L\; W"\; filenames="DEST\_PATH\_HA20182933200910091377101E00011.png,DEST\_PATH\_HA20182933200910091377101E00012.png"\; state="\{\{state\}\}">\; </figure-callout>}}_{\mathrm{<span\; class="notranslate">\; L</span>}}}{{}_{}}<span\; class="notranslate">\; \&CenterDot;</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&CenterDot;</span>\mathrm{<span\; class="notranslate">\; 255</span>}$

Among them, G(j) is the brightness value of the jth pixel in the area to be analyzed along the motion direction on the screen; W ₀ is the physical resolution in this direction; W _L is the TV line corresponding to the black and white stripes; the test stripes are When v pixel/field is moving, if the brightness of the pixel at position j is G(j) in a certain field, then the brightness of the pixel at position j+v is also G(j) in the next field.

The beneficial effects of the present invention are: according to the method for testing the clarity of moving images of display devices of the present invention, a more accurate model of the human visual system is provided. A more accurate "time-domain filtering" algorithm to replace the existing "field integration" algorithm; in addition, the existing test method's moving image definition evaluation parameters are inconsistent with the static image definition evaluation parameters, the present invention adopts the "modulation degree "Threshold algorithm obtains dynamic image clarity, takes "television line" as a parameter, which is consistent with static image clarity parameters, and can be compared between moving image and static image clarity.

Description of drawings

FIG. 1 is a schematic diagram of the time-domain frequency response of the human eye;

Figure 2 is a schematic diagram of the frequency response in the airspace of the human eye;

Fig. 3 is a schematic diagram of comparison results of human visual system simulation algorithms for LCD TVs;

Fig. 4 is a schematic diagram of comparison results of human visual system simulation algorithm of plasma TV;

Fig. 5 is a schematic diagram of the definition (right) of the moving image sharpness test stripes (left) and the degree of modulation;

Fig. 6 is a schematic diagram of horizontal brightness distribution of 720 TV lines of an LCD TV;

Fig. 7 is a schematic diagram of horizontal brightness distribution of 700 TV lines of an LCD TV;

Fig. 8 is a schematic diagram of comparison results between the subjective evaluation results of LCD TVs and the modulation degree algorithm;

Fig. 9 is a schematic diagram of a comparison result between the subjective evaluation result of the plasma TV and the modulation degree algorithm.

Detailed ways

Hereinafter, the method for testing the clarity of a moving image of a display device of the present invention will be described in detail with reference to FIGS. 1 to 9 .

The method for testing the moving image clarity of a display device of the present invention comprises the following steps:

Step 1: A video test sequence occurs. The video signal generator generates black and white test stripes with a certain moving speed, stripe spacing and thickness, that is, different definition TV lines, and drives the display device under test to correctly display the moving image.

Step 2: The high-speed camera shoots and records continuously. The displayed moving image is captured with a high-speed camera and recorded.

Step 3: "Motion Tracking" preprocessing. According to the "motion tracking" effect of human vision, the images of several consecutive frames (such as 64) are aligned according to the relative position invariant or the motion center alignment, and the brightness and chrominance data of the pixel coordinates with relative position invariant are taken out .

Wherein, in step 3, if the definition of moving images in the horizontal direction is tested, W ₀ is 1920 for China's high-definition television standard. W _L can take 700, 720, 740 and other TV lines.

Step 4: The time-domain filtering part in the correction of the human visual system model. Time-domain filtering is performed on the extracted luminance and chrominance data of pixel coordinates whose relative positions are unchanged.

Step 5: The spatial domain filtering part in the correction of the human visual system model. A test picture after time-domain filtering is selected, and the luminance and chrominance data of the pixels in the area to be analyzed are taken out to perform spatial domain filtering.

Step 6: Calculation of TV line by threshold value of modulation degree. The modulation degree is calculated for each sharpness stripe of the image corrected by the human visual system model, and the TV line that reaches the threshold of the modulation degree is the result of the moving image sharpness test, and the parameter is "TV line".

Wherein, steps 3 to 5 constitute the human visual system model simulation algorithm of the present invention, that is, the algorithm consists of three parts: preprocessing, time domain filtering and spatial domain filtering. The execution sequence of these three parts constitutes two implementation examples of this algorithm. An implementation example is performed in the order of preprocessing, time domain filtering and spatial domain filtering; another implementation example is performed in the order of preprocessing, spatial domain filtering and time domain filtering.

