WO2020168818A1 - Procédé, système et dispositif de traitement d'image, support d'informations et dispositif de tableau noir - Google Patents

Procédé, système et dispositif de traitement d'image, support d'informations et dispositif de tableau noir Download PDF

Info

Publication number
WO2020168818A1
WO2020168818A1 PCT/CN2019/127758 CN2019127758W WO2020168818A1 WO 2020168818 A1 WO2020168818 A1 WO 2020168818A1 CN 2019127758 W CN2019127758 W CN 2019127758W WO 2020168818 A1 WO2020168818 A1 WO 2020168818A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
pixels
blackboard
pixel
type
Prior art date
Application number
PCT/CN2019/127758
Other languages
English (en)
Chinese (zh)
Inventor
邹超洋
Original Assignee
广州视源电子科技股份有限公司
广州视睿电子科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 广州视源电子科技股份有限公司, 广州视睿电子科技有限公司 filed Critical 广州视源电子科技股份有限公司
Publication of WO2020168818A1 publication Critical patent/WO2020168818A1/fr

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/136—Segmentation; Edge detection involving thresholding
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10004—Still image; Photographic image

Definitions

  • This application relates to image processing technology, such as an image processing method, system, device, storage medium, and blackboard device.
  • Image threshold segmentation algorithm is a commonly used image segmentation method. Because of its simple implementation, small calculation amount, and stable performance, it has become the most basic and most widely used segmentation technology in image segmentation. It is widely used in extracting target objects from images. In the program.
  • commonly used threshold segmentation algorithms mainly include two ideas, one is a global threshold segmentation algorithm, and the other is a local threshold segmentation algorithm.
  • the global threshold segmentation algorithm it mainly uses different measurement indicators to find the optimal global threshold, so as to use this optimal global threshold to segment the image, but it is basically impossible for the target scene to have multiple peaks and valleys.
  • Adaptation that is, if there are multiple peaks and valleys in the image, the global threshold segmentation algorithm cannot segment the foreground and the background well. Therefore, it can be seen that for the image with reflection, the global threshold is used The segmentation algorithm cannot extract the target image from the image well, and the accuracy of the target image extraction is low.
  • embodiments of the present invention provide an image processing method, system, device, storage medium, and blackboard device, which can effectively extract a target image from a blackboard image.
  • an embodiment of the present invention provides an image processing method, including:
  • the threshold segmentation process of the image block is specifically performing pixel threshold segmentation on at least one image block of the first blackboard image Processing, so that the pixels of the at least one image block are divided into first type pixels or second type pixels;
  • the pixel is taken as the target pixel
  • a target image is extracted from the first blackboard image.
  • the threshold segmentation processing of the image block at least once is specifically: the threshold segmentation processing of the image block at least twice;
  • the attributes of the image blocks used in the threshold segmentation processing of each image block are different, and the attributes include the number, size, shape, and/or selection position of the image block.
  • the number of the image blocks is at least two, and the at least two image blocks are obtained by evenly dividing the first blackboard image.
  • the threshold segmentation algorithm used in the step of performing pixel threshold segmentation processing on the first blackboard image is a global threshold segmentation algorithm, and/or the threshold segmentation algorithm for the first blackboard image
  • the threshold segmentation algorithm used in the step of threshold segmentation processing for at least one image block is a global threshold segmentation algorithm.
  • the step of performing threshold segmentation processing on pixels of the first blackboard image, so that the pixels of the first blackboard image are divided into first type pixels or second type pixels It includes:
  • the step of performing image block threshold segmentation processing at least once on the first blackboard image includes:
  • At least one image block threshold segmentation process and binarization process After performing at least one image block threshold segmentation process and binarization process on the first blackboard image, at least one third blackboard image is obtained;
  • threshold segmentation processing and binarization processing of the image block include:
  • the step of determining that the number of times that the pixel of the first blackboard image is classified into the first type of pixel is within a first threshold range, then the step of using the pixel as a target pixel includes:
  • the fourth blackboard image After the second blackboard image and the at least one third blackboard image are summed to obtain the fourth blackboard image, it is determined that the pixel value of the fourth blackboard image is within the second threshold range, then the This pixel is regarded as the target pixel;
  • the second threshold value range is determined according to the first threshold value range.
  • an embodiment of the present invention also provides an image processing system, including:
  • the first acquiring unit is configured to acquire the first blackboard image
  • a first processing unit configured to perform threshold segmentation processing on pixels of the first blackboard image, so that the pixels of the first blackboard image are divided into first-type pixels or second-type pixels;
  • the second processing unit is configured to perform threshold segmentation processing of image blocks at least once on the first blackboard image; wherein the threshold segmentation processing of the image blocks is specifically: performing at least one image of the first blackboard image Performing threshold segmentation processing of pixels on the block, so that the pixels of the at least one image block are divided into first type pixels or second type pixels;
  • the third processing unit is configured to determine that the number of times that the pixels of the first blackboard image are divided into the first type of pixels is within a first threshold range, and then the pixels are used as target pixels;
  • the fourth processing unit is configured to extract a target image from the first blackboard image by using the target pixel.
