Accurate test method and system for picture fluency of cabin central control screen

Abstract

The invention discloses an accurate testing method and system for picture fluency of a central control screen of a cockpit, and belongs to the technical field of picture fluency performance testing of the central control screen, and the method comprises the steps: collecting video data of the central control screen in real time; performing frame-by-frame cutting to generate a continuous image frame sequence; calculating the actual frame rate of the video based on the timestamp sequence of the image frame sequence; if the actual frame rate is greater than or equal to the preset required frame rate, detecting a difference value of adjacent frames through a multi-feature fusion image difference calculation method, and judging whether the picture is changed or not in combination with a dynamic threshold value; and when the picture is changed, judging whether the picture is stuck or not according to the difference value of the timestamps of the two frames before and after the change in combination with a smooth qualified time threshold. According to the invention, through high-precision data acquisition, optimized frame rate calculation and multi-feature fusion lag detection, accurate testing of the video fluency of the central control screen is realized, powerful data support is provided for improving the display effect of the central control screen of the cockpit, and the visual experience of a user in the cockpit is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent cockpit center control screen smoothness performance testing, and more specifically, to a precise testing method and system for the smoothness of the cockpit center control screen. Background Art

[0002] Currently, when evaluating the display performance of the cockpit center control screen, smoothness is a key indicator. Existing smoothness testing methods have significant deficiencies in terms of accuracy and adaptability, and it is difficult to meet the increasing performance evaluation requirements of the cockpit center control screen. For example, in a complex cockpit environment, the center control screen may be affected by various factors, such as temperature changes, system load fluctuations, and differences in the quality of different video sources. Traditional testing methods often cannot accurately capture the video smoothness problems in these situations, resulting in inaccurate evaluation of the display effect of the center control screen and unable to provide effective data support for its performance optimization. Summary of the Invention

[0003] In view of this, the present invention provides a precise testing method and system for the smoothness of the cockpit center control screen, which can solve the problems of insufficient accuracy and poor adaptability existing in the existing smoothness testing methods for the cockpit center control screen.

[0004] To achieve the above object, the present invention adopts the following technical solutions:

[0005] In the first aspect, an embodiment of the present invention provides a precise testing method for the smoothness of the cockpit center control screen, including the following steps:

[0006] S1. Real-time collect video data of the cockpit center control screen through a high-frame industrial camera, and perform data transmission through an application layer protocol based on the TCP / IP protocol;

[0007] S2. Cut the received video data frame by frame to generate a sequence of continuous image frames, and mark each frame of image with a unique frame number and a timestamp accurate to milliseconds;

[0008] S3. Based on the timestamp sequence of the image frame sequence, calculate the actual frame rate of the video using a preset frame rate calculation formula;

[0009] S4. If the actual frame rate is lower than the preset required frame rate, it is determined that the video is not smooth; if it is greater than or equal to the preset required frame rate, detect the difference value between adjacent frames through a multi-feature fusion image difference calculation method, including the weighted fusion of gray difference, texture features, and edge features, and combine a dynamic threshold mechanism to determine whether the picture has changed;

[0010] S5. When it is determined that the picture has changed, based on the time difference between the timestamps of the two frames before and after the change, and combined with the smoothness qualified time threshold, determine whether the picture is stuck.

[0011] Furthermore, it also includes:

[0012] S6. If it is determined that the screen is stuck, record the time point and duration of the stuck event and store them in the database. At the same time, generate a visualization chart to analyze the stuck distribution.

[0013] Furthermore, the protocol in step S1 ensures the integrity and coherence of video data through checksum, sequence number mechanism, and data retransmission timer.

[0014] Furthermore, in step S2, the timestamp marking uses a high-precision system clock module with a timing accuracy of ±1 millisecond.

[0015] Furthermore, the preset frame rate calculation formula in step S3:

[0016]

[0017] where N is the total number of video frames; T start is the start time obtained from the corresponding timestamps of the video frame image sequence; T end is the end time obtained from the corresponding timestamps of the video frame image sequence; the timestamps are processed by Kalman filtering to eliminate the system clock jitter error, and the frame rate calculation error is controlled within ±0.1 FPS.

