Method and system for detecting stable picture of electronic equipment and image processing system

Abstract

The invention provides an electronic equipment stable picture detection method and system and an image processing system, and the method comprises the steps: carrying out the frame extraction processing of image frames in a video stream obtained based on a camera device, so as to obtain a plurality of first target images; for any two adjacent first target images, calculating a first difference degree of the two first target images on the whole; when the first difference degree is larger than or equal to a first threshold value, the first difference degree is stored in a data set, and otherwise, the two first target images are divided into a plurality of image blocks; performing difference analysis on the two image blocks at the corresponding positions of the two first target images, updating the first difference degree, and storing the updated first difference degree into a data set; and judging the image frame in a stable state based on the numerical value in the data set. According to the method, good adaptability to various screen types is achieved, the universality and precision of picture stability detection are greatly improved, and the detection effect in different application scenes is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a method and system for detecting a stable picture of an electronic device, and an image processing system. Background Art

[0002] In the performance evaluation of modern electronic devices, the fully stable screen detection algorithm has important application value, especially in the detection of screen stability after the APP click response of in-vehicle systems, POS machines, tablet computers, mobile phones and other devices. The so-called "fully stable" means that the screen of the device no longer changes after completing all loading and rendering tasks. The detection of this state is of great significance for evaluating the performance of the device, such as response speed and processing power.

[0003] At present, the common screens on the market are mainly divided into two categories: LCD (liquid crystal display) and OLED (organic light-emitting diode). Under high-speed cameras, videos recorded by LCD screens usually do not have obvious screen flickering, so there are relatively mature solutions for completely stable image detection on LCD screens. Most of these solutions rely on setting thresholds to judge the stability of the image, but when applied to a wider range of scenarios, the stability and accuracy of these methods are often difficult to guarantee.

[0004] However, for OLED screens, due to the serious flicker problem, existing detection methods are difficult to effectively deal with. On OLED devices, there are few and imperfect solutions for image stability detection, which directly affects the objective evaluation of the performance of these devices. In order to solve this problem and improve the accuracy of OLED device performance evaluation, it is particularly important to develop new detection methods.

[0005] Based on the above analysis, the current detection methods have many limitations when facing different types of screens, especially OLED screens. Summary of the invention

[0006] The object of the present invention is to provide a universal electronic device stable picture detection method, system and image processing system that are compatible with various screen types.

[0007] In order to achieve the above object, the present invention provides a method for detecting a stable picture of an electronic device, which comprises:

[0008] Start a video stream based on the image captured by the camera device on the electronic device, where the video stream includes a number of image frames;

[0009] Performing frame extraction processing on the image frames according to a time sequence and a preset frequency to obtain a plurality of first target images;

[0010] For any two adjacent first target images:

[0011] Calculate the difference between the two first target images as a whole to obtain a first difference degree;

[0012] When the first difference is greater than or equal to a first threshold, storing the first difference in a data set;

[0013] When the first difference is less than the first threshold, the two first target images are respectively divided into image blocks of the same number and size;

[0014] Performing difference analysis on the two image blocks at corresponding positions of the two first target images one by one, and calculating the partition difference degree;

[0015] Counting all the calculated differences of the partitions to obtain an overall difference;

[0016] updating the first difference degree based on the overall difference degree, and storing the updated first difference degree in the data set;

[0017] The image frame in a stable state is determined based on the value of the first difference in the data group.

[0018] Preferably, the method of calculating the difference between the two first target images as a whole includes:

[0019] Converting two adjacent first target images into grayscale images, and performing absolute value difference on the two grayscale images to obtain a grayscale difference image;

[0020] Modifying pixel values ​​on the grayscale difference image that are greater than a second threshold value into maximum values, and modifying pixel values ​​on the grayscale image that are less than or equal to the second threshold value into minimum values, so as to obtain a differential binary image;

[0021] The number of pixels on the differential binary image that are the maximum values ​​is counted to obtain the first difference degree.