  • an embodiment of the present invention also provides an image processing device, which includes:
  • At least one processor At least one processor
  • At least one memory for storing at least one program
  • the at least one processor When the at least one program is executed by the at least one processor, the at least one processor implements the image processing method.
  • an embodiment of the present invention also provides a storage medium in which instructions executable by a processor are stored, and the instructions executable by the processor are used to execute the image processing method when executed by the processor.
  • an embodiment of the present invention provides a blackboard device, including a blackboard, a camera, and a terminal device connected to the camera;
  • the camera is used to photograph the blackboard
  • the terminal equipment includes:
  • At least one processor At least one processor
  • At least one memory for storing at least one program
  • the at least one processor When the at least one program is executed by the at least one processor, the at least one processor implements the image processing method.
  • the embodiment of the present invention performs threshold segmentation processing on the first blackboard image and performs pixel points on each image block in the first blackboard image. According to the threshold segmentation processing, it is possible to determine whether the pixel is the target pixel according to the number of times the pixel is divided into the first type of pixel. It can be seen that the embodiment of the present invention has the effect of local threshold segmentation, and the pixel is divided The target pixel is determined for the number of times of the first type of pixel. Compared with the traditional solution, this can effectively solve the problem of image reflection and greatly improve the accuracy of extracting the target image from the blackboard image.
  • Fig. 1 is a flowchart of steps in a first specific embodiment of an image processing method according to an embodiment of the present invention
  • Figure 2 is a schematic diagram of an image obtained after a graph is segmented using a traditional local threshold segmentation algorithm
  • FIG. 3 is a flowchart of steps in a second specific embodiment of an image processing method according to an embodiment of the present invention.
  • FIG. 4 is a block diagram of a specific embodiment of an image processing system according to an embodiment of the present invention.
  • FIG. 5 is a block diagram of a specific embodiment of an image processing apparatus according to an embodiment of the present invention.
  • Fig. 6 is a structural block diagram of a specific embodiment of a blackboard device according to an embodiment of the present invention.
  • step numbers in the following embodiments they are set only for ease of elaboration, and there is no limitation on the order between the steps.
  • the execution order of the steps in the embodiments can be adapted according to the understanding of those skilled in the art.
  • the "up”, “down”, “left”, “right”, “front”, and “rear” mentioned in the following embodiments are only used to clearly describe the positional relationship, and are relative positional relationships, not absolute The position relationship can be adjusted adaptively according to the understanding of those skilled in the art.
  • the image processing solution provided by the embodiment of the present invention can be applied to a smart blackboard scene, so as to extract a target object (such as handwriting on the blackboard) from the blackboard image.
  • a target object such as handwriting on the blackboard
  • the smart blackboard scene it mainly includes a blackboard device.
  • the blackboard device includes a blackboard, a camera, and a terminal device connected to the camera.
  • the camera can be set on the blackboard and used to photograph the blackboard.
  • the terminal device is mainly used to perform image processing on the blackboard image taken by the camera, so as to realize the operation of displaying and recording the blackboard image.
  • an embodiment of the present invention provides an image processing method for extracting a target image from a blackboard image, as shown in FIG. 1, and the steps involved are as follows.
  • the first blackboard image it may be directly the original blackboard image captured by a camera, or it may be a blackboard image obtained after image preprocessing is performed on the original blackboard image captured by the camera (wherein, the The image preprocessing can include, but is not limited to, image filtering, de-manipulation, morphological image processing, etc.).
  • the former has the advantages of simple steps and high processing efficiency, while the latter has the advantages of improving recognition accuracy.
  • the image area obtained by the positioning is used as the first blackboard image, which not only improves processing efficiency but also improves accuracy. Therefore, as to which blackboard image is acquired as the first blackboard image, this can be set and selected according to the actual situation, and there is no excessive limitation here.
  • S102 Perform threshold segmentation processing on pixels of the first blackboard image, so that the pixels of the first blackboard image are divided into first type pixels or second type pixels. Performing this step is equivalent to that the corresponding pixel in the first blackboard image has been subjected to a pixel type division process. If the pixel is divided into the first type of pixel, then the pixel is divided into the first type of pixel. The number of times of a type of pixel is increased by 1. If the pixel is classified into a second type of pixel, the number of times the pixel is classified into a second type of pixel is 1.
  • the specific processing steps are: threshold judgment is performed on the pixel points in the image one by one through the feature threshold, so that according to the judgment result, the pixel points in the image are divided into at least Two types of pixels. For example, when a certain pixel point i is within the threshold range corresponding to the type A pixel point, the pixel point i is classified as the type A pixel point.
  • the feature mentioned in the feature threshold it refers to the feature of the image, such as the grayscale and color of the image. Therefore, if the image subjected to the threshold segmentation process is different, the obtained feature threshold is also different. It must be the same, and how to obtain the feature threshold will vary according to the algorithm used.