[0018] Furthermore, in step S4, when the navigation display scene is present, the preset required frame rate is 30 FPS;

[0019] when the high-definition video playback scene is present, the preset required frame rate is 60 FPS.

[0020] Furthermore, in step S4, if it is greater than or equal to the preset required frame rate, it includes:

[0021] S41. For two consecutive frame images I n and I n+1 , convert them into grayscale images G n and G n+1 ;

[0022] S42. Calculate the absolute value of the difference in grayscale values of each pixel point:

[0023] △G(x,y) = |G n (x,y) - G n+1 (x,y)|

[0024] where (x,y) is the pixel point coordinate;

[0025] S43. Sum up △G(x,y) for all pixel points and perform normalization processing to obtain the grayscale difference value D n ;

[0026] S44. For the grayscale image G n and G n+1 , use the gray-level co-occurrence matrix to extract the texture difference T at four directions and three pixel spacings n ;

[0027] S45. For the grayscale image G n and G n+1 , use the Canny edge detection to identify the contour change difference E of the objects in the image n ;

[0028] S46. Weightedly fuse the gray-level difference value D n , the texture difference T n and the contour change difference E n to obtain the comprehensive image difference value S n ;

[0029] S47. If the comprehensive image difference value S n is greater than the dynamic threshold, it is determined that the screen has changed; the dynamic threshold is the difference value calculated for the previous 20 frames by the exponentially weighted moving average method, and the smoothing coefficient is set to 0.2.

[0030] Furthermore, in the step S46, the weight distribution of the multi-feature fusion is: the gray-level difference weight is 0.4, the texture feature weight is 0.3, and the edge feature weight is 0.3.

[0031] Furthermore, in the step S5, for a 30FPS video, the smooth passing time threshold is 40ms;

[0032] For a 60FPS video, the smooth passing time threshold is 20ms.

[0033] In a second aspect, an embodiment of the present invention further provides a precise test system for the smoothness of the cockpit central control screen image, including:

[0034] A data acquisition module, configured to collect the video data of the cockpit central control screen in real time through a high-frame industrial camera, and perform data transmission through the application layer protocol based on the TCP / IP protocol;

[0035] A preprocessing module, configured to perform frame-by-frame cutting on the received video data to generate a continuous image frame sequence, and mark a unique frame number and a time stamp accurate to milliseconds for each frame of the image;

[0036] A frame rate analysis module, based on the time stamp sequence of the image frame sequence, calculates the actual frame rate of the video by using a preset frame rate calculation formula;

[0037] The screen change detection module is used to determine that the video is not smooth if the actual frame rate is lower than the preset required frame rate; if it is greater than or equal to the preset required frame rate, it detects the difference value between adjacent frames through an image difference calculation method of multi-feature fusion, including the weighted fusion of grayscale difference, texture features, and edge features, and determines whether the screen has changed in combination with a dynamic threshold mechanism;

[0038] The stuttering detection module is used to determine whether the screen is stuttering when it is determined that the screen has changed, based on the difference between the timestamps of two frames before and after the change, and in combination with the smooth pass time threshold.

[0039] Furthermore, it further includes:

[0040] The data storage and visualization module is used to record the time point and duration of the stuttering event and store them in the database if it is determined that the screen is stuttering, and at the same time generate a visualization chart to analyze the stuttering distribution.

[0041] It can be seen from the above technical solutions that compared with the prior art, the present invention has the following technical advantages:

[0042] The present invention uses a high-frame industrial camera installed in the middle position of the front row to collect video data, and through high-precision data collection, optimized frame rate calculation, multi-feature fusion stuttering detection, and visualization analysis, it realizes the accurate test of the video smoothness of the central control screen, provides strong data support for improving the display effect of the cockpit central control screen, and thus helps to improve the visual experience of users in the cockpit. Description of the Drawings

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention, and for those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0044] Figure 1 It is a flowchart of the accurate test method for the smoothness of the cockpit central control screen provided by the present invention.

[0045] Figure 2 It is a specific flowchart when it is greater than or equal to the preset required frame rate in step S4 provided by the present invention.