[0022] Preferably, the second threshold is greater than 0 and less than or equal to 10.

[0023] Preferably, the method for calculating the partition difference includes:

[0024] Converting two adjacent first target images into grayscale images, and performing absolute value difference on the two grayscale images to obtain a grayscale difference image;

[0025] Performing edge detection on the grayscale difference image, and generating a differential edge image according to the detection result;

[0026] For any pair of corresponding image blocks:

[0027] Extracting a local map of a corresponding position from the differential edge map;

[0028] The grayscale values ​​of all pixels on the local image are counted. If the grayscale values ​​of all pixels are minimum values, the partition difference is zero; if the grayscale value of any pixel is a maximum value, a difference analysis is performed on the current pair of image blocks through a pre-configured classification model based on a neural network model, and the partition difference is obtained according to the analysis result.

[0029] Preferably, the method for processing the two image blocks by using the classification model includes:

[0030] splicing the two first target images including the plurality of the image blocks to obtain a composite image;

[0031] Creating a three-channel pure black background image, the image specification of the background image being the same as that of the composite image;

[0032] splicing the background image and the composite image together vertically to obtain an image to be processed;

[0033] The to-be-processed graph is placed into the classification model to obtain the partition difference.

[0034] Preferably, in the data group, a search is performed from back to front to find out the minimum value at the last position to obtain a target value, and a previous frame in a group of the first target images corresponding to the target value is used as an image frame in which the picture is in a final stable state.

[0035] Preferably, after the target value is obtained, a predetermined number of the image frames are extracted frame by frame from a previous frame in a group of the first target images corresponding to the target value to obtain a plurality of second target images;

[0036] Processing the second target image based on the same detection method as the first target image, and updating the processing result into the data group;

[0037] In the data group, after the target value is obtained, a previous frame in a group of the second target images corresponding to the target value is used as an image frame in which the picture is in a final stable state.

[0038] The present invention also provides an image processing system, which includes an image processor and a camera device, wherein the camera device is used to obtain a screen startup video stream on an electronic device, and the video stream includes a plurality of image frames; the image processor is used to process the image frames through the above-mentioned electronic device stable screen detection method.

[0039] The present invention also provides an electronic device stable picture detection system, which comprises:

[0040] one or more processors;

[0041] Memory;

[0042] and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the programs include instructions for executing the electronic device stable picture detection method as described above.

[0043] The present invention also provides a computer-readable storage medium, which includes a computer program, and the computer program can be executed by a processor to implement the electronic device stable picture detection method as described above.

[0044] Compared with the prior art, the stable picture detection method provided by the above technical solution of the present invention obtains the video stream in a non-invasive manner, extracts the image frames in the video stream, and first quickly judges the difference between the two adjacent first target images from the overall perspective, and then divides the two first target images into blocks for the ones with smaller differences, analyzes the partition difference between each pair of image blocks, and then calculates the overall difference through statistics of all partition differences. This partition difference analysis method can accurately identify the stable state of the picture even when the noise is large and the frequency ripple is obvious, so that its application performance on OLED devices is significantly improved. At the same time, for LCD screens, the scheme can also be robustly executed, ensuring good adaptability to a variety of screen types. In general, the technical scheme achieves good adaptability to a variety of screen types through flexible difference calculation and update methods, greatly improves the versatility and accuracy of picture stability detection, and significantly improves the detection effect in different application scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 FIG. 4 is a flow chart of a stable picture detection method in an embodiment of the present invention.

[0046] Figure 2 FIG. 1 is a schematic diagram showing a stable image detection implementation principle in one embodiment of the present invention.

[0047] Figure 3 FIG. 2 is a schematic diagram showing a stable image detection implementation principle in another embodiment of the present invention.

[0048] Figure 4 4 is a flow chart of a method for calculating the difference between two first target images as a whole in an embodiment of the present invention.

