Writing trajectory extraction method, device, equipment and medium

Abstract

This application relates to the field of image processing technology, and discloses a method, apparatus, device, and medium for extracting writing trajectories. The method includes: acquiring an i-th initial image frame sent by a target camera device, where i is an integer greater than 0, and the target camera device is mounted on the back of a blackboard; performing residual filtering on the i-th initial image frame using a reference image frame corresponding to the target camera device to obtain an i-th standardized image frame; and filtering the i-th standardized image frame using a preset time window and dark channel processing method to obtain an i-th writing trajectory image frame. This avoids the problem of low accuracy in determining the writing trajectory caused by opaque occlusion objects located by infrared sensors, thus improving the accuracy of the determined writing trajectory.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to a method, apparatus, device and medium for extracting writing trajectories. Background Technology

[0002] In large classrooms or conference rooms, displaying the lecturer's writing on the blackboard on a screen simultaneously not only solves the problem of not being able to see the lecturer's writing clearly when the audience is too far away from the blackboard, but also saves the lecturer's writing in class as an electronic document, making it convenient to reproduce the lecture scene at any time.

[0003] To achieve this goal, the first step is to capture and process the instructor's writing trajectory in real time. Currently, infrared sensor positioning and elastic wave positioning are used to determine the writing trajectory. However, when there is an opaque object obstructing the infrared sensor's transmitter and receiver, it becomes difficult to effectively detect the writing position, resulting in low accuracy of the determined writing trajectory. While elastic wave positioning avoids the problem of opaque objects, its positioning is not precise enough, also leading to low accuracy of the determined writing trajectory. Summary of the Invention

[0004] The main purpose of this application is to provide a writing trajectory extraction method, apparatus, device and medium, which aims to solve the technical problem that the accuracy of the determined writing trajectory is not high in the current writing trajectory extraction methods.

[0005] To achieve the aforementioned objectives, this application proposes a method for extracting writing trajectories, the method comprising:

[0006] Obtain the i-th initial image frame sent by the target camera device, where i is an integer greater than 0, and the target camera device is installed on the back of the blackboard;

[0007] Using the reference image frame corresponding to the target camera device, residual filtering is performed on the i-th initial image frame to obtain the i-th normalized image frame;

[0008] Using a preset time window and dark channel processing method, the i-th standardized image frame is filtered to obtain the i-th writing trajectory image frame.

[0009] Further, the step of using the reference image frame corresponding to the target camera device to perform residual filtering on the i-th initial image frame to obtain the i-th normalized image frame includes:

[0010] The i-th initial image frame is obtained by subtracting the pixel value at the same pixel position from the reference image frame.

[0011] Furthermore, before the step of acquiring the i-th initial image frame sent by the target camera device, the method further includes:

[0012] Obtain the acquisition preparation signal of the target camera device;

[0013] In response to the acquisition preparation signal, the reference image frame sent by the target camera device is acquired.

[0014] Further, the step of acquiring the reference image frame sent by the target camera device includes:

[0015] Acquire each debug image frame sent by the target camera device;

[0016] Based on the unobstructed image frames determined from each of the debugging image frames, the reference image frame corresponding to the target camera device is used, and a debugging end signal is generated;

[0017] Based on the debugging completion signal, control the preset reminder device to remind you that the debugging is complete.

[0018] Further, the step of filtering the i-th standardized image frame using a preset time window and dark channel processing method to obtain the i-th writing trajectory image frame includes:

[0019] Store the i-th standardized image frame into a preset standardized image frame library;

[0020] Using the i-th standardized image frame as the end image frame, obtain each standardized image frame within the time window from the standardized image frame library to obtain the target image frame set;

[0021] The target image frame set is processed using the dark channel processing method to obtain the i-th writing trajectory image frame.

[0022] Further, the step of processing the target image frame set using the dark channel processing method to obtain the i-th writing trajectory image frame includes:

[0023] Pixel values ​​corresponding to the target pixel position are extracted from each of the normalized image frames in the target image frame set as pixel values ​​to be analyzed, wherein the target pixel position is any pixel position in the normalized image frame;

[0024] Find the smallest pixel value among all the pixel values ​​to be analyzed, and use it as the filtered pixel value;

[0025] The image frame composed of each of the filtered pixel values ​​is taken as the i-th writing trajectory image frame.