[0046] Figure 3 It is a block diagram of the accurate test system for the smoothness of the cockpit central control screen provided by the present invention. Detailed Embodiments

[0047] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0048] Embodiment 1:

[0049] An embodiment of the present invention discloses a precise test method for the smoothness of the cockpit central control screen, which is applicable to the evaluation of the smoothness performance of scenarios such as the cockpit central control screen and intelligent terminals; referring to Figure 1 as shown, it includes the following steps S1 to S6:

[0050] S1. Real-time collect the video data of the cockpit central control screen through a high-frame industrial camera, and perform data transmission through an application layer protocol based on the TCP / IP protocol;

[0051] For example, taking the test of the in-vehicle central control screen as an example, a high-frame industrial camera installed in the middle position of the front row records the video of the central control screen and transmits the data to the test system in real time. Through a reliable wired data transmission method, such as using a high-speed Ethernet connection, the video data of the central control screen captured by the high-frame industrial camera is quickly and completely transmitted to the test system. During the transmission process, a custom application layer protocol based on the TCP / IP protocol is used. This protocol adds a strict checksum and sequence number mechanism during data packet encapsulation to ensure that there is no data loss or error during transmission. At the same time, a data retransmission timer is set. When a data packet is found to be lost or damaged, it can be automatically retransmitted within 50 milliseconds to ensure the integrity and coherence of the data.

[0052] S2. Cut the received video data frame by frame to generate a sequence of continuous image frames, and mark each frame of image with a unique frame number and a timestamp accurate to milliseconds;

[0053] In this step, the received video is cut frame by frame to convert the video into a sequence of continuous image frames. At the same time, each frame of image is marked with a unique frame number and the corresponding timestamp, and the timestamp is accurate to milliseconds. The precise marking of the timestamp is realized by using a high-precision system clock module, and its timing accuracy can reach ±1 millisecond for subsequent time-related calculations. During the frame numbering process, a 64-bit integer encoding is used to ensure that there is no number overflow during long-term video recording, providing a stable basis for large-scale video data processing.

[0054] S3. Based on the timestamp sequence of the image frame sequence, use a preset frame rate calculation formula to calculate the actual frame rate of the video;

[0055] In this step, assume the total number of video frames is N. By analyzing the timestamps of video frames, the start time T and end time T of the video are calculated. start and the end time T end . An optimized frame rate calculation method is adopted, that is:

[0056]

[0057] The above preset frame rate calculation formula also takes into account the time deviation in the actual video recording process, and is more accurate than the traditional frame rate calculation method. During the calculation process, the timestamps are filtered multiple times to remove the small deviations caused by system clock jitter, further improving the accuracy of frame rate calculation. For example, using the Kalman filter algorithm to process the timestamp sequence can effectively estimate and correct the uncertainty of the system clock, making the error of the frame rate calculation result controlled within ±0.1 FPS.

[0058] In the above steps S2 and S3, the proposed optimized frame rate calculation method and accurate timestamp marking and analysis make the frame rate calculation more accurate and the time-related calculations more precise, providing more reliable basic data for the smoothness test. In actual tests, accurate frame rate calculation and timestamp analysis are crucial for judging the smoothness of the video. Traditional frame rate calculation methods may be affected by factors such as system clock error and data transmission delay, resulting in inaccurate calculation results. This algorithm effectively reduces the interference of these factors through fine processing of timestamps and an optimized frame rate calculation formula, improving the credibility of the test results and laying a solid foundation for subsequent stutter detection and analysis.

[0059] S4. If the actual frame rate is lower than the preset required frame rate, it is determined that the video is not smooth; if it is greater than or equal to the preset required frame rate, the difference value between adjacent frames is detected through a multi-feature fusion image difference calculation method, including the weighted fusion of gray difference, texture features, and edge features, and combined with a dynamic threshold mechanism to determine whether the picture has changed;

[0060] In this step, the calculated frame rate FPS is compared with the required frame rate FPS preset by the system require . If FPS ≥ FPS require , the video initially meets the smoothness requirements at the frame rate level, and the subsequent detailed detection steps are continued;

[0061] If FPS < FPS require , it is directly determined that the video as a whole does not meet the basic smoothness requirements, relevant information is recorded, and the current detection process is ended (it can be decided whether to re-collect and detect the video according to actual needs). In actual applications, for different types of cockpit center control screen application scenarios, such as navigation display, multimedia playback, etc., different FPS can be set according to their functional characteristics and user experience requirementsrequire Values, for example, the navigation display scenario can be set to 30 FPS, while the high-definition video playback scenario can be set to 60 FPS.