  • the method for obtaining the feature threshold includes, but is not limited to, algorithms with the largest inter-class variance (such as the Otsu algorithm), The maximum variance threshold algorithm, the bimodal selection threshold algorithm, etc., are not too limited here. It can be seen that, for the step of performing the threshold segmentation processing on the pixels of the first blackboard image, so that the pixels of the first blackboard image are divided into the first type of pixels or the second type of pixels, the In order to threshold the pixels in the first blackboard image according to the characteristic threshold of the first blackboard image, and divide the pixels in the first blackboard image into the first type of pixels or the second type of pixels according to the judgment result point.
  • algorithms with the largest inter-class variance such as the Otsu algorithm
  • the maximum variance threshold algorithm such as the Otsu algorithm
  • the bimodal selection threshold algorithm etc.
  • the first type of pixel it is essentially the target type of pixel, and among all the pixels of the image, except for the first type of pixel, the pixel is the second type of pixel, for example, a total of There are three threshold ranges, namely Threshold Range 1, Threshold Range 2, and Threshold Range 3. Through these three threshold ranges, the pixels in the image can be divided into three types of pixels, and the pixel values fall into Threshold Range 1. The pixels in are type 1 pixels, the pixels whose pixel values fall within the threshold range 2 are type 2 pixels, and the pixels whose pixel values fall within the threshold range 3 are type 3 pixels.
  • type 1 pixel is Target type pixels
  • type 1 pixels are the first type pixels
  • type 2 pixels and type 3 pixels belong to the second type pixels
  • type 1 pixels and type 3 pixels are the target type pixels
  • type 1 pixel and type 3 pixel are the first type pixel
  • type 2 pixel is the second type pixel. If the number of pixel types to be divided is more, the same is true.
  • S103 Perform at least one image block threshold segmentation process on the first blackboard image; wherein the threshold segmentation process of the image block is specifically: performing pixel point segmentation on at least one image block of the first blackboard image. Threshold segmentation processing, so that the pixels of the at least one image block are divided into first type pixels or second type pixels. Regarding the threshold segmentation process, the first type pixel point, and the second type pixel point described in this step, they are the same as the analytical definition of the above step S102.
  • this step is mainly to perform threshold segmentation processing on at least one image block in the first blackboard image, that is, in this step, the object to be subjected to threshold segmentation processing is the image block in the first blackboard image Therefore, at this time, the characteristic threshold of the corresponding image block will be different from the characteristic threshold of the corresponding first blackboard image (of course, there may also be the case where the characteristic threshold of the corresponding image block is the same as the characteristic threshold of the corresponding first blackboard image), that is, this The type of pixel points will be different.
  • the threshold segmentation processing of the image block at least once, when the threshold segmentation processing of the image block is performed once, if the pixels in the image block (which are essentially the pixels in the first blackboard image) are divided into first Type pixel, the number of times that the pixel is divided into the first type pixel is increased by 1. If the pixel in the image block is divided into the second type pixel, then the pixel is divided into the second type pixel Increase the number of points by 1.
  • the pixel i in the first blackboard image is divided into the first type of pixel after the step S102 is performed, after the step S103 is performed (assuming that the threshold segmentation process of the image block is performed at least once at this time) Specifically, it is the threshold segmentation processing of the image block for three times), and it is determined that the pixel i is divided into the first type of pixel during the threshold segmentation processing of the three image blocks. At this time, the pixel i is divided into the first type of pixel.
  • the number of pixels of a type is 4.
  • the selection and acquisition methods from the first blackboard image can be selected according to the actual situation, and when the threshold segmentation processing of each image block is performed, the selection and acquisition methods of the image blocks can be the same. Not exactly the same or completely different.
  • the image block performs threshold segmentation processing, and then perform the remaining times of image block threshold segmentation processing, the selected acquisition position, size, and shape of the image block can be adjusted, for example, compared to the image block acquired the first time, the second time,
  • the selection and acquisition position of the 3rd and/or 4th image block can be moved up/down/left/right by the corresponding distance, or the size or shape of the image block can be adjusted, such as the size becomes larger or smaller, and the shape changes from square Or the rectangle becomes round or irregular. If the selection and acquisition method of the image block remains unchanged, the number of executions of the threshold segmentation processing of the image block may be 1.
  • the number of image blocks should be selected as multiple, and the multiple The extraction positions of the image blocks on the first blackboard image are set as evenly as possible everywhere in the first blackboard image.
  • the multiple image blocks are preferably obtained by dividing the first blackboard image.
  • the first blackboard image is divided into image regions, and the image regions obtained after the division are equivalent to image blocks. It can be seen that, for the step S103, it is equivalent to performing local threshold segmentation processing on the first blackboard image.
  • step S102 and step S103 the execution order between the two can be interchanged, that is, step S103 is executed first and then step S102 is executed.
  • the number of times the pixels in the first blackboard image are classified into the first type of pixels is counted, and then the counted times are thresholded to determine whether the times are If it falls within the first threshold range, it means that the pixel is the target pixel.
