Device and method for automatically measuring delay of image processing system

Abstract

The invention provides an automatic measuring device and method for delay of an image processing system, a specific identifier is inserted at a set position of a generated test image and output to a system to be tested, the output image of the system to be tested is collected, the corresponding specific identifier is extracted and judged at the position, and the sending and receiving time of the specific identifier is calculated, so that the delay of the system to be tested is measured. The delay time of the to-be-measured system is obtained according to the delay error which can be accurately metered in the input and output process; the whole process is completely and automatically completed by a real-time processor, the calculation precision can reach the nanosecond level, the personnel operation difficulty in a traditional delay measurement method is reduced, the construction complexity of a traditional measurement device is simplified, the difference interference of instruments and devices and the operation error of people in a loop are effectively reduced, and the measurement accuracy is improved. Therefore, the delay time of the image processing system can be accurately and automatically measured.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and specifically to an automatic measurement device and method for delay of an image processing system, which is applicable to an image processing system with a video input and video output physical interface, and can accurately and automatically measure the full-link delay time of the system. Background Art

[0002] The full-link delay of the image processing system mainly includes image acquisition, data processing and imaging output, which also includes time overhead such as data caching, multi-threaded collaboration and task processing. Therefore, image processing systems with real-time requirements need to test and quantify the delay time. Especially for high-performance image processing systems, the measurement of delay time has reached the nanosecond level.

[0003] There are three main methods for measuring link latency in traditional image processing:

[0004] The first method is to generate a time-delay image sequence after transmitting the source image sequence, and then use a frame of the source image as the image to be tested, match the best image in the time-delay image, collect the acquisition time of the two, and calculate the difference as the delay time. In order to make the result more accurate, multiple measurements are taken to calculate the average value as the final result;

[0005] The second method uses a digital or disk pointer timer as the image input. After passing through the system to be tested, the delayed timer image is displayed in real time on the display. A high-speed camera is used to capture the timer at the actual input end and the delayed timer image displayed on the display multiple times in the same frame. After observing and comparing the time difference between the two timers in the same picture, the delay time of the image processing system is calculated.

[0006] The third method uses a device with a source clock to generate a clock image and input it into the system under test. The clock image and the output clock image of the system under test are then simultaneously acquired through a computer with an acquisition card. The delay time of the image processing system is calculated by observing and comparing the time difference between the two timers in the same screen.

[0007] The first method requires matching the best image to the image to be tested. The accuracy of the matching result depends on the matching algorithm. The matching result may have errors, resulting in errors in the final delay difference. When measuring the delay, the second method is prone to blurring the counter value or pointer position due to the performance of the shooting device. It is necessary to obtain multiple sets of shooting results and average them to obtain the delay value. At the same time, this method cannot accurately calculate the delay of the system to be tested because the delay time of the image input and the display screen is included in the image processing system. The third method uses a computer with an acquisition card to simultaneously obtain the clock image and the output clock image of the system to be tested. Although it solves the problem of blurry clock value display, in order to obtain the clock image and the output image of the system to be tested in the same picture, the acquisition delay of the measuring device is introduced, and there is a defect of insufficient measurement accuracy.

[0008] The key problem with the above three methods is that the delay error introduced into the test equipment is uncontrollable, resulting in low delay measurement accuracy and difficulty in meeting the high-precision measurement requirements at the nanosecond level. The main reason for the uncontrollable delay error is that the existing methods mainly use a device structure consisting of CPU + memory + graphics card + acquisition card, and the internal data processing method is that the CPU generates a test image, which is transmitted to the graphics card through the memory, and the graphics card outputs it to the device under test; the test image processed by the device under test is captured by the acquisition card and output to the memory, and then transmitted to the CPU or graphics card by the memory; in the process of generating the test image, the time for the CPU to access the memory and the data interaction time between the memory and the graphics card are difficult to measure accurately; in the process of receiving the test image, the data interaction time between the acquisition card and the memory and the time for the CPU to access the memory are difficult to measure accurately. Summary of the invention

[0009] In view of the problems existing in the prior art, the present invention proposes an automatic measurement device and method for the delay of an image processing system. A specific mark is inserted at a specific position of a generated test image and output to the system to be tested. The same specific mark is extracted at the position by collecting the output image of the system to be tested, and the time of sending and receiving the specific mark is calculated. The delay error of the automatic measurement device can be accurately obtained, and the delay measurement and calculation process of the system to be tested is completely automatically completed by a real-time processor, and the calculation accuracy can reach nanosecond level.