[0062] Among them, in step S4, when it is greater than or equal to the preset required frame rate, refer to Figure 2 As shown, the specific process includes:

[0063] S41. For two consecutive frames of images I n and I n+1 , convert them into grayscale images G n and G n+1 ;

[0064] S42. Calculate the absolute value of the difference in grayscale values of each pixel point:

[0065] △G(x,y) = |G n (x,y) - G n+1 (x,y)|

[0066] where (x,y) is the pixel point coordinate;

[0067] S43. Sum up △G(x,y) of all pixel points and perform normalization processing to obtain the grayscale difference value D of the entire frame of image n ;

[0068] S44. For the grayscale images G n and G n+1 , use the gray-level co-occurrence matrix to extract the texture difference T at four directions and three pixel spacings n ;

[0069] S45. For the grayscale images G n and G n+1 , use Canny edge detection to identify the contour change difference E of the objects in the image n ;

[0070] S46. Combine the grayscale difference value D n , the texture difference T n and the contour change difference E n through weighted fusion to obtain the comprehensive image difference value S n ;

[0071] S47. If the comprehensive image difference value S n is greater than the dynamic threshold, it is determined that the picture has changed; the dynamic threshold is the difference value calculated for the previous 20 frames by the exponential weighted moving average method, and the smoothing coefficient is set to 0.2.

[0072] Among them, adaptive picture change detection involves multi-feature fusion analysis of adjacent two frames of images. The features mentioned here include grayscale difference, texture features, and edge features.

[0073] The gray - level difference refers to the difference in pixel gray - level values between two frames of images. If the gray - level difference between two frames of images is large, it indicates that there are obvious changes in the picture content. However, relying solely on the gray - level difference is not enough. For example, in the case of light changes or color gradients, the gray - level difference may be misjudged as a picture change.

[0074] Next is the texture feature. The gray - level co - occurrence matrix (GLCM) can be used to extract texture information in four directions and three pixel spacings. The gray - level co - occurrence matrix is a method for describing the texture features of an image. By statistically analyzing the spatial relationship of pixel pairs, it can reflect attributes such as the uniformity and contrast of the texture. This can capture the structural changes of the image more comprehensively, such as the icon blinking in a dynamic map or the special - effect scenes in a video.

[0075] Then comes the edge feature, using the Canny edge - detection algorithm. Edge detection can identify the contour changes of objects in an image, such as the changes in road lines in a navigation interface or the movement of objects in a multimedia video. Setting high and low thresholds (30 and 100) can reduce noise interference while retaining important edges.

[0076] The weights of the above three features are 0.4 for gray - level difference, 0.3 for texture feature, and 0.3 for edge feature. When comprehensively judging the picture change, the contribution of the gray - level difference is slightly larger, but the texture and edge features also play important roles. This weight distribution is determined through a large number of experiments to balance the detection effects in different scenarios.

[0077] Next is the dynamic threshold setting. The exponential weighted moving average method is adopted to adjust the current threshold according to the difference values of the previous 20 frames. This can adapt to the dynamic changes of video content. For example, when the picture changes rapidly, the threshold is increased to avoid misjudgment; when the picture is relatively static, the threshold is decreased to improve the detection sensitivity. The smoothing coefficient is set to 0.2, indicating that the difference value of the current frame has a small impact on the threshold, and it mainly relies on historical data to maintain the stability of the threshold.

[0078] Generally speaking, "picture change detection" accurately identifies the changes in the picture by comprehensively considering various image features and combining the dynamically adjusted threshold, thereby improving the accuracy and reliability of the test.

[0079] The calculation process is as follows:

[0080] First, in steps S41 - S3, the video data is analyzed frame by frame. An algorithm based on image difference and threshold judgment is adopted. For two consecutive frames of images I n and I n+1 , first, they are converted into grayscale images G n and G n+1, the color image is converted into a grayscale image by the weighted average method, which can effectively retain the main information of the image. Then, calculate the absolute value of the difference in grayscale values of each pixel:

[0081] △G(x,y) = |G n (x,y) - G n+1 (x,y)| (2)

[0082] where (x,y) are the coordinates of the pixel. Sum up △G(x,y) for all pixels and perform normalization to obtain the difference value D of the entire frame of the image n , and the normalization is achieved by dividing by the total number of pixels in the image, making the difference value between 0 and 1, which is convenient for subsequent judgment.