  • the type classification processing is the processing of dividing the pixel point i into a first type pixel point or a second type pixel point
  • the pixel point When the number of times i is divided into the first type pixel is greater than or equal to a preset threshold (such as 2), that is, when it falls within the first threshold range, it means that the pixel i is the target pixel.
  • a preset threshold such as 2
  • the total number of times, d is expressed as a percentage, d is greater than or equal to 50%; b is expressed as a value greater than or equal to c.
  • d it is also possible to set d to be less than 50%, but if d is greater than or equal to 50%, the judgment accuracy of the target pixel will be higher, which is suitable for scenes with unsatisfactory external environmental factors.
  • the target pixels extract a target image from the first blackboard image. If the target pixel corresponds to the foreground of the blackboard image, that is, the handwriting, then the extracted target image is the foreground image. If the target pixel corresponds to the background of the blackboard image, then the extracted target image is Background image.
  • the blackboard image is first subjected to the threshold segmentation processing of the entire image, and then the pixels of the first blackboard image are divided into the first type of pixels or the second type of pixels, and then After threshold segmentation is performed on the image block of the blackboard image, the pixels of the image block are divided into the first type pixel point or the second type pixel point to achieve the effect of local threshold segmentation of the blackboard image, and then pass the pixel point
  • the number of times that pixels are classified into the first type is thresholded to determine the target pixel. In this way, the determined target pixel is used to extract the corresponding image from the first blackboard image, that is, the target image. Its validity and accuracy High, and it can solve the problem of judging the foreground as the background due to the reflection of the image.
  • the at least one image block threshold segmentation process is specifically: at least two image block threshold segmentation processes; each of the at least two image block threshold segmentation processes
  • the attributes of the image blocks used in the threshold segmentation process are different, and the attributes include the number, size, shape, and/or selection position of the image block. Wherein, the difference means completely different or partly different.
  • the step of image block threshold segmentation processing is performed at least twice, and in each processing, the attribute parameters of the image block are not the same (that is, each time the image block threshold segmentation processing is performed When the image block selection and extraction method has been adjusted), this means that each time the threshold segmentation process of the image block is performed, for the threshold judgment of the same pixel, the characteristic threshold used will be different each time , Which is equivalent to the feature threshold value used is randomly changed. It can be seen that in this case, by counting the number of times the pixel is divided into the first type of pixel and the threshold judgment, the confirmation of the target pixel can be further improved Accuracy, thereby improving the accuracy and efficiency of target extraction.
  • the attributes of the image block will be different, so that the characteristic threshold will change randomly, so that even if one or two of the pixels belonging to the foreground are judged as The background pixel, but as long as the pixel is judged as the foreground pixel 3 or more times (for example, the pixel has undergone a total of 5 types of classification processing), then the pixel will eventually be judged as the foreground pixel pixel.
  • the method of this embodiment to achieve the extraction of the target image, its accuracy and effectiveness are further improved, and it can solve the problem that the traditional local threshold segmentation method is easy to judge the partial area of the filled figure as the background image. , That is, to solve the problem that the traditional local threshold segmentation method cannot handle the problem of whether the current neighborhood is divided into the global foreground or the global background.
  • the number of the image blocks is at least two, and the at least two image blocks are obtained by evenly dividing the first blackboard image. Since the handwriting written on the blackboard may spread all over the blackboard, it is difficult to determine the image area with reflective image area and/or the image area where the filled figure is located. Therefore, in order to better adapt to the blackboard scene, use the even division Way to obtain at least two image blocks, so that the foreground handwriting of the blackboard image can be obtained well. Moreover, the image blocks are obtained by dividing in an average manner, which can reduce the workload of program design and utilize the design, modification and adjustment of the staff.
  • the threshold segmentation algorithm used in the step of performing pixel threshold segmentation processing on the first blackboard image is a global threshold segmentation algorithm, and/or, the first blackboard image
  • the threshold segmentation algorithm used in the step of threshold segmentation processing for at least one image block of a blackboard image is a global threshold segmentation algorithm. That is equivalent to the step of performing pixel threshold segmentation processing on the first blackboard image, which specifically includes using a first threshold segmentation algorithm to perform pixel threshold segmentation processing on the first blackboard image;
  • the step of performing pixel threshold segmentation processing on at least one image block of the first blackboard image specifically includes using a second threshold segmentation algorithm to perform pixel pixel processing on at least one image block of the first blackboard image.
  • Point threshold segmentation processing wherein the first threshold segmentation algorithm and/or the second threshold segmentation algorithm are global threshold segmentation algorithms. Since in this embodiment, the first threshold segmentation algorithm and/or the second threshold segmentation algorithm adopts a global threshold segmentation algorithm, it can further reduce the traditional local threshold segmentation algorithm that can easily determine the partial area of the filled figure as a background. The problem of the image guarantees the accuracy of the segmentation of the foreground image and the background image.