[0010] The technical solution of the present invention is:

[0011] The automatic measuring device for delay of an image processing system comprises a test image generating module, an identification encoding module, an identification decoding module, a counter module, a first interface module, and a second interface module; the automatic measuring device uses the first interface module as an output interface and the second interface module as an input interface, and is connected to the system to be tested via a video cable;

[0012] The test image generation module generates a test image and outputs it to the identification coding module. The identification coding module sets a specific identification in the specified coordinate area of ​​the test image according to the setting of the host computer, and generates a delayed counting trigger signal to the counter module while setting the specific identification, and the counter module starts counting; the identification coding module outputs the test image with the specific identification to the first interface module; the first interface module outputs the test image with the specific identification to the system to be tested;

[0013] After the test image is processed in the system to be tested, it is output from the system to be tested to the second interface module; the second interface module performs protocol parsing on the test image processed by the system to be tested, and outputs the parsed test image to the identification decoding module. The identification decoding module determines whether it is the specific identification in the specified coordinate area of ​​the received image according to the settings of the upper computer. If it is a specific identification, it generates a delayed counting stop signal to the counter module, and the counter module stops counting.

[0014] Furthermore, the test image generating module generates a test image signal having the same input resolution, timing, reference clock and frame rate according to the input resolution, timing, reference clock and frame rate parameters of the system to be tested.

[0015] Furthermore, the frequency of the counter module is higher than a set required value, and the set required value is determined according to a delay measurement accuracy requirement.

[0016] Furthermore, the specific identifier is a frame number.

[0017] The present invention also proposes a method for automatically measuring the delay of an image processing system using the above device:

[0018] Obtaining the delay error from the time when the test image generation module generates the test image to the time when the first interface module outputs the test image with the specific identification to the system to be tested, which is composed of two parts of delay errors that can be accurately measured; one part of the delay error t1 is the delay error from the time when the identification encoding module sets the specific identification to the time when the test image is output to the first interface module, and the delay error t1 is accurately determined according to the range of the specified coordinate area; the other part of the delay error t2 is the delay error generated by the interface protocol conversion process of the first interface module;

[0019] Obtain the delay error from the time when the system under test outputs the processed test image to the second interface module to the time when the counter module stops counting, which is composed of two parts of delay errors that can be accurately measured; one part of the delay error t3 is the delay error generated by the interface protocol conversion process of the second interface module; the other part of the delay error t4 is the time required for the identification decoding module to determine the specified coordinate area, and the delay error t4 is accurately determined according to the range of the specified coordinate area;

[0020] Obtaining a count value of the counter, and converting the count value to obtain an initial delay time T1;

[0021] The final image processing system delay is T = T1-Δti, i = 1, 2, 3, 4.

[0022] Furthermore, within a set time period, different specific identifiers are loaded in test images of different frames. Each specific identifier can be used to obtain the delay time of the current frame test image passing through the system to be tested. The delay times obtained using different specific identifiers are counted to obtain the delay time distribution of the system to be tested within the set time period.

[0023] Beneficial Effects

[0024] The present invention proposes an automatic measurement device and method for the delay of an image processing system. A specific mark is inserted at a set position of a generated test image and output to the system to be tested. By collecting the output image of the system to be tested, the corresponding specific mark is extracted and judged at the position, and the time of sending and receiving the specific mark is calculated. Based on the accurately measurable delay error in the input and output process, the delay time of the system to be tested is obtained. The whole process is completely completed automatically by a real-time processor, and the calculation accuracy can reach nanosecond level, which reduces the difficulty of human operation in traditional delay measurement methods, simplifies the complexity of setting up traditional measurement devices, and effectively reduces the difference interference of instruments and devices and the operation error of people in the loop, so that the delay time of the image processing system can be accurately and automatically measured.