[0049] Figure 5 Flow chart of a method for calculating partition differences in an embodiment of the present invention.

[0050] Figure 6 4 is a flow chart of a method for processing an image block by using a classification model in an embodiment of the present invention. DETAILED DESCRIPTION

[0051] In order to explain the technical content, structural features, achieved objectives and effects of the present invention in detail, the following is a detailed description in conjunction with the implementation methods and the accompanying drawings.

[0052] This embodiment discloses a method for detecting a stable screen of an electronic device, which is used to detect a stable screen when the electronic device starts a certain screen. Through this detection, the time taken by the electronic device to completely start the screen can be calculated, thereby evaluating the performance of the electronic device.

[0053] Currently, common methods for determining stable images are mainly divided into two categories: intrusive and non-intrusive.

[0054] Intrusive methods usually require installing software on the device under test to record the screen in the background, and then analyzing the recorded video to determine the image stability. The advantage of this method is that the video screen is less disturbed and is easy to process and analyze by algorithms. However, this method will occupy device resources and may affect the normal operation of the device, thereby affecting the objectivity of the test results.

[0055] A non-intrusive method uses a high-speed camera (or an electronic device with a camera function) to record a video, split the video into frame images, and perform differential processing on adjacent images to detect stability. However, this method requires manual setting of thresholds, which easily introduces uncertainty. In addition, although algorithms such as SSIM (Structural Similarity Index) perform well under certain conditions, they are susceptible to interference from noise and frequency ripples, especially in the stability judgment of OLED screens, and it is difficult to provide reliable results.

[0056] Based on this, the stable picture detection method in this embodiment is non-intrusive and not limited by noise and frequency ripple interference, and can be applied to various types of screens such as LCD and OLED.

[0057] like Figure 1 and Figure 2 The stable picture detection method in this embodiment includes the following steps:

[0058] S1: Based on the camera device, a video stream is obtained at a frame rate of 240 (4.17ms / frame) on the screen of the electronic device. The video stream includes the entire process of the screen from the start to the complete start, and the video stream includes several image frames. It should be noted that the working parameters (frame rate) of the camera device can be freely adjusted according to demand and are not limited here.

[0059] S2: According to the time sequence and the preset frequency, the image frames are subjected to frame extraction processing to obtain a plurality of first target images. Figure 2 , the video stream has 41 image frames, which are sorted in the order of generation, from frame 1 to frame 41, frame 1 is the first frame (that is, the first frame of the video stream), and frame 41 is the last frame (that is, the last frame of the video stream). The frame extraction frequency is set to 8, starting from the first frame, then the first target image extracted includes frame 1, frame 9, frame 17, frame 25, frame 33, and frame 41.

[0060] For the extracted first target images, from front to back, a difference analysis is performed on each of the two adjacent first target images. Specifically, for any two adjacent first target images, the following process is performed:

[0061] S3: Calculate the difference between the two first target images as a whole to obtain a first difference degree.

[0062] S4: Determine whether the first difference is greater than or equal to a preset first threshold, if yes, proceed to S5, if no, proceed to 60.

[0063] S5: The first difference is stored in a data group. In the data group, the first difference is arranged in chronological order, with the first difference generated first being arranged first and the first difference generated later being arranged later.

[0064] S60: Divide the two first target images into image blocks of the same number and size respectively, and the image blocks at the same position on the two first target images correspond to each other one by one.

[0065] S61: performing difference analysis on two image blocks at corresponding positions of two first target images one by one, and calculating the partition difference, that is, calculating the difference between the two corresponding image blocks.

[0066] S62: Count all calculated partition differences to obtain an overall difference. It should be noted that the partition differences may be counted in an accumulation or average manner.

[0067] S63: updating the first difference degree based on the overall difference degree, and storing the updated first difference degree in the data set.