In the following, an implementation example performed in the order of preprocessing, time domain filtering and spatial domain filtering will be introduced in detail:

Step 31: Preprocessing.

Before the temporal filtering algorithm, the data needs to be preprocessed. According to the "motion tracking" effect of human vision, the images of several consecutive frames (for example, 64) need to be aligned in such a way that the relative position remains unchanged or the center of motion is aligned, and taken out The luminance and chrominance data of pixel coordinates with relative position unchanged, and these data are composed into an array.

For example, the luminance value of the pixel to be analyzed in the first frame of the consecutive frames is r(x, y), where x and y represent the horizontal and vertical coordinates in the pixel image, respectively. The movement speed of the center of the image is v pixels/high-speed camera frame, and the movement in the horizontal direction means that the center points of the continuous frames of these high-speed cameras move v pixels in one direction per frame. Considering the convenience of calculation, v can be rounded to an integer. The luminance data at the points where the relative pixel positions of these consecutive frames are constant are formed into an array R(i, j), for example, when 64 analysis frames are taken:

R(0,j)=r0(x,y), r0(x,y) is the first graph.

R(1,j)=r1(x,y+v), r1(x,y+v) is the second graph.

R(2, j)=r2(x, y+2*v), r2(x, y+2*v) is the third picture.

R(3, j)=r3(x, y+3*v), r3(x, y+3*v) is the fourth picture.

R(63, j)=r64(x, y+63*v), r64(x, y+63*v) is the 64th picture.

Step 32: Time domain filtering.

Only considering the test of the definition of moving images in the horizontal direction, the preprocessed array R(i, j) performs time-domain filtering according to the following formula, and two implementation examples can be time-domain and frequency-domain:

Implementation example in time-frequency domain: first transform pixel data from Fourier to frequency domain, multiply by filter function, and then inverse Fourier transform back to time domain to obtain pixel array R1(i, j) after time domain filtering.

R ₁ (i,j)=F ^-1 (F(R(i,j)) T(j))

In the above formula, F is the Fourier transform, F ^-1 is the inverse Fourier transform, and T is the time domain filter function.

Time-domain implementation example: Convolute pixel data with a time-domain filter function for direct filtering.

${\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&CircleTimes;</span>\mathrm{<span\; class="notranslate">\; W</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span>$

In the above formula, W is the time domain filter function.

Step 33: Spatial filtering.

Take an image after time-domain filtering and perform spatial filtering. For example, the array R1(0, j) after time-domain filtering is selected to perform spatial filtering according to the following formula. There are two implementation examples: spatial domain and spatial frequency domain:

Example of implementation in the spatial frequency domain: first transform the pixel data from the Fourier transform to the frequency domain, multiply by the filter function, and then inverse Fourier transform back to the spatial domain to obtain the pixel array R2(0, j) after the spatial domain filtering.

R ₂ (0, j) = F ^-1 (F(R ₁ (0, j))·S(j))

In the above formula, F is the Fourier transform, F ^-1 is the inverse Fourier transform, and S is the spatial filter function.

Example of spatial domain implementation: Convolute pixel data with spatial domain filter function and filter directly.

${\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&CircleTimes;</span>\mathrm{<span\; class="notranslate">\; P</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span>$

In the above formula, P is the spatial filtering function.

In addition, in the above step 6, the threshold value of the modulation degree is used to calculate the TV line, and the specific execution steps are as follows:

Step 61: Take out the luminance and chrominance data of a TV line (for example, 700 lines) in the image processed by the human visual system model, and form an array R ₂ (0, j).

Step 62: Find all the maximum and minimum points of the luminance distribution of the TV line to be analyzed with the position. If the test pattern consists of m black stripes and n white stripes alternately, there should be m minima and n maxima.

Step 63: According to the luminance data of these maximum and minimum points, the modulation degree is calculated by the formula mentioned above. There should be m to n modulation degrees, and the worst value of these modulation degrees is taken as the TV line degree of modulation.

Step 64: Repeat steps 61 to 63 to calculate the modulation degrees of all the sampling points that need to be tested. Sampling is performed at certain analysis intervals (eg, 700, 720, 740 TV lines).

Step 65: Find the TV line whose modulation degree is closest to 5%, which is the calculation result.

