A scanning filling point recognition method and system based on weight distribution

Abstract

The present invention relates to the field of image scanning and recognition technology, and specifically to a scanning filling point recognition method and system based on weight distribution. The method includes obtaining a blank answer sheet scanned image, and establishing an answer sheet scanning template based on the blank answer sheet scanned image; obtaining a candidate's answer sheet scanned image; obtaining a rectangular filling frame in the candidate's answer sheet scanned image based on the answer sheet scanning template; assigning values ​​to the pixel points in the rectangular filling frame according to a preset pixel value assignment strategy; obtaining the area of ​​each rectangular filling frame, and performing equal area segmentation on the rectangular filling frame to obtain a number of segmentation modules; counting the pixel values ​​of the segmentation modules; calculating the recognition coefficient of the rectangular filling frame based on the preset segmentation module weight coefficient and the segmentation module pixel value, and when the recognition coefficient is less than the preset threshold, the rectangular filling frame is determined to be filled. The present invention can more quickly identify the filling information of objective questions through the weight distribution of filling points and image multi-valued processing, with a high accuracy rate.

Description

Technical Field

[0001] The present invention relates to the technical field of image scanning and recognition, and in particular to a scanning and filling point recognition method and system based on weight distribution. Background Art

[0002] With the development of education, computer marking is becoming more and more popular. The current online marking system based on paper-based examinations grays the paper answer sheets of examinees to obtain processed images, and then identifies the objective question filling area of ​​the processed image and determines whether the filling is valid. Patent CN112733644A provides a filling point recognition system and method based on scanning recognition, which can quickly identify the filling point information of objective questions without a lot of manual intervention, thereby improving work efficiency and accuracy.

[0003] However, the recognition accuracy of the above recognition method is still somewhat insufficient, for example, valid filling points may be judged as invalid filling points.

[0004] Therefore, it is urgent to develop a new scanning and filling point recognition method to improve recognition accuracy and reduce manual intervention. Summary of the invention

[0005] In order to solve the problems in the above-mentioned background technology, the present invention provides a scanning filling point recognition method and system based on weight distribution. Through the weight distribution of filling points and image multi-valued processing, the filling point information of objective questions can be recognized more quickly with high accuracy.

[0006] The first object of the present invention is to provide a scanning filling point recognition method based on weight distribution, comprising the following steps:

[0007] Obtain a scanned image of a blank answer sheet, and create an answer sheet scanning template according to the scanned image of the blank answer sheet;

[0008] Obtain a scanned image of the examinee's answer sheet;

[0009] According to the answer sheet scanning template, a rectangular filling frame in the scanned image of the examinee's answer sheet is obtained;

[0010] Assign values ​​to the pixel points in the rectangular filling frame according to the preset pixel value assignment strategy;

[0011] Obtain the area of ​​each rectangular filling frame, and divide the rectangular filling frame into equal-area segments to obtain a number of segmentation modules;

[0012] Count the pixel values ​​of the segmentation module;

[0013] The recognition coefficient of the rectangular fill frame is calculated by weighted summing up the preset segmentation module weight coefficient and the pixel value of the segmentation module, and the segmentation module weight coefficient at the center of the rectangular fill frame is the largest. When the recognition coefficient is less than the preset threshold, the rectangular fill frame is judged to be filled.

[0014] In one embodiment, counting pixel values ​​of the segmentation module includes:

[0015] The pixel value corresponding to the largest number of pixels in the segmentation module is selected as the pixel value in the segmentation module.

[0016] In other embodiments, counting pixel values ​​of the segmentation module includes:

[0017] Divide the segmentation module into several pixel groups;

[0018] Count the number of pixels with the same pixel value in a pixel group;

[0019] Select the pixel value corresponding to the largest number of pixels as the pixel value of the pixel group;

[0020] Count the number of pixel groups with the same pixel value in the segmentation module;

[0021] The pixel value corresponding to the pixel group with the largest number is selected as the pixel value of the segmentation module.

[0022] Furthermore, the pixel value corresponding to the largest number of pixel points is selected as the pixel value of the pixel group, including:

[0023] There are at least two groups of pixel points with the same and maximum number, and the smallest pixel value among the multiple groups is selected as the pixel value of the pixel group.