[0026] Furthermore, the target camera device is any one of the cameras in the camera array, which is mounted on the back of the blackboard;

[0027] After the step of filtering the i-th standardized image frame using a preset time window and dark channel processing method to obtain the i-th writing trajectory image frame, the method further includes:

[0028] By using the method of taking the minimum pixel value in the overlapping area, the i-th writing trajectory image frame is stitched together to obtain the i-th stitched writing trajectory image frame.

[0029] This application also proposes a writing trajectory extraction device, the device comprising:

[0030] The data acquisition module is used to acquire the i-th initial image frame sent by the target camera device, where i is an integer greater than 0, and the target camera device is installed on the back of the blackboard;

[0031] The first filtering module is used to perform residual filtering on the i-th initial image frame using the reference image frame corresponding to the target camera device to obtain the i-th normalized image frame;

[0032] The second filtering module is used to filter the i-th standardized image frame using a preset time window and dark channel processing method to obtain the i-th writing trajectory image frame.

[0033] This application also proposes a computer device including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of any of the methods described above.

[0034] This application also proposes a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the methods described above.

[0035] This application discloses a writing trajectory extraction method, apparatus, device, and medium. The method acquires the i-th initial image frame sent by a target camera device, where i is an integer greater than 0, and the target camera device is installed on the back of a blackboard. Using a reference image frame corresponding to the target camera device, residual filtering is applied to the i-th initial image frame to obtain the i-th standardized image frame. A preset time window and dark channel processing method are then used to filter the i-th standardized image frame to obtain the i-th writing trajectory image frame. By installing the target camera device on the back of the blackboard, the problem of low accuracy in determining the writing trajectory caused by opaque obstructions from infrared sensor positioning is avoided. The target camera device can accurately capture images on the blackboard, improving the accuracy of the determined writing trajectory. By first performing residual filtering and then using a preset time window and dark channel processing method to filter the standardized image frame obtained from the residual filtering, the accuracy of the determined writing trajectory is further improved. Attached Figure Description

[0036] Figure 1 This is a flowchart illustrating a writing trajectory extraction method according to an embodiment of this application;

[0037] Figure 2 This is a schematic diagram illustrating the change of the initial frame image over time.

[0038] Figure 3 This is a schematic block diagram of the writing trajectory extraction device according to an embodiment of this application;

[0039] Figure 4 This is a schematic block diagram of the structure of a computer device according to an embodiment of this application.

[0040] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0042] Reference Figure 1 This application provides a method for extracting writing trajectories, the method comprising:

[0043] S1: Obtain the i-th initial image frame sent by the target camera device, where i is an integer greater than 0, and the target camera device is installed on the back of the blackboard;

[0044] Specifically, the i-th initial image frame sent by the target camera device is obtained through a communication connection with the target camera device.

[0045] The initial image frame is an image or video frame captured by the target camera device on the blackboard.

[0046] The target camera is installed on the back of the blackboard, meaning the blackboard is located between the target camera and the lecturer. The target camera is facing the lecturer and is shooting the blackboard from the back.

[0047] The information captured by the target camera device is often a mixture of various visual information. Besides the handwritten notes on the blackboard, it also contains many unwanted features, such as indoor and outdoor lighting, shadows, and hands holding pens. If these features are directly projected onto the display screen, they will greatly affect the viewing experience. Therefore, how to remove unwanted visual information while preserving the writing trajectory is the problem this application aims to solve. Analysis shows that the images captured by the target camera device contain not only the writing trajectory we want to preserve, but also lighting, shadows, Gaussian white noise, hand shadows moving along the writing trajectory, and human shadows.

[0048] Figure 2 The diagram illustrates three signal variation curves formed by the signal output from the blackboard captured by the target camera device. The first signal variation curve is the signal variation curve of Gaussian white noise, the second signal variation curve is the signal variation curve of hand shadow and human shadow, and the third signal variation curve is the signal variation curve of chalk writing (i.e., writing trajectory). Among them, the Gaussian white noise signal is a random signal, the hand shadow and human shadow signals can be regarded as high-intensity square waves (i.e., high-intensity square wave signals), and the writing trajectory signal is a step signal. The intensity of the hand shadow and human shadow signals is higher than the intensity of the writing trajectory signal.

[0049] It is understandable that light and shadow (i.e., human shadows and hand shadows) can be regarded as constant signals that do not change over time within a certain period of time.