[0083] Secondly, in step S44, texture information is obtained by calculating the gray-level co-occurrence matrix. When calculating the gray-level co-occurrence matrix, the parameter settings of 4 directions (0°, 45°, 90°, 135°) and 3 pixel spacings are adopted, which can comprehensively capture the texture difference T of the image n ;

[0084] T n = |C n - C n+1 | (3)

[0085] where C n and C n+1 are the texture contrast values of two frames, which are calculated by the gray-level co-occurrence matrix GLCM in 4 directions (0°, 45°, 90°, 135°) and 3 pixel spacings.

[0086] Then, in step S45, the Canny edge detection algorithm is used to extract the contour change difference E n , and the high and low thresholds are set to (30, 100) to reduce noise interference while ensuring the accuracy of edge detection. Extract the edge pixels of two frames and calculate the complement of the edge overlap rate:

[0087]

[0088] In step S46, these features are weighted and fused to obtain a comprehensive image difference value S n :

[0089] S n = D n × α + T n × β + E n × γ (5)

[0090] Among them, α is the grayscale difference weight set to 0.4, β is the texture feature weight set to 0.3, and γ is the edge feature weight set to 0.3. These weights are verified and optimized through a large amount of experimental data and can achieve good detection effects under different video contents.

[0091] Finally, in step S47, the comprehensive image difference value S n is compared with the dynamic threshold. When it is greater than the threshold, it is determined that the picture has changed. Introducing the dynamic threshold mechanism is to more accurately judge the picture change. According to the overall content of the video and the difference situation of the previous frames, the threshold T for judging the picture change is dynamically adjusted. For example, the method of exponentially weighted moving average is adopted to calculate the current threshold T based on the difference values of the previous 20 frames, so that the threshold can adapt to the changes of different video contents. In the exponentially weighted moving average calculation, the smoothing coefficient is set to 0.2, which can maintain a certain stability while quickly responding to the changes of video contents, and avoid excessive fluctuations of the threshold.

[0092]

[0093] Among them, T prev is the threshold of the previous frame (or previous moment). S n is the comprehensive difference value between the current frame and the previous frame (calculated through grayscale, texture, and edge features). is the smoothing coefficient (usually set to 0.2), which controls the weight ratio of the old and new thresholds.

[0094] If S n > T, it is determined that the picture has changed; if S n ≤ T, it is regarded as the picture being static or changing normally.

[0095] Through the adaptive dynamic threshold adjustment mechanism, it can adapt to the normal display fluctuations of the central control screen under different test environments, effectively reduce the occurrence of misjudgment, and improve the reliability of the system in complex environments. Under the influence of factors such as different vehicle models, different lighting conditions, and different aging degrees of the central control screen, the display stability of the central control screen will be different. The method of fixed threshold is difficult to adapt to these changes and is prone to misjudgment. The dynamic threshold setting mechanism of this algorithm can dynamically adjust the threshold according to the actual performance of the central control screen, ensure the accuracy of judgment, and improve the applicability and reliability of the algorithm.

[0096] In steps S41 to S47, a multi-feature fusion method is adopted, which improves the accuracy of detecting changes in the central control screen image. Compared with traditional single-feature detection methods, it can capture subtle screen changes more accurately. Traditional single-feature detection methods may only focus on certain aspects of an image, such as relying solely on edge detection or only on texture analysis, and are easily affected by factors such as image noise and lighting changes, resulting in misjudgments or missed detections. However, this algorithm can analyze image changes more comprehensively by fusing multiple features, enhancing the robustness and accuracy of the algorithm.

[0097] S5. When it is determined that the screen has changed, based on the difference between the timestamps of two consecutive frames before and after the change, and combined with the smooth passing time threshold, determine whether the screen is stuck. Here, the screen changing means: when operating a function in the central control screen, the switching and jumping changes before the function page. It can record: the start time and end time of the switched and jumped screen frames, and combined with the smooth passing time threshold, determine whether the screen is stuck.