[0068] S64: judging the image frame in a stable state based on the value of the first difference in the data group. For example, when the first difference is a minimum value, it means that there is no difference between the two corresponding first target images, and the picture enters a stable state, thereby finding a stable picture.

[0069] In this embodiment, the video stream is obtained in a non-invasive manner, and after the image frames in the video stream are extracted, the difference between the two adjacent first target images is first quickly determined as a whole, and then the two first target images with smaller differences are divided into blocks, and the partition difference between each pair of image blocks is analyzed, and then the overall difference is calculated by statistics of all partition differences. This partition difference analysis method can accurately identify the stable state of the picture even when the noise is large and the frequency ripple is obvious, so that its application performance on OLED devices is significantly improved. At the same time, for LCD screens, this solution can also be robustly executed, ensuring good adaptability to various screen types.

[0070] In general, this detection method achieves good adaptability to various screen types through flexible difference calculation and update methods, greatly improves the versatility and accuracy of picture stability detection, and significantly improves the detection effect in different application scenarios.

[0071] On the other hand, Figure 4 , the method for calculating the difference between the two first target images as a whole includes the following process:

[0072] S30: Convert two adjacent first target images into grayscale images.

[0073] S31: Performing an absolute value difference on the two grayscale images to obtain a grayscale difference image.

[0074] S32: modifying pixel values ​​on the grayscale difference image that are greater than the second threshold value to maximum values, and modifying pixel values ​​on the grayscale image that are less than or equal to the second threshold value to minimum values, so as to obtain a differential binary image.

[0075] S33: Count the number of pixels with maximum values ​​on the differential binary image to obtain a first difference degree.

[0076] Further, the second threshold is greater than 0 and less than or equal to 10.

[0077] Specifically, taking the second threshold value of 7 as an example, the maximum value is 255 and the minimum value is 0. Then, on the grayscale difference image, the pixel values ​​greater than or equal to 7 are modified to 255, and the pixel values ​​less than 7 are modified to 0. Finally, the difference binary image formed only has two values: 255 and 0. Therefore, the first difference degree can be obtained by counting the number of 255 on the difference binary image. If the number of pixel values ​​of 255 on the difference binary image is 185, then the first difference degree is 185.

[0078] On the other hand, Figure 5 , the methods for calculating partition differences include:

[0079] S610: Convert two adjacent first target images into grayscale images.

[0080] S611: Perform absolute value difference on the two grayscale images to obtain a grayscale difference image.

[0081] S612: Perform edge detection on the grayscale difference image, and generate a difference edge image according to the detection result.

[0082] The edge of an image refers to the part of the image where the brightness changes significantly. The grayscale profile of this area can be regarded as a step, that is, a sharp change from one grayscale value in a small buffer area to another grayscale value with a large grayscale difference. The edge of an image concentrates most of the image information. Edge detection is mainly the measurement, detection and positioning of the grayscale change of the image.

[0083] Generally speaking, the purpose of edge detection is to significantly reduce the data size of the image while retaining the original image properties. There are many algorithms for edge detection, and Canny edge detection is an algorithm that uses multiple levels of edge detection.

[0084] For any pair of corresponding image blocks, the following process is performed:

[0085] S613: extracting the local map of the corresponding position from the differential edge map. By extracting the local map corresponding to each pair of image blocks, we can focus on analyzing the edge changes in the image block area and ignore the influence of other areas, thereby improving detection efficiency and accuracy.

[0086] S614: Count the grayscale values ​​of all pixels on the local image. If the grayscale values ​​of all pixels are minimum values ​​(ie, 0), the partition difference is zero; if the grayscale value of any pixel is a maximum value (ie, 255), enter S615.

[0087] S615: Perform difference analysis on the current pair of image blocks through a pre-configured classification model based on a neural network model, and obtain the partition difference degree according to the analysis result. Specifically, when the classification model determines that the difference probability of the current two image blocks is greater than or equal to the preset threshold value 0.85, the distinction difference degree of the current two image blocks is set to 1; otherwise, when the classification model determines that the difference probability of the current two image blocks is less than the preset threshold value, the distinction difference degree of the current two image blocks is set to 0.