[0050] It is understandable that the salience of a person and / or hand is higher than that of a writing trajectory; therefore, the signal intensity of the hand shadow and the person's shadow is higher than that of the writing trajectory.

[0051] It is understandable that since a person and / or hand only stay in a certain position for a period of time before moving away, the signal of the hand shadow and the person's shadow is a square wave.

[0052] S2: Using the reference image frame corresponding to the target camera device, perform residual filtering on the i-th initial image frame to obtain the i-th normalized image frame;

[0053] from Figure 2It is known that the signals corresponding to light and shadow are constant signals, and the intensity of the signal does not change with time at any given moment. Moreover, the intensity of the signals corresponding to hand shadows and human shadows is higher than the intensity of the signal of the writing trajectory. Therefore, based on this prior, residual filtering is used to filter out light and shadow.

[0054] Specifically, the image frame captured by the target camera device that does not contain human shadows, hand shadows, or writing traces is used as the reference image frame corresponding to the target camera device; the reference image frame corresponding to the target camera device is subtracted from the i-th initial image frame to obtain an image frame that has been filtered out of light and shadow, and this image frame that has been filtered out of light and shadow is used as the i-th normalized image frame. That is to say, the i-th normalized image frame only contains signals that exist from the reference image frame onwards.

[0055] It is understandable that the shooting range of the reference image frame is the same as that of the i-th initial image frame, so that the same pixel position of the reference image frame and the i-th initial image frame corresponds to the same position on the blackboard.

[0056] S3: Using a preset time window and dark channel processing method, the i-th standardized image frame is filtered to obtain the i-th writing trajectory image frame.

[0057] from Figure 2 It is known that, apart from the signal from the writing trajectory, the other types of signals do not maintain their intensity continuously within a time window; there will always be a moment when the signal intensity is zero. However, once the step signal formed by the writing trajectory appears, the total signal intensity remains unchanged. Based on this characteristic, we can effectively separate the step signal from other signals.

[0058] Specifically, for each pixel in each standardized image frame within the time window, a minimum value is taken in the time dimension, thereby enabling the dark channel processing method to filter out signals other than step signals. Figure 2 The three dashed boxes in the image represent a time window applied to different signals. Within this time window, the minimum value of Gaussian white noise is zero, the minimum value of high-intensity square waves such as hand shadows and human shadows is zero, while the signal intensity of the writing trajectory remains unchanged. Therefore, the signal of the writing trajectory is preserved through dark channel processing. At this point, unwanted signals such as lighting, shadows, Gaussian white noise, hand shadows, and human shadows have been separated, resulting in the i-th image frame that retains only the writing trajectory.

[0059] It is understood that the i-th standardized image frame is the last standardized image frame among all the standardized image frames within the time window.

[0060] This embodiment avoids the problem of low accuracy in determining the writing trajectory caused by opaque objects obstructing the infrared sensor positioning by installing the target camera device on the back of the blackboard; the target camera device can accurately capture images on the blackboard, improving the accuracy of the determined writing trajectory; by first performing residual filtering, and then using a preset time window and dark channel processing method to filter the standardized image frames of the residual filtering, the accuracy of the determined writing trajectory is further improved; the entire process of writing trajectory extraction only requires two steps, residual filtering and dark channel processing, which consumes low computing resources and does not rely on the subsequent detection and segmentation of the handwriting edges, thus improving the timeliness of writing trajectory extraction.

[0061] In one embodiment, the step of using the reference image frame corresponding to the target camera device to perform residual filtering on the i-th initial image frame to obtain the i-th normalized image frame includes:

[0062] S21: Subtract the pixel values ​​at the same pixel positions from the i-th initial image frame and the reference image frame to obtain the i-th standardized image frame.

[0063] Specifically, the pixel values ​​at the same pixel position of the i-th initial image frame are subtracted from those of the reference image frame. The value obtained by subtraction is the value after filtering out constant signals such as illumination and shadow.

[0064] For example, the pixel value of the pixel in row a and column b of the reference image frame is subtracted from the pixel value of the pixel in row a and column b of the i-th initial image frame to obtain the pixel value of the pixel in row a and column b of the i-th standardized image frame.