  • the threshold segmentation processing is performed on the pixels of the first blackboard image, so that the pixels of the first blackboard image are divided into the first type of pixels or the second type of pixels
  • This step S102 includes:
  • the step S1021 is implemented by performing threshold segmentation processing and binarization processing on the first blackboard image, which not only achieves the pixel processing of the first blackboard image
  • the purpose of threshold segmentation is to mark the pixels classified as the first type as 255 or 1 (that is, the pixels of the first type are set as white pixels), and for the pixels classified as the second type Marked as 0 (that is, the second type of pixel is set as a black pixel), so that the pixel value after the pixel binarization can facilitate the subsequent statistics of the number of times the pixel is divided into the first type of pixel .
  • the pixel value range is 0 to 255, and the pixel value corresponding to the white pixel is 255; while in the color space represented by floating point numbers, the pixel value range It is 0 to 1, and the pixel value corresponding to the white pixel is 1.
  • the pixel value 1 should be selected to represent the white pixels, that is, the color space of the image should be represented by a floating point number, so that when the first blackboard image is determined
  • the pixel value of the pixel in the binarized image is 1, it means that in the current binarized image, the pixel is divided into the first type of pixel, that is, the foreground pixel, which also indicates the pixel
  • the number of times the pixels are classified into the first type plus one.
  • the step S103 of performing image block threshold segmentation processing on the first blackboard image at least once includes:
  • At least one third blackboard image is obtained.
  • the threshold segmentation processing and binarization processing of the image block include:
  • the above step S1031 is implemented by performing threshold segmentation processing and binarization processing on the at least one image block, which not only achieves the threshold segmentation processing of pixels on the image block It also realizes that the pixels classified into the first type are marked as 255 or 1 (that is, the pixels of the first type are set as white pixels), and the pixels classified as the second type are marked as 0 (That is, the second-type pixel is set as a black pixel). In this way, the pixel value after the pixel is binarized can facilitate subsequent statistics on the number of times the pixel is classified into the first-type pixel.
  • a pixel value of 1 should be optionally used to represent white pixels, that is, a floating point number should be selected to represent the color space of the image, so that when the pixel in the binary image of the image block is determined When the value is 1, it means that in the current binarized image, the pixel is divided into the first type of pixel, that is, the foreground pixel, and it also indicates the number of times the pixel is divided into the first type of pixel plus 1.
  • the second threshold range is determined according to the first threshold range.
  • the at least A third blackboard image is also a binarized image of the first blackboard image. Therefore, by performing pixel value summation processing on the second blackboard image and the at least one third blackboard image, it is equivalent to counting the number of times each pixel is judged to be the first type of pixel to obtain each pixel The total number of times H is judged to be the first type of pixel. At this time, a threshold judgment can be performed on the total number of times H, so as to finally determine whether the corresponding pixel is the target pixel.
  • the second threshold range is the same as the first threshold range. If the first type of pixel is not set as a white pixel, but with another color, that is, the second threshold The pixel value corresponding to one type of pixel is not 1. In this case, the first threshold range needs to be multiplied by a corresponding multiple to obtain the required second threshold range.
  • the number of executions of the threshold segmentation processing of the image block can be selected from 2 to 4 times.
  • the embodiment of the present invention provides an image processing method, which is specifically applied to the writing scene of a smart blackboard device, and is mainly used to extract the handwriting image (ie, the target foreground image) on the blackboard image.
  • the specific steps are As follows.
  • the first blackboard image is an original blackboard image captured by a camera or a blackboard image obtained after preprocessing the original blackboard image captured by a camera.
  • S202 Perform threshold segmentation processing on pixels of the first blackboard image, so that the pixels of the first blackboard image are divided into first type pixels or second type pixels.
  • the step S202 is specifically that the Otsu algorithm is used to obtain the optimal feature threshold of the first blackboard image, and then the optimal feature threshold is used to threshold the pixels in the first blackboard image one by one, thereby
  • the pixels in the first blackboard image are divided into foreground type pixels (i.e., first type pixels) and background type pixels (second type pixels), so as to realize the first blackboard image.
  • Global threshold segmentation then the pixels classified into the foreground type are set as white pixels, and the pixels classified as the background type are set as black pixels, so that the first blackboard image is binarized to obtain the first Two blackboard image.
  • the second blackboard image is a binary image, that is, the first segmented mask image (mask1).
  • S203 Perform at least one image block threshold segmentation process on the first blackboard image; wherein the threshold segmentation process of the image block is specifically: performing pixel point segmentation on at least one image block of the first blackboard image. Threshold segmentation processing, so that the pixels of the at least one image block are divided into first type pixels or second type pixels.
  • the number of executions n of the threshold segmentation processing of the image block can be selected as 3, and each time the threshold segmentation processing of the image block is executed, the size of the image block used is different each time, that is, the first