The method for testing the clarity of the moving image of the display device of the present invention is realized by means of a video generator, a high-speed camera, a human visual system model simulation algorithm and a "TV line" calculation method based on a modulation degree threshold.

When the video signal generator sends the test signal to the TV to be tested, the TV will display the moving sharpness test chart, and the video signal can generate a moving image of 0-6 pixels/field. At this time, you can use a high-speed camera to take continuous shots and record the pictures of the shooting results. The shooting speed of the high-speed camera is 500 frames per second, RGB color mode, and the resolution is 1280×1024. The output interface of the video signal generator is YPbPr component signal, and the resolution is 1920×1080i, 50Hz. Test image resolution ranges from 400 TV lines to 900 TV lines. The high-speed camera continuously records 64 images at a time, corresponding to 6.4 TV images, and the algorithm is realized by C++ software. Use this system to test a liquid crystal TV (LCD) and a plasma TV (PDP) respectively.

Considering that the present invention is mainly embodied in the simulation algorithm of the human visual system model decoupled from the time domain and the space domain and the "TV line" discrimination algorithm based on the modulation degree, the comparison is carried out in two processes, and these two points are tested separately, and the mutual interference is eliminated as much as possible. the influence between.

1. Firstly, the simulation algorithm of the human visual system model proposed by the present invention decoupled from time domain and space domain is compared with the original framing, field integration and direct subjective evaluation algorithms.

Since the evaluation parameter obtained by the original field integration algorithm is the perceivable boundary blur width (PEBW), which is different from the "television line" parameter system of the present invention, in order to make the results comparable, we use a subjective method in the last step of the comparison. Evaluation, that is, the final image obtained by the field integration algorithm and the final image obtained by the time-domain filtering algorithm proposed by the present invention are subjectively evaluated to obtain the "television line" index of moving image clarity, which is directly compared with the human eye on the motion on the TV. Compared with the "TV line" obtained from the subjective evaluation of the sharpness test chart, the closer to the direct subjective evaluation result, the more correct the algorithm is for the simulation and modeling of the human visual system. As a comparison reference, we also provide the original score The results of the subjective evaluation after frame image averaging to confirm the role of the human visual system model in the final image processing.

2. Compare and verify the "TV line" discrimination algorithm based on the modulation degree of the present invention.

The two sides of the comparison are the "TV line" calculated by the present invention using the modulation algorithm and the "TV line" directly obtained by subjective evaluation of the motion definition test chart on the TV by human eyes. Because the evaluation parameters of algorithms such as "field integration" and "high-speed/tracking camera" mentioned above are all perceptible boundary blur width (PEBW), they are different from the evaluation parameters of "television line" proposed by the present invention, because those algorithms do not have A method of calculating "television line" from the image corrected by the human visual system model is proposed, so it cannot participate in the comparison. The comparison method and results of the two parts are as follows:

(1) Human visual system model simulation algorithm.

Participating in the comparison are the original framing, field integration, direct subjective evaluation and the time domain filtering algorithm proposed by the present invention.

(1.1) Original framing. Subjectively evaluate the 64 images taken continuously by the high-speed camera one by one, and average the obtained "TV line" results.

(1.2) Field integral. Align the 10 consecutive images out of the 64 images continuously taken by the high-speed camera in such a way that the relative position remains unchanged or the center of motion is aligned, and the corresponding pixels are integrated and averaged, and the processed images obtained by subjective evaluation are " TV line".

(1.3) Direct subjectivity. The "TV line" obtained by the subjective evaluation of the human eye directly on the motion definition test chart displayed on the TV.

(1.4) Algorithm of the present invention. Algorithm for human visual system model decoupling in time domain and space domain. Take the 64 images continuously taken by the high-speed camera, take out the luminance and chrominance data of the pixel to be analyzed according to the principle of constant relative position, and then follow the time-domain frequency response characteristics and space-domain response of the human eye as shown in Figure 1 and Figure 2 respectively Characteristic filtering, and then extract a processed image, and perform subjective evaluation to obtain the "TV line".

Figure 3 shows the comparison results of the human visual system model simulation algorithm implemented on a certain LCD TV test results, and Figure 4 shows the comparison results after it is implemented on a certain plasma TV test results.