[0065] This embodiment achieves residual filtering by subtracting the pixel value at the same pixel position from the i-th initial image frame and the reference image frame, thus filtering out constant signals such as illumination and shadows.

[0066] In one embodiment, before the step of acquiring the i-th initial image frame sent by the target camera device, the method further includes:

[0067] S11: Obtain the acquisition preparation signal of the target camera device;

[0068] Specifically, when the target camera device is powered on and completes initialization, a data acquisition preparation signal is generated.

[0069] S12: In response to the acquisition preparation signal, acquire the reference image frame sent by the target camera device.

[0070] Specifically, upon receiving the acquisition preparation signal, the reference image frame is determined from the image frames sent by the target camera device.

[0071] Optionally, upon receiving the acquisition preparation signal, the first image frame sent by the target camera device is used as the reference image frame, and the image frames after the first image frame are used as the initial image frames.

[0072] This embodiment provides a basis for residual filtering by acquiring the reference image frame sent by the target camera device based on the acquisition preparation signal.

[0073] In one embodiment, the step of obtaining the reference image frame sent by the target camera device includes:

[0074] S121: Acquire each debug image frame sent by the target camera device;

[0075] Specifically, by establishing a communication connection with the target camera device, the system acquires each debug image frame sent by the target camera device. It is understood that each debug image frame is all image frames captured by the target camera device within a preset time period.

[0076] S122: Determine an unobstructed image frame from each of the debug image frames, use it as the reference image frame corresponding to the target camera device, and generate a debug end signal;

[0077] Specifically, the maximum pixel value is extracted from the pixel values ​​corresponding to the pixel position to be processed in each of the debugging image frames, and the extracted pixel value is used as the pixel value corresponding to the pixel position to be processed in the reference image frame, wherein the pixel position to be processed is any pixel position in the debugging image frame; when the reference image frame corresponding to the target camera device is generated, a debugging end signal is generated.

[0078] S123: Based on the debugging end signal, control the preset reminder device to remind you that the debugging is over.

[0079] Specifically, based on the debugging completion signal, a preset reminder device is controlled to remind the user that the debugging is complete, so that the user can start writing on the blackboard.

[0080] Optionally, the reminder device is an indicator light.

[0081] This embodiment determines unobstructed image frames from each of the debugging image frames, thereby improving the accuracy of the determined reference image frames; moreover, it controls a preset reminder device to remind users of the end of debugging based on the debugging end signal, which is beneficial for users to write on the blackboard after debugging is completed and for extracting the writing trajectory.

[0082] In one embodiment, the step of filtering the i-th standardized image frame using a preset time window and dark channel processing method to obtain the i-th writing trajectory image frame includes:

[0083] S31: Store the i-th standardized image frame into a preset standardized image frame library;

[0084] S32: Using the i-th standardized image frame as the end image frame, obtain each standardized image frame within the time window from the standardized image frame library to obtain the target image frame set;

[0085] Specifically, the i-th standardized image frame is used as the ending image frame. That is, the capture time of the initial image frame corresponding to the i-th standardized image frame is the latest capture time among all capture times corresponding to the target image frame set. The target image frame set includes the i-th standardized image frame.

[0086] S33: Using the dark channel processing method, the target image frame set is processed to obtain the i-th writing trajectory image frame.

[0087] Specifically, the dark channel processing method is used to process the target image frame set, and the processed image frame is used as the i-th writing trajectory image frame.

[0088] This embodiment preserves the writing trajectory signal through dark channel processing. Thus, unwanted signals such as lighting, shadows, Gaussian white noise, hand shadows, and human shadows are separated, resulting in the i-th writing trajectory image frame that retains only the writing trajectory.

[0089] In one embodiment, the step of processing the target image frame set using the dark channel processing method to obtain the i-th writing trajectory image frame includes:

[0090] S331: Extract the pixel value corresponding to the target pixel position from each of the normalized image frames in the target image frame set, as the pixel value to be analyzed, wherein the target pixel position is any pixel position in the normalized image frame;

[0091] S332: Find the smallest pixel value among all the pixel values ​​to be analyzed, and use it as the filtered pixel value;

[0092] Specifically, the smallest pixel value is found from all the pixel values ​​to be analyzed, thereby enabling the extraction of writing traces using dark channel processing. The smallest pixel value is then used as the filtered pixel value. In other words, dark channel processing extracts the pixel value closest to black.