[0025] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0027] Figure 1 is a connection diagram for delay measurement in an embodiment of the present invention;

[0028] Figure 2 is a flowchart of delay measurement in an embodiment of the present invention;

[0029] Figure 3 It is a schematic diagram of the human-machine control interface of the measuring device of the present invention. DETAILED DESCRIPTION

[0030] Embodiments of the present invention are described in detail below. The embodiments are exemplary and intended to be used to explain the present invention, but should not be construed as limiting the present invention.

[0031] The present embodiment proposes an automatic measurement device for delay of an image processing system, comprising a test image generating module, an identification encoding module, an identification decoding module, a counter module, a first interface module, and a second interface module; the automatic measurement device uses the first interface module as an output interface and is connected to a video input interface of a system to be tested via a video cable; the automatic measurement device uses the second interface module as an input interface and is connected to a video output interface of the system to be tested via a video cable.

[0032] The test image generation module generates a test image and outputs it to the identification encoding module, wherein the test image generation module generates a test image signal with the same input resolution, timing, reference clock, and frame rate according to the input resolution, timing, reference clock, and frame rate parameters of the system to be tested. In the specific implementation of this embodiment, according to the input and output video parameters of the system to be tested and the connection number of the video input and output interface, through the human-computer interaction interface (such as Figure 3 After loading the test image and completing the configuration, click the "Start" button on the human-computer interaction interface to observe whether the system under test can normally capture video and whether the automatic measurement device of this embodiment can normally display the output video of the system under test.

[0033] The identification coding module sets a specific identification in the specified coordinate area of ​​the test image according to the setting of the host computer. In this embodiment, the specific identification adopts the frame number. The identification coding module outputs the test image with the frame number to the first interface module; the first interface module outputs the test image with the frame number to the system under test.

[0034] The identification encoding module generates a delay count trigger signal to the counter module while setting the specific identification, and the counter module starts counting. In order to meet the accuracy requirements, the frequency of the counter module is higher than the set requirement value, and the set requirement value is determined according to the delay measurement accuracy requirements. In this embodiment, the frequency of the counter module is higher than 100Mhz. Under 100Mhz, the output delay error between the automatic measuring device and the system to be measured does not exceed 2 clock cycles, that is, the output delay error is not greater than 20ns.

[0035] The above-mentioned process from the test image generation module generating the test image to the first interface module outputting the test image with a specific identification to the system to be tested includes two parts of delay errors, and these two parts of delay errors can be accurately measured; one part of the delay error t1 is the delay error from the identification encoding module setting the specific identification to the output of the test image to the first interface module. This part of the delay error is accurately determined according to the range of the specified coordinate area. If the specified coordinate area consists of 2 pixel points, the delay error is 2 clocks; the other part of the delay error t2 is the delay error generated by the interface protocol conversion process of the first interface module, and this part of the error is well known in the industry.

[0036] After the test image is processed in the system to be tested, it is output from the system to be tested to the second interface module; the second interface module performs protocol parsing on the test image processed by the system to be tested, and outputs the parsed test image to the identification decoding module. The identification decoding module determines whether it is the specific identification in the specified coordinate area of ​​the received image according to the settings of the upper computer. If it is a specific identification, it generates a delayed counting stop signal to the counter module, and the counter module stops counting.

[0037] The above process from the system under test outputting the processed test image to the second interface module to the counter module stopping counting also includes two parts of delay errors, and these two parts of delay errors can be accurately measured; one part of the delay error t3 is the delay error caused by the interface protocol conversion process of the second interface module, and this part of the error is well known in the industry; the other part of the delay error t4 is the time required for the identification decoding module to determine the specified coordinate area, and this time is accurately determined according to the range of the specified coordinate area. If the specified coordinate area consists of 2 pixel points, the delay error is 2 clocks.

[0038] In this way, the initial delay time T1 is converted according to the count value of the counter, and the final high-precision delay time T = T1-Δti, i = 1, 2, 3, 4. For example, the count value C of the counter is extracted, and the frequency F set by the reference clock is automatically calculated. The initial delay time T1 = C / F is automatically calculated. Each automatically calculated delay time is continuously displayed in the human-computer interaction window until the "Stop" button is clicked, and the delay time corresponding to the frame number is stored and recorded in a file format, and the file can be opened for reference.