[0088] In this embodiment, the canny edge detection algorithm is used to extract edge information of the grayscale difference image generated based on the two first target images, thereby avoiding interference from factors such as color and more accurately determining whether there is a difference in the image block area.

[0089] For image blocks with edge differences, a pre-trained classification model is used to further determine the degree of difference to avoid misjudgment due to factors such as noise.

[0090] According to the classification result and the preset threshold, the number of image blocks with differences is counted, and finally it is determined whether the two first target images are similar.

[0091] In short, in this embodiment, by decomposing the image into smaller units and combining edge detection and classification models, more precise and accurate image difference detection is achieved, thereby more reliably determining whether the OLED screen has achieved a stable picture.

[0092] Furthermore, if Figure 6 , the method of processing two image blocks through the classification model includes:

[0093] S6140: splicing two first target images including a plurality of image blocks to obtain a composite image;

[0094] S6141: Create a three-channel pure black background image, the image specifications of the background image are the same as the composite image;

[0095] S6142: splicing the background image and the composite image together to obtain an image to be processed;

[0096] S6143: Place the graph to be processed into the classification model to obtain the partition difference.

[0097] For example, if the size of the first target image is 64*64, then the size of the composite image after the two first target images are spliced ​​is 64*128, and thus the size of the background image is also 64*128. Finally, the size of the image to be processed obtained by splicing the background image and the composite image is 128*128.

[0098] It should also be noted that, in this embodiment, a 64x128 three-channel pure black image is created in order to meet the input requirements of the classification model while retaining the image information of the original single-channel composite image to avoid information loss.

[0099] On the other hand, after comparing all the first target images, the data group is searched from back to front to find the minimum value at the last position to obtain the target value, and the previous frame in a group of first target images corresponding to the target value is used as the image frame in which the picture is in the final stable state.

[0100] For example, as shown in the figure, the target value queried in the data group is "0" at position A, and a group of first target images corresponding to the target value are frame 25 and frame 33. Then, frame 25 is used as the image frame in which the picture is in the final stable state.

[0101] On the other hand, after obtaining the target value, a refined correction test can be performed as follows:

[0102] In the image frames, a predetermined number of image frames are extracted frame by frame starting from a previous frame in a group of first target images corresponding to the target value, so as to obtain a plurality of second target images;

[0103] Processing the second target image based on the same detection method as the first target image, and updating the processing result into the data group;

[0104] In the data group, after the target value is obtained, the previous frame in a group of second target images corresponding to the target value is used as an image frame in which the picture is in a final stable state.

[0105] In summary, the present invention discloses a method for detecting a stable image of an electronic device, such as Figure 3 , the following is a specific example to illustrate the execution process of the detection method.

[0106] 1. There are 41 image frames of the video stream of the electronic device startup screen acquired by the camera device, from frame 1 to frame 41.

[0107] 2.1. The image frames are extracted at a frequency of 8, and the first target image formed includes frame 1, frame 9, frame 17, frame 25, frame 33, and frame 41.

[0108] 2.2. From front to back, extract two adjacent first target images one by one for difference analysis. Specifically, perform difference analysis on frame 1-frame 9, frame 9-frame 17, frame 17-frame 25, frame 25-frame 33, and frame 33-frame 41. For any pair of first target images, convert them into grayscale images and perform absolute value difference to obtain grayscale difference images.

[0109] 2.3 Change the pixel values ​​on the grayscale difference image that are greater than or equal to 7 to 255, and change the pixel values ​​that are less than 7 to 0, to obtain a differential binary image.

[0110] 2.4 Perform edge detection on the grayscale difference image to obtain a differential edge image.

[0111] 2.5 Count the number of pixels N1 whose value is 255 in the differential binary image.