[0093] It is understood that by repeating steps S331 to S332, the filtered pixel value corresponding to each pixel position in the standardized image frame can be determined.

[0094] S333: The image frame composed of each of the filtered pixel values ​​is taken as the i-th writing trajectory image frame.

[0095] This embodiment uses the minimum pixel value in the time dimension as the pixel value in the writing trajectory image frame, thereby achieving the extraction of the writing trajectory using dark channel processing. This allows the dark channel processing method to filter out signals other than step signals. Thus, unwanted signals such as illumination, shadows, Gaussian white noise, hand shadows, and human shadows are separated, resulting in the i-th writing trajectory image frame that retains only the writing trajectory.

[0096] In one embodiment, the target camera device is any one of the cameras in a camera array, which is mounted on the back of the blackboard;

[0097] After the step of filtering the i-th standardized image frame using a preset time window and dark channel processing method to obtain the i-th writing trajectory image frame, the method further includes:

[0098] S4: Using the method of taking the minimum pixel value in the overlapping area, the i-th writing trajectory image frame is stitched together to obtain the i-th writing trajectory stitched image frame.

[0099] For example, each i-th writing trajectory image frame has three writing trajectory image frames. The c-th row and d-th column of the first writing trajectory image frame and the m-th row and n-th column of the second writing trajectory image frame correspond to the same writing point on the blackboard. Then, the minimum pixel value is extracted from the pixel value of the c-th row and d-th column of the first writing trajectory image frame and the pixel value of the m-th row and n-th column of the second writing trajectory image frame. The extracted pixel value is used as the pixel value of the corresponding pixel position of the writing point in the i-th writing trajectory stitched image frame.

[0100] This embodiment uses the method of taking the minimum pixel value in the overlapping area, thereby accurately extracting the writing trajectory at the pixel position in multiple writing trajectory stitched image frames, improving the accuracy of the determined i-th writing trajectory stitched image frame; by using a camera array, it is beneficial to reduce the shooting range of each camera and improve the shooting resolution of each camera, thus improving the accuracy of the determined writing trajectory stitched image frame.

[0101] Reference Figure 3 This application also proposes a writing trajectory extraction device, the device comprising:

[0102] The data acquisition module 100 is used to acquire the i-th initial image frame sent by the target camera device, where i is an integer greater than 0, and the target camera device is installed on the back of the blackboard;

[0103] The first filtering module 200 is used to perform residual filtering on the i-th initial image frame using the reference image frame corresponding to the target camera device to obtain the i-th normalized image frame;

[0104] The second filtering module 300 is used to filter the i-th standardized image frame using a preset time window and dark channel processing method to obtain the i-th writing trajectory image frame.

[0105] This embodiment avoids the problem of low accuracy in determining the writing trajectory caused by opaque objects obstructing the infrared sensor positioning by installing the target camera device on the back of the blackboard; the target camera device can accurately capture images on the blackboard, improving the accuracy of the determined writing trajectory; by first performing residual filtering, and then using a preset time window and dark channel processing method to filter the standardized image frames of the residual filtering, the accuracy of the determined writing trajectory is further improved.

[0106] Reference Figure 4 This application also provides a computer device, which may be a server, and its internal structure may be as follows: Figure 4 As shown. The computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computational and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores data such as writing trajectory extraction methods. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements a writing trajectory extraction method. The writing trajectory extraction method includes: acquiring the i-th initial image frame sent by a target camera device, where i is an integer greater than 0, and the target camera device is mounted on the back of the blackboard; performing residual filtering on the i-th initial image frame using a reference image frame corresponding to the target camera device to obtain the i-th normalized image frame; and filtering the i-th normalized image frame using a preset time window and dark channel processing method to obtain the i-th writing trajectory image frame.

[0107] This embodiment avoids the problem of low accuracy in determining the writing trajectory caused by opaque objects obstructing the infrared sensor positioning by installing the target camera device on the back of the blackboard; the target camera device can accurately capture images on the blackboard, improving the accuracy of the determined writing trajectory; by first performing residual filtering, and then using a preset time window and dark channel processing method to filter the standardized image frames of the residual filtering, the accuracy of the determined writing trajectory is further improved.