  • the size of the image block used in the second execution of the threshold segmentation process of the image block is size 1
  • the size of the image block used in the second execution of the threshold segmentation process of the image block is size 2
  • the size of the image block used in the segmentation process is size 3. Size 1, size 2 and size 3 are all different, and because the image block described in this embodiment is obtained by evenly dividing the first blackboard image Therefore, by adjusting the number of divided image blocks, the size of the image blocks can be adjusted.
  • threshold segmentation processing is performed on each image block separately, that is, at this time, the image block is used as the input image to perform the threshold segmentation processing of the pixels in the image block. That is to say, the optimal feature threshold obtained at this time is obtained after processing the image block, that is, threshold segmentation is performed on an image block, which can be selected as using Otsu algorithm to obtain the optimal feature of the image block Threshold, and then through the optimal feature threshold, threshold judgment is performed on the pixels in the image block one by one, so that according to the judgment result, the pixels in the image block are divided into foreground type pixels (i.e.
  • the sum of mask1 and mask_i is to add up the pixel values of the same pixel in the second blackboard image and the three third blackboard images.
  • the fourth blackboard image ie, the second mask image mask2
  • the value after the sum of the pixel values corresponding to the same pixel, such as pixel j in the fourth blackboard image, its pixel value is the pixel value of pixel j in mask1, the pixel value of pixel j in mask_1, mask_2
  • the threshold range used can be [2 , 4], at this time, after the fourth blackboard image is thresholded by using the threshold range, the pixels falling within the threshold range, that is, pixels with a pixel value greater than or equal to 2, are set to white
  • the pixel point, that is, the pixel value is set to 1
  • the pixel point that does not fall into the threshold range, that is, the pixel point with a pixel value less than 2 is set as a black pixel point, that is, the pixel value is set to 0, which is equivalent to
  • the fifth blackboard image is obtained, and the fifth blackboard image is the third mask image mask3.
  • the pixels at the corresponding positions are obtained from the first blackboard image.
  • the final required handwriting image since the pixel values of the foreground pixels in mask3 are all 1, according to the pixels with the pixel value of 1 in mask3, the pixels at the corresponding positions are obtained from the first blackboard image. The final required handwriting image.
  • this embodiment can achieve the local threshold segmentation of the first blackboard image by thresholding the image block, and at the same time, the attributes of the image block used are different each time the threshold segmentation process of the image block is performed, so as to achieve multiple Scale threshold segmentation, which can not only solve the problem of the blackboard image being unable to segment the foreground and background due to reflections, but also solve the traditional local threshold segmentation because the size of the selected neighborhood is too small to cover the entire image. Lead to the problem of segmenting a certain part of the graph into the background. It can be seen that by using the method of the embodiment of the present invention, the foreground and background can be segmented effectively and accurately, and it is very suitable for the scene of the smart blackboard device. Because the blackboard is easy to reflect light, the blackboard image often has a reflective area, and In class, I often draw filling figures on the blackboard.
  • an embodiment of the present invention also provides an image processing system, including:
  • the first acquiring unit 301 is configured to acquire a first blackboard image
  • the first processing unit 302 is configured to perform threshold segmentation processing on pixels of the first blackboard image, so that the pixels of the first blackboard image are divided into first type pixels or second type pixels;
  • the second processing unit 303 is configured to perform threshold segmentation processing of an image block at least once on the first blackboard image; wherein the threshold segmentation processing of the image block is specifically performed on at least one of the first blackboard image Performing threshold segmentation processing of pixels on the image block, so that the pixels of the at least one image block are divided into first type pixels or second type pixels;
  • the third processing unit 304 is configured to determine that the number of times that the pixels of the first blackboard image are classified into the first type of pixels is within a first threshold range, and then use the pixel as a target pixel;
  • the fourth processing unit 305 is configured to use the target pixels to extract a target image from the first blackboard image.
  • the blackboard image is first subjected to the threshold segmentation processing of the entire image, and then the pixels of the first blackboard image are divided into the first type of pixels or the second type of pixels, and then After threshold segmentation is performed on the image block of the blackboard image, the pixels of the image block are divided into the first type pixel point or the second type pixel point to achieve the effect of local threshold segmentation of the blackboard image, and then pass the pixel point
  • the number of times that pixels are classified into the first type is thresholded to determine the target pixel. In this way, the determined target pixel is used to extract the corresponding image from the first blackboard image, that is, the target image. Its validity and accuracy High, and it can solve the problem of judging the foreground as the background due to the reflection of the image.
  • the threshold segmentation processing of the image block at least once is specifically: the threshold segmentation processing of the image block at least twice.
  • the attributes of the image blocks used in the threshold segmentation processing of each image block are different, and the attributes include the number, size, shape, and/or selection position of the image block.
  • the number of the image blocks is at least two, and the at least two image blocks are obtained by evenly dividing the first blackboard image.