[0024] Further, the pixel value corresponding to the pixel group with the largest number is selected as the pixel value of the segmentation module, including:

[0025] There are at least two groups of pixel groups with the same and largest number, and the pixel value of the smallest pixel group among the multiple groups is selected as the pixel value of the segmentation module.

[0026] Further, a blank answer sheet scan image is obtained, and an answer sheet scan template is established according to the blank answer sheet scan image, including:

[0027] On the scanned image of the blank answer sheet, select the positioning mark points and the rectangular fill box, mark the coordinate position of the selected rectangular fill box, and generate an XML file answer sheet scan template.

[0028] Further, according to the answer sheet scanning template, a rectangular filling frame in the scanned image of the examinee's answer sheet is obtained, including:

[0029] Obtain the answer sheet scan template and the examinee's answer sheet scan image;

[0030] According to the positioning mark points, the scanned image of the examinee's answer sheet and the scanned image of the blank answer sheet are placed in the same coordinate system;

[0031] According to the coordinate position of the rectangular filling frame of the answer sheet scanning template, the position information of the rectangular filling frame in the examinee's answer sheet scanned image is obtained.

[0032] Furthermore, the pixel points in the rectangular filling frame are assigned values ​​according to a preset pixel value assignment strategy, including:

[0033] Divide the grayscale value range of the rectangular fill frame into N equal parts.

[0034] Divide the grayscale value range of the rectangular filling frame from 0 to n into N equal-length grayscale value intervals;

[0035] With a step size of (n+1) / N, select N grayscale values ​​in the grayscale value range of 0 to n, and make the N grayscale values ​​correspond to the N grayscale value intervals one by one;

[0036] Revalue the pixel value of the pixel point in the rectangular filling frame so that the pixel value of the pixel point is equal to the gray value corresponding to the gray value interval.

[0037] Furthermore, the recognition coefficient of the rectangular filling frame is calculated according to the preset segmentation module weight coefficient, including:

[0038] The pixel values ​​of several segmentation modules are counted as A1, A2, A3...A n ;

[0039] The weight coefficients of several segmentation modules are B1, B2, B3…B n ;

[0040] Define the recognition coefficient of the rectangular filling frame as R;

[0041] Then R = A1*B1+A2*B2+...+A n *B n ;

[0042] A second object of the present invention is to provide a scanning filling point recognition system based on weight distribution, comprising:

[0043] A scanning template making module is used to obtain a scanned image of a blank answer sheet and to create a scanning template for the answer sheet based on the scanned image of the blank answer sheet;

[0044] An acquisition module obtains a scanned image of the examinee's answer sheet;

[0045] The answer sheet information recognition module is used to obtain the rectangular filling frame in the scanned image of the examinee's answer sheet according to the answer sheet scanning template;

[0046] A processing module, used for assigning values ​​to the pixel points in the rectangular filling frame according to a preset pixel value assignment strategy;

[0047] A segmentation module is used to obtain the area of ​​each rectangular filling frame and segment the rectangular filling frame into equal areas to obtain a plurality of segmentation modules;

[0048] A statistics module is used to count the pixel values ​​of the segmentation module; wherein: counting the pixel values ​​of the segmentation module includes:

[0049] Selecting the pixel value corresponding to the largest number of pixel points in the segmentation module as the pixel value of the segmentation module; or

[0050] Dividing the segmentation module into a plurality of pixel groups;

[0051] Counting the number of pixels with the same pixel value in the pixel group;

[0052] Select the pixel value corresponding to the largest number of pixels as the pixel value of the pixel group;

[0053] Counting the number of pixel groups with the same pixel value in the segmentation module;

[0054] Select the pixel value corresponding to the pixel group with the largest number as the pixel value of the segmentation module;

[0055] The calculation module is used to calculate the recognition coefficient of the rectangular fill frame by weighted summing up the preset segmentation module weight coefficient and the pixel value of the segmentation module, and the segmentation module weight coefficient at the center of the rectangular fill frame is the largest. When the recognition coefficient is less than the preset threshold, the rectangular fill frame is judged to be filled.