[0108] One embodiment of this application also provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements a writing trajectory extraction method, including the steps of: acquiring the i-th initial image frame sent by a target camera device, where i is an integer greater than 0, and the target camera device is installed on the back of a blackboard; using a reference image frame corresponding to the target camera device, performing residual filtering on the i-th initial image frame to obtain the i-th standardized image frame; and using a preset time window and dark channel processing method to filter the i-th standardized image frame to obtain the i-th writing trajectory image frame.

[0109] The writing trajectory extraction method described above avoids the problem of low accuracy in determining the writing trajectory caused by opaque obstructions from infrared sensor positioning by installing the target camera device on the back of the blackboard; the target camera device can accurately capture images on the blackboard, improving the accuracy of the determined writing trajectory; by first performing residual filtering, and then using a preset time window and dark channel processing method to filter the standardized image frames of the residual filtering, the accuracy of the determined writing trajectory is further improved.

[0110] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media provided in this application and in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual-speed SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0111] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0112] The above description is only a preferred embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural changes made based on the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method of extracting a writing trajectory, characterized by, The method comprises: obtaining an i-th initial image frame sent by a target camera device, wherein i is an integer greater than 0, and the target camera device is installed on the back of a blackboard; performing residual filtering on the i-th initial image frame by using a reference image frame corresponding to the target camera device to obtain an i-th standardized image frame; performing filtering on the i-th standardized image frame by using a preset time window and a dark channel processing method to obtain an i-th writing track image frame; the step of performing filtering on the i-th standardized image frame by using the preset time window and the dark channel processing method to obtain the i-th writing track image frame comprises: storing the i-th standardized image frame into a preset standardized image frame library; obtaining each standardized image frame within the time window from the standardized image frame library by using the i-th standardized image frame as an ending image frame to obtain a target image frame set; processing the target image frame set by using the dark channel processing method to obtain the i-th writing track image frame.

2. The writing trajectory extraction method according to claim 1, characterized by, the step of performing residual filtering on the i-th initial image frame by using the reference image frame corresponding to the target camera device to obtain the i-th standardized image frame comprises: subtracting pixel values of the same pixel position of the i-th initial image frame from the reference image frame to obtain the i-th standardized image frame.

3. The writing trajectory extraction method according to claim 1, characterized by, before the step of obtaining the i-th initial image frame sent by the target camera device, the method further comprises: obtaining a collection preparation signal of the target camera device; in response to the collection preparation signal, obtaining the reference image frame sent by the target camera device.

4. The writing trajectory extraction method according to claim 3, characterized by, the step of obtaining the reference image frame sent by the target camera device comprises: obtaining each debugging image frame sent by the target camera device; determining an image frame without occlusion from each debugging image frame as the reference image frame corresponding to the target camera device, and generating a debugging end signal; controlling a preset reminding device to perform debugging end reminding according to the debugging end signal.

5. The writing trajectory extraction method according to claim 1, characterized by, the step of processing the target image frame set by using the dark channel processing method to obtain the i-th writing track image frame comprises: extracting a pixel value corresponding to a target pixel position from each standardized image frame in the target image frame set as an analysis pixel value, wherein the target pixel position is any pixel position in the standardized image frame; finding a minimum analysis pixel value from each analysis pixel value as a filtered pixel value; combining each filtered pixel value into an image frame as the i-th writing track image frame.

6. The writing trajectory extraction method according to claim 1, characterized by, the target camera device is any camera in a camera array, and the camera array is installed on the back of the blackboard; after the step of performing filtering on the i-th standardized image frame by using the preset time window and the dark channel processing method to obtain the i-th writing track image frame, the method further comprises: splicing each i-th writing track image frame by using an overlapping region minimum pixel value method to obtain an i-th writing track spliced image frame.

7. A writing trajectory extraction apparatus for implementing the writing trajectory extraction method according to any one of claims 1 to 6, characterized by, The device comprises: a data acquisition module configured to acquire an i-th initial image frame sent by a target camera, wherein i is an integer greater than 0, and the target camera is installed on the back of the blackboard; a first filtering module configured to perform residual filtering on the i-th initial image frame by using a reference image frame corresponding to the target camera, to obtain an i-th standardized image frame; a second filtering module configured to perform filtering on the i-th standardized image frame by using a preset time window and a dark channel processing method, to obtain an i-th writing track image frame.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor executes the computer program to implement the steps of the method in any one of claims 1 to 6.

9. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 6.

Images