[0112] 2.6 If N1 is 185, which is greater than the first threshold value 180, it means that the two first target images forming the differential binary image are different. Therefore, the first difference degree 185 of the current two first target images is stored in the data group. In addition, when N1 is greater than the first threshold value, the first difference degree can also be uniformly set to the maximum value 255 to facilitate the final query analysis. If N1 is 100, which is less than the first threshold value 180, the difference degree analysis is continued through the following step 2.7.

[0113] 2.7 The current pair of first target images is separated into a 64*64 grid to obtain a number of image blocks.

[0114] 2.8 For any pair of corresponding image blocks, first extract the local image of the corresponding position from the differential edge map, and count the grayscale value of the local image.

[0115] 2.9 If the grayscale values ​​on the local image are all minimum values ​​0, then the partition difference between the pair of image blocks is set to 0.

[0116] 2.10 If the grayscale value on the local image has at least one maximum value of 255, the partition difference between the pair of image blocks is set to 1.

[0117] 2.11 Count the partition differences of all image blocks in an accumulative manner, and generate a first difference corresponding to the pair of first target images based on the statistical results.

[0118] 3. Query the data set and search from back to front until the last minimum value is "0" at position A.

[0119] 4. Since the "0" at position A is obtained by comparing frame 25 and frame 33, frame 25 can be regarded as the image frame in which the picture is in the final stable state. In addition, refinement can also be performed based on the following steps.

[0120] 5. Starting from frame 25, move forward 16 frames to extract 9 frames from 17 to 25 from the image frames, thereby obtaining several second target images, which include: frame 17, frame 18, frame 19, frame 20, frame 21, frame 22, frame 23, frame 24, and frame 25.

[0121] 6. Process the second target image based on the same processing method as the first target image, and update the data set based on the processing result.

[0122] 7. Based on the updated data set, the queried target position is at position B "0", and this target value is obtained by comparing frame 24 and frame 25. Therefore, frame 24 is finally used as the image frame in which the picture is in the final stable state.

[0123] In another preferred embodiment of the present invention, an image processing system is also disclosed, which includes an image processor and a camera device, wherein the camera device is used to obtain a screen startup video stream on an electronic device, wherein the video stream includes a plurality of image frames. The image processor processes the image frames based on the electronic device stable screen detection method in the above embodiment to detect an image frame in which the startup screen on the electronic device is in a final stable state.

[0124] The present invention also discloses a stable picture detection system for electronic devices, which includes one or more processors, a memory and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the program includes instructions for executing the stable picture detection method for electronic devices as described above. The processor can adopt a general central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, which are used to execute relevant programs to realize the functions required to be executed by the modules in the stable picture detection system for electronic devices of the embodiment of the present application, or to execute the stable picture detection method for electronic devices of the method embodiment of the present application.

[0125] The present invention also discloses a computer-readable storage medium, which includes a computer program, and the computer program can be executed by a processor to complete the above-mentioned electronic device stable picture detection method. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more available media integrated. The available medium can be a read-only memory (ROM), or a random access memory (RAM), or a magnetic medium, such as a floppy disk, a hard disk, a tape, a magnetic disk, or an optical medium, such as a digital versatile disc (DVD), or a semiconductor medium, such as a solid-state drive (SSD), etc.

[0126] The embodiment of the present application also discloses a computer program product or a computer program, which includes a computer instruction stored in a computer-readable storage medium. A processor of an electronic device reads the computer instruction from the computer-readable storage medium, and the processor executes the computer instruction, so that the electronic device executes the above-mentioned electronic device stable picture detection method.

[0127] The above disclosure is only the preferred embodiment of the present invention, which certainly cannot be used to limit the scope of rights of the present invention. Therefore, equivalent changes made according to the scope of the patent application of the present invention are still within the scope covered by the present invention.