[0056] The present invention provides a scanning filling point recognition method based on weight distribution. The recognition coefficient of the rectangular filling box is calculated according to the preset segmentation module weight coefficient. By comparing the recognition coefficient with the preset threshold, the filling validity of the rectangular filling box of the objective question can be identified more quickly. The weight coefficient calculation method improves the accuracy of scanning recognition, and there are few situations where inaccurate recognition requires manual intervention, which effectively ensures the fairness of the test results. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0058] Figure 1 This is a flow chart (I) of a scanning filling point recognition method based on weight distribution according to an embodiment of the present application;

[0059] Figure 2 This is a flow chart (II) of the scanning filling point recognition method based on weight distribution according to an embodiment of the present application;

[0060] Figure 3 This is a flow chart of the scanning filling point recognition method based on weight distribution according to an embodiment of the present application (III);

[0061] Figure 4 This is a module composition diagram of the scanning filling point recognition system based on weight distribution in an embodiment of the present application. DETAILED DESCRIPTION

[0062] The following will be combined with the accompanying drawings in the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the protection scope of the present invention.

[0063] Figure 1 , Figure 2 The flowcharts (I) and (II) of a scanning filling point recognition method based on weight distribution in some embodiments of the present application are as follows. The recognition method comprises the following steps:

[0064] S101, obtaining a scanned image of a blank answer sheet, and establishing an answer sheet scan template according to the scanned image of the blank answer sheet;

[0065] S102, obtaining a scanned image of the examinee's answer sheet;

[0066] S103, obtaining a rectangular filling frame in the scanned image of the examinee's answer sheet according to the answer sheet scan template;

[0067] S104, assigning values ​​to the pixel points in the rectangular filling frame according to a preset pixel value assignment strategy;

[0068] S105, obtaining the area of ​​each rectangular filling frame, and dividing the rectangular filling frame into equal-area segments to obtain a plurality of segmentation modules;

[0069] S106, counting pixel values ​​of the segmentation module;

[0070] S107, calculating the recognition coefficient of the rectangular fill frame by weighted summing up the preset segmentation module weight coefficient and the pixel value of the segmentation module, and when the recognition coefficient is less than a preset threshold, determining the rectangular fill frame as being filled.

[0071] Among them, in step S101, on the scanned image of the blank answer sheet, the positioning mark points and the rectangular filling box are selected, the coordinate position of the selected rectangular filling box is marked, and an XML file answer sheet scanning template is generated and stored in the data server for batch image recognition of the examinees' answer sheets.

[0072] In step S102, a high-speed scanner is used to batch scan the examinees' answer sheets into examinees' answer sheet scan images, and the obtained examinees' answer sheet scan images are stored in a data server.

[0073] In step S103, the answer sheet scan template and the examinee's answer sheet scan image are obtained from the data server, and according to the positioning mark points in step S101, the examinee's answer sheet scan image and the blank answer sheet scan image are placed in the same coordinate system, and according to the coordinate position of the rectangular fill-in box of the answer sheet scan template, the position information of the rectangular fill-in box in the examinee's answer sheet scan image is obtained.

[0074] In step S104, the image of the examinee's answer sheet is converted into an image in BMP format with n grayscale levels, and the grayscale value range of the rectangular fill box is divided into N equal parts (N value is generally an integer between 2 and 5, and N value can be adjusted according to business conditions), and the cutting point is N-1. The grayscale value range of the rectangular fill box from 0 to n is divided into N equal-length grayscale value intervals; with a step size of (n+1) / N, N grayscale values ​​are selected in the grayscale value range from 0 to n, and the N grayscale values ​​are made to correspond one to one with the N grayscale value intervals; the pixel value of the pixel in the rectangular fill box is re-valued, and the pixel value of the pixel is made equal to the grayscale value corresponding to the grayscale value interval.