[0098] In this step, when it is detected that the screen has changed, record the timestamp Tpre of the previous frame before the change and the timestamp Tpost of the next frame after the change, and calculate the interval time △T = Tpost - Tpre (unit: ms). Compare the interval time △T with the smooth passing time T smooth for comparison.

[0099] If △T ≤ T smooth , it is determined that the video plays smoothly during this time period; if △T > T smooth , it is determined to be stuck. In practical applications, for videos with different frame rates, different T smooth values can be set according to experience and user perception research. For example, for a 30 FPS video, T smooth can be set to 40 ms, and for a 60 FPS video, T smooth can be set to 20 ms. In the stuck judgment logic part, if the time interval exceeds the threshold, it is determined to be stuck.

[0100] For example, assume that the grayscale difference D n of the current frame is 0.8; the texture difference T n is 0.5; the contour change difference E n is 0.6, then:

[0101] S n = 0.8 * 0.4 + 0.5 * 0.3 + 0.6 * 0.3 = 0.32 + 0.15 + 0.18 = 0.65

[0102] If the previous threshold is T prev = 0.6, then the new threshold is:

[0103] T = 0.2×0.6 + 0.8×0.65 = 0.12 + 0.52 = 0.64

[0104] At this time, S n = 0.65 > T = 0.64, it is determined that the screen has changed significantly, which may be a freeze.

[0105] S6. If it is determined that the screen is frozen, record the time point and duration of the freeze event, and store them in the database. At the same time, generate a visualization chart to analyze the freeze distribution.

[0106] In this step, when it is determined to be frozen, the system automatically increments the freeze count by 1, and details such as the time point when the freeze occurred and the duration of the freeze are recorded. These freeze data are stored in the MySQL database table, and the table structure includes fields such as freeze serial number, start time of freeze, end time of freeze, and duration of freeze. In the database storage process, a transaction processing mechanism is adopted to ensure atomicity and consistency when writing freeze data in a multi-threaded concurrent manner, and to avoid data loss or incorrect writing. The InnoDB engine of the MySQL database is used to implement transaction processing to ensure data integrity and reliability.

[0107] Furthermore, to facilitate users to intuitively understand the fluency of the video, the system provides data analysis and visualization functions. Through statistical analysis of the freeze data, visualization charts such as a chart showing the change of the freeze count over time and a histogram of the freeze duration distribution are generated. Users can clearly view these charts on the interface of the test system, so as to have a comprehensive understanding of the fluency of the central control screen video. In the process of chart drawing, for example, visualization libraries such as Echarts are used to quickly generate high-quality charts and provide interactive functions such as mouse hovering to view detailed data, zooming, and panning, which is convenient for users to conduct in-depth analysis and comparison.

[0108] In this step S6, not only the basic information of the freeze is recorded, but also various visualization charts are generated through data analysis, providing users with intuitive and comprehensive fluency evaluation results, facilitating users to quickly understand the fluency status of the central control screen video. Compared with the traditional method of simply recording the freeze count, it has stronger practicability and user-friendliness. In the actual process of testing and optimizing the cockpit central control screen, developers and testers need to understand the distribution and pattern of freezes in detail in order to make targeted improvements. The visualization function provided by this algorithm can present complex freeze data in the form of intuitive charts, helping users quickly discover problems, improving the efficiency of testing and optimization, and making the evaluation and improvement of the fluency of the central control screen video more convenient and effective.

[0109] The precise test method for the smoothness of the cockpit center control screen provided by the present invention innovatively integrates multiple features such as grayscale, texture, and edges in the detection of screen changes, and adopts a dynamic threshold mechanism. Compared with the traditional single-feature detection and fixed-threshold judgment methods, it greatly improves the accuracy and adaptability of stutter detection, and can better handle various complex display contents of the center control screen. Traditional methods are prone to misjudgment or missed judgment when facing complex image changes on the center control screen, such as icon flickering in dynamic map navigation and special effect scenes in multimedia videos. However, through multi-feature fusion and dynamic threshold adjustment, this method can more accurately capture the real changes of the screen, effectively distinguish normal screen switching and stutter phenomena, and provide a more reliable basis for the evaluation of the video smoothness of the center control screen.