[0039] In addition, we found that as the task requirements of the system to be tested become more and more stringent, its working scenes and environments change greatly. The delay time of the system to be tested is different under different task loads (internal factors) and different environmental factors (external factors). However, the traditional solution can only obtain the delay time at a specific time point, and cannot obtain the delay time distribution, which is not conducive to accurate analysis of the performance of the system to be tested. Utilizing the present invention, different specific identifiers can be loaded in test images of different frames within a set time period. For example, clicking the "Start" button on the human-computer interaction interface triggers the insertion of an increasing frame number into the specified coordinate area of ​​each consecutive test image frame. And each frame number corresponds to a counter. Through the corresponding counter, the delay time of the current frame test image passing through the system to be tested is obtained. The delay time corresponding to different frame numbers is statistically utilized to obtain the delay time distribution of the system to be tested within the set time period, which is helpful for accurate analysis of the performance of the system to be tested.

[0040] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations on the present invention. A person skilled in the art may change, modify, replace and modify the above embodiments within the scope of the present invention without departing from the principles and intent of the present invention.

Claims

1. An automatic measurement device for delay of an image processing system, characterized in that: It includes a test image generation module, an identification encoding module, an identification decoding module, a counter module, a first interface module, and a second interface module; the automatic measuring device uses the first interface module as an output interface and the second interface module as an input interface, and connects the automatic measuring device and the system to be tested through a video cable; The test image generation module generates a test image and outputs it to the identification coding module. The identification coding module sets a specific identification in the specified coordinate area of ​​the test image according to the setting of the host computer, and generates a delayed counting trigger signal to the counter module while setting the specific identification, and the counter module starts counting; the identification coding module outputs the test image with the specific identification to the first interface module; the first interface module outputs the test image with the specific identification to the system to be tested; After being processed in the system under test, the test image is output from the system under test to the second interface module; The second interface module performs protocol analysis on the test image processed by the system to be tested, and outputs the analyzed test image to the identification decoding module. The identification decoding module determines whether it is the specific identification in the specified coordinate area of ​​the received image according to the settings of the upper computer. If it is a specific identification, it generates a delayed counting stop signal to the counter module, and the counter module stops counting.

2. The automatic measurement device for image processing system delay according to claim 1, characterized in that: The test image generation module generates a test image signal with the same input resolution, timing, reference clock and frame rate according to the input resolution, timing, reference clock and frame rate parameters of the system to be tested.

3. The automatic measurement device for delay of an image processing system according to claim 1, characterized in that: The frequency of the counter module is higher than a set required value, and the set required value is determined according to the delay measurement accuracy requirement.

4. The automatic measurement device for delay of an image processing system according to claim 1, characterized in that: The specific identifier is a frame number.

5. A method for automatically measuring the delay of an image processing system using the device according to any one of claims 1 to 4, characterized in that: Obtaining the delay error from the time when the test image generation module generates the test image to the time when the first interface module outputs the test image with the specific identification to the system to be tested, which is composed of two parts of delay errors that can be accurately measured; one part of the delay error t1 is the delay error from the time when the identification encoding module sets the specific identification to the time when the test image is output to the first interface module, and the delay error t1 is accurately determined according to the range of the specified coordinate area; the other part of the delay error t2 is the delay error generated by the interface protocol conversion process of the first interface module; Obtain the delay error from the time when the system under test outputs the processed test image to the second interface module to the time when the counter module stops counting, which is composed of two parts of delay errors that can be accurately measured; one part of the delay error t3 is the delay error generated by the interface protocol conversion process of the second interface module; the other part of the delay error t4 is the time required for the identification decoding module to determine the specified coordinate area, and the delay error t4 is accurately determined according to the range of the specified coordinate area; Obtaining a count value of the counter, and converting the count value to obtain an initial delay time T1; The final image processing system delay is T = T1-Δti, i = 1, 2, 3, 4.

6. The method according to claim 5, characterized in that: Within a set time period, different specific identifiers are loaded into test images of different frames. Each specific identifier can be used to obtain the delay time of the current frame test image passing through the system to be tested. The delay times obtained using different specific identifiers are counted to obtain the delay time distribution of the system to be tested within the set time period.

Images