Comparison results: the algorithm of the decoupled human visual system model proposed by the present invention, which considers the characteristics of time domain and space domain, is closer to the result of direct subjective evaluation than other algorithms such as "field integration", so the simulation proposed by the present invention The algorithm is closest to the real human visual system.

Result analysis: the reason that the algorithm proposed by the present invention is closer to the real visual system of the human eye has the following two aspects: one is that under the condition of computing power permission, the algorithm proposed by the present invention can select a relatively long analysis cycle, such as this paper takes 64 A high-speed camera sampling period corresponds to 6 TV field periods. Compared with the "field integration method" which has only one TV field analysis period, it can effectively avoid the "flicker", "block" and other phenomena "missing" problems caused by the "field integration method" due to the short analysis period, so the analysis results More comprehensive. Another reason is that the algorithm proposed by the present invention takes into account the characteristics of the human eye in time domain and space domain, and uses "time domain filtering" to replace "field integration" and "persistence of vision" in the simulation process of time domain characteristics. It is the limitation of time-domain frequency response bandwidth, which is time-domain filtering in essence, and "field integration" is only a phenomenon. Therefore, the algorithm of the present invention is closer to the essence of the human visual system, and thus closer to the subjective evaluation result.

(2) Calculation method of "television line" based on threshold value of modulation degree.

Existing high-speed cameras, tracking cameras and other moving image sharpness test methods all use perceptible boundary blur width (PEBW) as a parameter, and an inventive point of the test method proposed by the present invention is to use "television line" as a parameter system, Therefore, it is not possible to directly compare the advantages and disadvantages of the algorithm with high-speed cameras and tracking cameras. However, it can be compared with the direct subjective evaluation results to evaluate whether the moving image definition "TV line" calculated by this algorithm is related to the direct subjective evaluation results. If there is correlation, the size of the test error can be evaluated, and the potential user of this algorithm can decide whether it can be used as an objective calculation method of the moving image definition "TV line" according to the size of the error.

运动图像的调制度为：

其中，V₀ ⁰为静止测试图中的黑条纹对应的亮度值；V₁ ⁰为静止测试图中的白条纹对应的亮度值；V₀ ¹为运动测试图中的黑条纹对应的亮度值；V₁ ¹为运动测试图中的白条纹对应的亮度值。The modulation degree M refers to the ratio of the brightness difference of the black and white stripes to the brightness sum of the black and white stripes, as shown in Fig. 5 . According to the illustration, the modulation degree of a still image is:

The modulation degree of the moving image is:

Wherein, V ₀ ⁰ is the brightness value corresponding to the black stripes in the static test chart; V ₁ ⁰ is the brightness value corresponding to the white stripes in the static test chart; V ₀ ¹ is the brightness value corresponding to the black stripes in the motion test chart; V ₁ ¹ is the brightness value corresponding to the white stripes in the motion test image.

The modulation degree discrimination threshold is set at 5%, that is, if the modulation degree corresponding to a certain TV line of the image processed by the temporal filtering algorithm of the human visual system model in the previous step is greater than 5%, then this TV line is considered to be acceptable. Recognizable, otherwise it is unrecognizable, at the boundary position between recognizable and unrecognizable, it is the "TV line" of moving image clarity at this speed of motion. Figures 6 and 7 show the horizontal luminance distribution at 720 TV lines and 740 TV lines when an LCD TV is displaying an image with 1 pixel/field motion. It is calculated that the modulation degree of 700 lines is 7.3%, and the modulation degree of 720 lines is 2.9%, then the definition of the LCD TV when displaying 1 pixel/field moving image is 700 TV lines.

Figure 8 shows the comparison results of the "TV line" calculation method based on the threshold value of the modulation degree implemented on a certain LCD TV test results, and Figure 9 shows the comparison results after it is implemented on a certain plasma TV test results.

Comparison results: The TV line discrimination algorithm based on the "modulation degree" threshold proposed by the present invention has correlation with the subjective evaluation algorithm, and can be used to calculate the moving image definition "TV line" within a certain error range, which is an effective Calculation method.

In summary, according to the method for testing the clarity of moving images of display devices of the present invention, a more accurate model of the human visual system is provided. The "time-domain filtering" algorithm to replace the existing "field integration" algorithm; in addition, the current test method's moving image clarity evaluation parameters are inconsistent with the static image definition evaluation parameters, the present invention adopts the "modulation degree" threshold The algorithm obtains the sharpness of the dynamic image, and takes "television line" as a parameter, which is consistent with the parameter of the sharpness of the static image, and can be compared between the sharpness of the moving image and the static image.