[0110] Embodiment 2:

[0111] The embodiment of the present invention also provides a precise test system for the smoothness of the cockpit center control screen. Referring to Figure 3 as shown, it includes:

[0112] A data acquisition module, which is used to collect video data of the cockpit center control screen in real time through a high-frame industrial camera and transmit the data through an application layer protocol based on the TCP / IP protocol;

[0113] A preprocessing module, which is used to cut the received video data frame by frame to generate a sequence of continuous image frames, and mark a unique frame number and a timestamp accurate to milliseconds for each frame of image;

[0114] A frame rate analysis module, which calculates the actual frame rate of the video based on the timestamp sequence of the image frame sequence by using a preset frame rate calculation formula;

[0115] A screen change detection module, which is used to determine that the video is not smooth if the actual frame rate is lower than the preset required frame rate; if it is greater than or equal to the preset required frame rate, it detects the difference value between adjacent frames through an image difference calculation method of multi-feature fusion, including the weighted fusion of grayscale difference, texture features, and edge features, and determines whether the screen has changed in combination with a dynamic threshold mechanism;

[0116] A stutter detection module, which is used to determine whether the screen is stuttering when it is determined that the screen has changed, according to the difference between the timestamps of the two frames before and after the change and in combination with a smooth pass time threshold;

[0117] A data storage and visualization module, which is used to record the time point and duration of the stutter event and store them in the database if it is determined that the screen is stuttering, and at the same time generate a visualization chart to analyze the stutter distribution.

[0118] In this embodiment, a high-frame industrial camera (for example, frame rate ≥ 120 FPS) is used to transmit data through a high-speed Ethernet. The application layer protocol supports packet verification and automatic retransmission within 50 ms. The FFmpeg tool is used for frame cutting, the timestamp is marked by a high-precision clock module (±1 ms), and the frame number is encoded with a 64-bit integer. The Kalman filter is applied to the timestamp sequence to eliminate system clock jitter. The calculation formula is:

[0119]

[0120] In the process of stutter detection, a grayscale difference algorithm is adopted to obtain the grayscale difference value of the entire frame image. Texture and edge features are extracted based on the gray-level co-occurrence matrix (4 directions, 3 pixel spacings) and Canny edge detection (threshold 30 - 100). After weighted fusion of the three, a dynamic threshold (exponentially weighted moving average, ) is used to determine whether there is a stutter.

[0121] Finally, the stutter data is stored in a MySQL database (InnoDB engine), and visual charts are generated through Echarts to support interactive analysis.

[0122] For example, in the cockpit test of a certain vehicle model, the system successfully detected 3 stutters (duration > 50 ms) in the navigation interface due to high GPU load in a high-temperature environment, and located the problem time period through visual charts. After the hardware was optimized under the guidance of developers and testers, the number of stutters was reduced to 0.

[0123] In this embodiment, the accuracy of stutter detection is improved and the false positive rate is reduced through multi-feature fusion and dynamic threshold mechanism. The frame rate calculation error is controlled within ±0.1 FPS, and the timestamp accuracy reaches ±1 ms. The stutter distribution can also be intuitively displayed through visual charts, which is beneficial for developers and testers to understand the distribution and pattern of stutters in detail, so as to make targeted improvements.

[0124] In this specification, each embodiment is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The same or similar parts among the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0125] The foregoing description of the disclosed embodiments enables those skilled in the art to practice or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for accurately testing the fluidity of a cockpit central control screen, characterized in that: The following steps are involved: S1. Use a high-frame industrial camera to collect video data from the cockpit central control screen in real time, and transmit the data through an application layer protocol based on the TCP / IP protocol; S2, cutting the received video data frame by frame to generate a continuous image frame sequence, and marking each frame with a unique frame number and a timestamp accurate to milliseconds; S3, based on the timestamp sequence of the image frame sequence, using a preset frame rate calculation formula to calculate the actual frame rate of the video; S4. If the actual frame rate is lower than the preset required frame rate, the video is determined to be not smooth; if it is greater than or equal to the preset required frame rate, the difference values ​​of adjacent frames are detected by a multi-feature fusion image difference calculation method, including grayscale difference, weighted fusion of texture features and edge features, and a dynamic threshold mechanism is combined to determine whether the picture has changed; S5. When it is determined that the picture has changed, whether the picture is stuck is determined based on the difference in the timestamps of the two frames before and after the change and in combination with the smoothness qualified time threshold.