  • the threshold segmentation algorithm used in the step of performing pixel threshold segmentation processing on the first blackboard image is a global threshold segmentation algorithm, and/or, the first blackboard image
  • the threshold segmentation algorithm used in the step of threshold segmentation processing for at least one image block of a blackboard image is a global threshold segmentation algorithm.
  • the first processing unit 302 includes:
  • the first processing module is configured to perform threshold segmentation processing on pixels of the first blackboard image, so that the pixels of the first blackboard image are divided into first type pixels or second type pixels, and then The pixels of the first type are set as white pixels, and the pixels of the second type are set as black pixels, thereby obtaining a second blackboard image.
  • the second processing unit 303 includes:
  • the second processing module is configured to perform at least one image block threshold segmentation process and binarization process on the first blackboard image to obtain at least one third blackboard image.
  • the threshold segmentation processing and binarization processing of the image block include:
  • the third processing unit 304 includes:
  • the third processing module is configured to perform sum processing on the second blackboard image and the at least one third blackboard image to obtain the fourth blackboard image, and then determine that the pixel value of the fourth blackboard image is in the first Within the second threshold range, the pixel is taken as the target pixel; wherein, the second threshold range is determined according to the first threshold range.
  • an embodiment of the present invention also provides an image processing device, which includes:
  • At least one processor 401 At least one processor 401;
  • the at least one processor 401 When the at least one program is executed by the at least one processor 401, the at least one processor 401 implements the steps of the image processing method described in the foregoing method embodiment.
  • the blackboard image is first subjected to the threshold segmentation processing of the entire image, and then the pixels of the first blackboard image are divided into the first type of pixels or the second type of pixels, and then After threshold segmentation is performed on the image block of the blackboard image, the pixels of the image block are divided into the first type pixel point or the second type pixel point to achieve the effect of local threshold segmentation of the blackboard image, and then pass the pixel point
  • the number of times that pixels are classified into the first type is thresholded to determine the target pixel. In this way, the determined target pixel is used to extract the corresponding image from the first blackboard image, that is, the target image.
  • an embodiment of the present invention also provides a storage medium in which instructions executable by a processor are stored, and the instructions executable by the processor are used to execute the one described in the foregoing method embodiment when executed by the processor.
  • This image processing method steps That is to say, the content in the above method embodiment is applicable to this storage medium embodiment, and the specific functions implemented by this storage medium embodiment are the same as the above method embodiment, and the beneficial effects achieved are the same as those achieved by the above method embodiment. The beneficial effects are also the same.
  • an embodiment of the present invention also provides a blackboard device, including a blackboard 501, a camera 502, and a terminal device 503 connected to the camera 502.
  • the camera 502 is used to photograph the blackboard 501.
  • the terminal device 503 includes:
  • At least one processor At least one processor
  • At least one memory is used to store at least one program.
  • the at least one processor When the at least one program is executed by the at least one processor, the at least one processor implements the steps of an image processing method described in the foregoing method embodiment.
  • the entire image of the blackboard image is first subjected to threshold segmentation processing, and then the pixels of the first blackboard image are divided into first type pixels or second type pixels, and then After threshold segmentation is performed on the image block of the blackboard image, the pixels of the image block are divided into the first type of pixel or the second type of pixel to achieve the effect of the local threshold segmentation of the blackboard image.
  • the number of times the pixels are divided into the first type is used for threshold judgment to determine the target pixel.
  • the determined target pixel is used to extract the corresponding image from the first blackboard image, that is, the target image, which is highly effective and accurate , And can solve the problem of judging the foreground as the background due to the reflection of the image.
  • the terminal device 503 it is realized by a combination of software and hardware, and it can be a computer, a mobile phone, an interactive smart tablet, a display device with smart processing functions (such as a smart TV, a smart display), and other devices.
  • the memory it may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage devices.
  • processor and the memory can be connected by a bus, and the processor and the memory can be integrated in the same circuit board or independently arranged in different circuit boards, and the connection between the processor and the memory can be
  • the fixed and non-detachable connection can also be a detachable connection.
  • the communication method between the terminal device 503 and the camera 502 can be a wired connection (such as a serial port wired connection, a universal serial bus (Universal Serial Bus, USB) interface wired connection, etc.), or a wireless connection (such as Infrared, Bluetooth, Zigbee, wireless fidelity (Wireless Fidelity, Wifi, etc.), these communication connection modes are not too limited in this embodiment, and can be selected according to actual conditions/requirements.
  • a wired connection such as a serial port wired connection, a universal serial bus (Universal Serial Bus, USB) interface wired connection, etc.
  • a wireless connection such as Infrared, Bluetooth, Zigbee, wireless fidelity (Wireless Fidelity, Wifi, etc.
  • the camera 502 is directly arranged on the blackboard 501 (for example, above the blackboard 501), so that the writing board can be directly installed during use, which is convenient for operation and use.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Image Analysis (AREA)