The above is a detailed description of the present invention for those skilled in the art to understand the present invention, but it is conceivable that other changes and modifications can be made without departing from the scope covered by the claims of the present invention. Variations and modifications are within the scope of the present invention.

Claims (7)

1. a method for testing definition of moving images of display devices is characterized in that, comprises the following steps:

Steps A: the exercise test striped of display device or display device is taken and noted to high speed camera continuously;

Step B: carry out the motion tracking preliminary treatment, motion tracking effect according to human eye vision, the image of a plurality of successive frames is constant or align in the mode of centre of motion alignment according to relative position, and take out the brightness and the chroma data of the constant pixel coordinate of relative position;

Step C: the brightness and the chroma data of the pixel coordinate that the relative position of described taking-up is constant carry out time-domain filtering;

Step D: choose the width of cloth test picture behind the time-domain filtering, take out the brightness and the chroma data of the pixel in zone to be analyzed, carry out spatial domain filtering;

Step e: judge according to threshold modulation whether certain tv line can be discerned by subjectivity; Wherein, modulation degree is meant the ratio of luminance difference with the black and white strip brightness sum of black and white strip; The modulation degree of certain tv line correspondence of process filtering processed images is greater than 5%, think that then this tv line can be identified, otherwise can not be identified, can discern and unrecognizable boundary position, be exactly the tv line of the definition of moving images under this movement velocity.

2. method for testing definition of moving images of display devices as claimed in claim 1 is characterized in that, further comprises the following steps: before described steps A

Produce chequered with black and white test striped certain movement speed, that fringe spacing and thickness do not wait by flying-spot video generator, and drive tested display device and correctly show this moving image.

3. method for testing definition of moving images of display devices as claimed in claim 2 is characterized in that, in described step C, described time-domain filtering is the processing method that adopts low pass filter filtering; In described step D, described spatial domain filtering is the processing method that adopts band-pass filter.

4. method for testing definition of moving images of display devices as claimed in claim 1 is characterized in that described step e comprises the following steps:

Step e 1: brightness and the chroma data of described step D being handled a certain tv line in the image of back take out, and form array;

Step e 2: find all maximum points and the minimum point of tv line to be analyzed,, m minimum and n maximum should be arranged then if test pattern alternately is made of m black streaking and n informal voucher line with the brightness of position distribution;

Step e 3: according to the brightness data of these maximum value minimum points, calculate modulation degree, should have m, get worst-case value in these modulation degree as the modulation degree of this tv line to n modulation degree by formula;

Step e 4: repeating step E1 is to the modulation degree of step e 3 all sampling points that need test of calculating, and sampling is according to certain analysis spacing;

Step e 5: finding modulation degree near 5% tv line, is exactly result of calculation.

5. method for testing definition of moving images of display devices as claimed in claim 4 is characterized in that, the modulation degree of rest image is:

The modulation degree of moving image is:

Wherein, V ₀ ⁰Brightness value for the black streaking correspondence in the static resolution chart; V ₁ ⁰Brightness value for the informal voucher line correspondence in the static resolution chart; V ₀ ¹Brightness value for the black streaking correspondence among the exercise test figure; V ₁ ¹Brightness value for the informal voucher line correspondence among the exercise test figure.

6. method for testing definition of moving images of display devices as claimed in claim 4, it is characterized in that, the evaluating of definition of moving images measurement result is the evaluating that can directly compare with the still image definition, comprises in " line ", " line to ", " line right/screen size " or " tv line " at least a.

7. method for testing definition of moving images of display devices as claimed in claim 4 is characterized in that, when the test striped was static, the brightness of the pixel in zone to be analyzed on this direction of motion changed by sinusoidal wave distribution:

Wherein, G (j) is this direction of motion of screen upper edge, in the brightness number of j the pixel in zone to be analyzed; W ₀It is this direction physical resolution; W _LIt is the tv line of this black and white strip correspondence; When the test striped moves with v pixel/field, if the brightness of the pixel of position j is G (j) certain the time, then next the time, the brightness of position j+v place pixel also is G (j).

Images