Claims

1. A method for detecting a stable image of an electronic device, characterized in that: include: Start a video stream based on the image captured by the camera device on the electronic device, where the video stream includes a number of image frames; Performing frame extraction processing on the image frames according to a time sequence and a preset frequency to obtain a plurality of first target images; For any two adjacent first target images: Calculate the difference between the two first target images as a whole to obtain a first difference degree; When the first difference is greater than or equal to a first threshold, storing the first difference in a data set; When the first difference is less than the first threshold, the two first target images are respectively divided into image blocks of the same number and size; Performing difference analysis on the two image blocks at corresponding positions of the two first target images one by one, and calculating the partition difference degree; Counting all the calculated differences of the partitions to obtain an overall difference; updating the first difference degree based on the overall difference degree, and storing the updated first difference degree in the data set; The image frame in a stable state is determined based on the value of the first difference in the data group.

2. The method for detecting a stable image of an electronic device according to claim 1, characterized in that: The method of calculating the difference between the two first target images as a whole comprises: Converting two adjacent first target images into grayscale images, and performing absolute value difference on the two grayscale images to obtain a grayscale difference image; Modifying pixel values ​​on the grayscale difference image that are greater than a second threshold value into maximum values, and modifying pixel values ​​on the grayscale image that are less than or equal to the second threshold value into minimum values, so as to obtain a differential binary image; The number of pixels on the differential binary image that are the maximum values ​​is counted to obtain the first difference degree.

3. The method for detecting a stable image of an electronic device according to claim 2, characterized in that: The second threshold is greater than 0 and less than or equal to 10.

4. The method for detecting a stable image of an electronic device according to claim 1, characterized in that: The method for calculating the partition difference includes: Converting two adjacent first target images into grayscale images, and performing absolute value difference on the two grayscale images to obtain a grayscale difference image; Performing edge detection on the grayscale difference image, and generating a differential edge image according to the detection result; For any pair of corresponding image blocks: Extracting a local map of a corresponding position from the differential edge map; The grayscale values ​​of all pixels on the local image are counted. If the grayscale values ​​of all pixels are minimum values, the partition difference is zero; if the grayscale value of any pixel is a maximum value, a difference analysis is performed on the current pair of image blocks through a pre-configured classification model based on a neural network model, and the partition difference is obtained according to the analysis result.

5. The method for detecting a stable image of an electronic device according to claim 4, characterized in that: The method for processing the two image blocks by using the classification model includes: splicing the two first target images including the plurality of the image blocks to obtain a composite image; Creating a three-channel pure black background image, the image specification of the background image being the same as that of the composite image; splicing the background image and the composite image together vertically to obtain an image to be processed; The to-be-processed graph is placed into the classification model to obtain the partition difference.

6. The method for detecting a stable image of an electronic device according to claim 1, characterized in that: In the data group, search from back to front to find out the minimum value at the last position to obtain the target value, and use the previous frame in a group of the first target images corresponding to the target value as the image frame in which the picture is in the final stable state.

7. The method for detecting a stable image of an electronic device according to claim 6, characterized in that: After the target value is obtained, a predetermined number of the image frames are extracted frame by frame from a previous frame in a group of the first target images corresponding to the target value to obtain a plurality of second target images; Processing the second target image based on the same detection method as the first target image, and updating the processing result into the data group; In the data group, after the target value is obtained, a previous frame in a group of the second target images corresponding to the target value is used as an image frame in which the picture is in a final stable state.

8. An image processing system, characterized in that: It comprises an image processor and a camera device, wherein the camera device is used to obtain a screen startup video stream on an electronic device, the video stream comprising a plurality of image frames; the image processor is used to process the image frames by the electronic device stable screen detection method according to any one of claims 1 to 7.

9. A stable picture detection system for electronic equipment, characterized in that: include: one or more processors; Memory; and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, the programs including instructions for executing the electronic device stable picture detection method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: It includes a computer program, which can be executed by a processor to complete the electronic device stable picture detection method as described in any one of claims 1 to 7.

Images