Abstract

L'invention concerne un procédé, un système et un dispositif de traitement d'image, un support d'informations et un dispositif de tableau noir. Ledit procédé consiste : à acquérir une première image de tableau noir (S101) ; à effectuer un traitement de segmentation de seuil d'un point de pixel sur la première image de tableau noir (S102) ; à effectuer au moins un traitement de segmentation de seuil d'un bloc d'image sur la première image de tableau noir (S103) ; s'il est déterminé que le nombre de fois où le point de pixel de la première image de tableau noir est classé dans un premier type de points de pixel se trouve dans une première plage de seuil, à prendre le point de pixel comme point de pixel cible (S104) ; et à utiliser le point de pixel cible afin d'extraire une image cible de la première image de tableau noir (S105).
PCT/CN2019/127758 2019-02-20 2019-12-24 Procédé, système et dispositif de traitement d'image, support d'informations et dispositif de tableau noir WO2020168818A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910127218.6A CN109978889A (zh) 2019-02-20 2019-02-20 一种图像处理方法、系统、装置、存储介质及黑板装置
CN201910127218.6 2019-02-20

Publications (1)

Publication Number Publication Date
WO2020168818A1 true WO2020168818A1 (fr) 2020-08-27

Family

ID=67077138

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/127758 WO2020168818A1 (fr) 2019-02-20 2019-12-24 Procédé, système et dispositif de traitement d'image, support d'informations et dispositif de tableau noir

Country Status (2)

Country Link
CN (1) CN109978889A (fr)
WO (1) WO2020168818A1 (fr)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109978889A (zh) * 2019-02-20 2019-07-05 广州视源电子科技股份有限公司 一种图像处理方法、系统、装置、存储介质及黑板装置
CN118072192B (zh) * 2024-04-25 2024-06-21 烟台哈尔滨工程大学研究院 基于gee的大型藻有害藻华长时序自动提取方法

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101699469A (zh) * 2009-11-09 2010-04-28 南京邮电大学 课堂录像中教师黑板书写动作的自动识别方法
CN103019466A (zh) * 2012-11-16 2013-04-03 厦门大学 基于红外探测的投影互动系统
US20150098648A1 (en) * 2013-10-09 2015-04-09 Casio Computer Co., Ltd. Image Processing Device Capable of Segmenting an Image into Plural Areas
CN107809563A (zh) * 2017-11-30 2018-03-16 浙江大华技术股份有限公司 一种板书检测系统、方法及装置
CN108550157A (zh) * 2018-04-04 2018-09-18 安徽海青智能科技有限公司 一种教学板书的无遮挡处理方法
CN109978889A (zh) * 2019-02-20 2019-07-05 广州视源电子科技股份有限公司 一种图像处理方法、系统、装置、存储介质及黑板装置

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9740957B2 (en) * 2014-08-29 2017-08-22 Definiens Ag Learning pixel visual context from object characteristics to generate rich semantic images
CN107945200B (zh) * 2017-12-14 2021-08-03 中南大学 图像二值化分割方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101699469A (zh) * 2009-11-09 2010-04-28 南京邮电大学 课堂录像中教师黑板书写动作的自动识别方法
CN103019466A (zh) * 2012-11-16 2013-04-03 厦门大学 基于红外探测的投影互动系统
US20150098648A1 (en) * 2013-10-09 2015-04-09 Casio Computer Co., Ltd. Image Processing Device Capable of Segmenting an Image into Plural Areas
CN107809563A (zh) * 2017-11-30 2018-03-16 浙江大华技术股份有限公司 一种板书检测系统、方法及装置
CN108550157A (zh) * 2018-04-04 2018-09-18 安徽海青智能科技有限公司 一种教学板书的无遮挡处理方法
CN109978889A (zh) * 2019-02-20 2019-07-05 广州视源电子科技股份有限公司 一种图像处理方法、系统、装置、存储介质及黑板装置

Also Published As

Publication number Publication date
CN109978889A (zh) 2019-07-05

Similar Documents

Publication Publication Date Title
WO2020199776A1 (fr) Procédé et dispositif de calcul de taux de vacance d'étagère et support d'informations
WO2019233264A1 (fr) Procédé de traitement d'image, support de stockage lisible par ordinateur et dispositif électronique
US9071745B2 (en) Automatic capturing of documents having preliminarily specified geometric proportions
TWI619080B (zh) 指紋重疊區域面積的計算方法及電子裝置
US9305360B2 (en) Method and apparatus for image enhancement and edge verification using at least one additional image
WO2017215527A1 (fr) Procédé, dispositif et support de stockage informatique de détection de scénario hdr
WO2018120038A1 (fr) Procédé et dispositif de détection de cible
CN110796600B (zh) 一种图像超分重建方法、图像超分重建装置及电子设备
CN108769634B (zh) 一种图像处理方法、图像处理装置及终端设备
US10885321B2 (en) Hand detection method and system, image detection method and system, hand segmentation method, storage medium, and device
WO2022166865A1 (fr) Procédé et appareil d'élimination d'ombre pour image de texte, et dispositif électronique
US9401027B2 (en) Method and apparatus for scene segmentation from focal stack images
CN104486552A (zh) 一种获取图像的方法及电子设备
WO2021195873A1 (fr) Procédé et dispositif d'identification d'une région d'intérêt dans une image graphique d'essai sfr, et support
WO2017088637A1 (fr) Procédé et appareil permettant de localiser un contour d'image dans un fond naturel
WO2020168818A1 (fr) Procédé, système et dispositif de traitement d'image, support d'informations et dispositif de tableau noir
WO2020232910A1 (fr) Procédé et appareil de comptage de cibles basés sur un traitement d'image, dispositif et support de stockage
US20130201358A1 (en) Efficient Line Detection Method
CN111294516A (zh) 一种矾花图像处理方法、系统、电子设备及介质
CN105049706A (zh) 一种图像处理方法及终端
CN113283439A (zh) 基于图像识别的智能计数方法、装置及系统
CN111695374B (zh) 监控视角中斑马线区域的分割方法、系统、介质及设备
CN117611483A (zh) 一种图像采集方法、装置、设备及存储介质
WO2017128646A1 (fr) Procédé et dispositif de traitement d'image
WO2024016686A1 (fr) Procédé et appareil de détection de coins

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19916153

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19916153

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 19916153

Country of ref document: EP

Kind code of ref document: A1