2. According to claim 1, a method for accurately testing the fluidity of the cockpit central control screen is characterized in that: Also includes: S6. If it is determined that the screen is stuck, the time point and duration of the stuck event are recorded and stored in the database, and a visual chart is generated to analyze the stuck distribution.

3. According to claim 1, a method for accurately testing the fluidity of the cockpit central control screen is characterized in that: The protocol in step S1 ensures the integrity and consistency of the video data through checksum, sequence number mechanism and data retransmission timer.

4. According to claim 1, a method for accurately testing the fluidity of the cockpit central control screen is characterized in that: The timestamp in step S2 uses a high-precision system clock module with a timing accuracy of ±1 millisecond.

5. The method for accurately testing the fluidity of the cockpit central control screen according to claim 1 is characterized in that: The preset frame rate calculation formula in step S3 is: Where N is the total number of video frames; T start is the start time obtained from the corresponding timestamp of the video frame image sequence; T end The end time is obtained from the corresponding timestamp of the video frame image sequence; the timestamp is processed by Kalman filtering to eliminate the system clock jitter error, so that the frame rate calculation error is controlled within ±0.1FPS.

6. The method for accurately testing the fluidity of the cockpit central control screen according to claim 1 is characterized in that: In step S4, when navigating the display scene, the preset required frame rate is 30FPS; When playing high-definition video, the default frame rate is 60FPS.

7. The method for accurately testing the fluidity of the cockpit central control screen according to claim 1 is characterized in that: In the step S4, if the frame rate is greater than or equal to the preset required frame rate, the following steps are performed: S41, for two consecutive frames of image I n and I n+1 , convert it into a grayscale image G n and G n+1 ; S42, calculate the absolute value of the gray value difference of each pixel: △G(x,y)=|G n (x,y)-G n+1 (x,y)| Where (x, y) is the pixel coordinate; S43, summing up the △G(x,y) of all pixel points and performing normalization processing to obtain the grayscale difference value D of the entire frame image n ; S44, grayscale image G n and G n+1 , use the gray level co-occurrence matrix to extract the texture difference T in four directions and three pixel spacings n ; S45, grayscale image G n and G n+1 , use Canny edge detection to identify the contour change difference E of the object in the image n ; S46, grayscale difference value D n , texture difference T n and contour change difference E n , perform weighted fusion to obtain the comprehensive image difference value S n ; S47, if the comprehensive image difference value S n If it is greater than a dynamic threshold, it is determined that the picture has changed; the dynamic threshold is the difference value of the previous 20 frames calculated by the exponential weighted moving average method, and the smoothing coefficient is set to 0.

2.

8. The method for accurately testing the fluidity of the cockpit central control screen according to claim 7 is characterized in that: In step S46, the weight distribution of multi-feature fusion is: grayscale difference weight 0.4, texture feature weight 0.3, edge feature weight 0.

3.

9. The method for accurately testing the fluidity of the cockpit central control screen according to claim 1 is characterized in that: In step S5, for a 30FPS video, the smooth pass time threshold is 40ms; For 60FPS videos, the smoothness qualification time threshold is 20ms.

10. A precise test system for the fluidity of the cockpit central control screen, characterized in that: include: The data acquisition module is used to collect the video data of the cockpit central control screen in real time through a high-frame industrial camera, and transmit the data through an application layer protocol based on the TCP / IP protocol; A preprocessing module is used to cut the received video data frame by frame, generate a continuous image frame sequence, and mark each frame with a unique frame number and a timestamp accurate to milliseconds; A frame rate analysis module calculates the actual frame rate of the video using a preset frame rate calculation formula based on the timestamp sequence of the image frame sequence; The picture change detection module is used to determine that the video is not smooth if the actual frame rate is lower than the preset required frame rate; if it is greater than or equal to the preset required frame rate, the difference value of adjacent frames is detected by the image difference calculation method of multi-feature fusion, including the weighted fusion of grayscale difference, texture feature and edge feature, and combined with the dynamic threshold mechanism to determine whether the picture has changed; The freeze detection module is used to determine whether the picture is frozen based on the difference in the timestamps of the two frames before and after the change and the smooth qualified time threshold when the picture changes.

Images