[0075] For example: the grayscale value range of the rectangular fill box is 0 to 255. These 256 grayscale values ​​are divided into N equal parts. There are (N-1) cutting points, and the values ​​from small to large are 255 / N, 2*255 / N, 3*255 / N, ..., (N-1)*255 / N (the cutting point values ​​are rounded to integers). Then the numerical ranges of the N segments are 0~255 / N, 255 / N~2*255 / N, 2*255 / N~3*255 / N, ..., (N-2)*255 / N~(N-1)*255 / N, (N-1)*255 / N~255, and the first and last two numerical values ​​0~255 / N and (N-1)*255 / N~255 are 0 and 255 respectively, and the remaining (N-2) numerical ranges are the middle value of their respective numerical ranges (510*N-765) / (2*N) (rounded to the nearest integer). For example, the value of N is 3, and it is evenly cut into 3 segments. The number of cutting points is 2, and the values ​​of the cutting points are 85 and 170 respectively, which are further divided into 3 value ranges, namely 0-85, 85-170, and 170-255. The first and last value ranges 0-85 and 170-255 are 0 and 255 respectively, and the remaining value range is the middle value 128 of the value range (rounded to the integer).

[0076] In step S105, the coordinate positions of the four corners of a single rectangular fill frame are obtained, the horizontal and vertical lengths of the rectangular fill frame are calculated, the area of ​​the rectangular fill frame is calculated, and the area is divided into Y equal parts, and cut into Y modules of the same area size (the Y value is generally an integer between 2 and 5, and the Y value can be adjusted according to business conditions). If the Y value is 3, then they are segmentation module 1, segmentation module 2, and segmentation module 3 respectively.

[0077] Step S106 includes step S1061: selecting the pixel value corresponding to the largest number of pixel points in the segmentation module as the pixel value in the segmentation module.

[0078] In step S107, the pixel values ​​of several segmentation modules are counted as A1, A2, A3, ... n ; The weight coefficients of several segmentation modules are B1, B2, B3…B n ; Define the recognition coefficient of the rectangular filling frame as R; then R= A1*B1+A2*B2+...+A n *B n Among them, B n-1 The value is B nThe weight coefficient of the segmentation module in the center of the rectangular filling box is the largest. If the R value obtained is greater than the preset threshold, it is determined that the rectangular filling box is not filled; if the R value is less than the preset threshold, it is determined that the rectangular filling box is filled. The preset threshold can be adjusted according to business conditions.

[0079] Reference Figure 3 , Figure 3 This is a flow chart (III) of a scanning filling point recognition method based on weight distribution in some embodiments of the present application. This embodiment is different from the above embodiment in that the method of counting pixel groups of the segmentation module is different, specifically:

[0080] In step S106, this embodiment includes step S1062: dividing the segmentation module into Z small rectangular blocks of the same shape (the value of Z can be adjusted according to business conditions), each small rectangular block is equivalent to a pixel group, counting the number of pixel points with the same pixel value in the pixel group, and selecting the pixel value corresponding to the pixel point with the largest number as the pixel value of the pixel group.

[0081] Furthermore, if there are at least two groups of pixel points with the same number and the largest number, the smallest pixel value among the multiple groups is selected as the pixel value of the pixel group.

[0082] Step S1063: Count the pixel values ​​of the small rectangular blocks in the segmentation module, that is, the pixel values ​​of the pixel groups; count the number of pixel groups with the same pixel value in the segmentation module; select the pixel value corresponding to the pixel group with the largest number as the pixel value of the segmentation module.

[0083] Furthermore, if there are at least two groups of pixel groups with the same number and the largest number, the pixel value of the smallest pixel group among the multiple groups is selected as the pixel value of the segmentation module.

[0084] The pixel value A of each segmentation module is counted above.

[0085] For example: if the number of 0 pixel values ​​in a single small rectangular block is the largest, then mark the small rectangular block as a 0 pixel value rectangular block, that is, the pixel value of the pixel group is 0. The next step is to count the number of rectangular blocks with the same pixel value in the segmentation module. If the number of 0 pixel value rectangular blocks is the largest, then take 0 pixel value as the pixel value of the segmentation module. Similarly, according to the above method, count the pixel values ​​of several segmentation modules.

[0086] From the above, it can be seen that the scanning filling point recognition method based on weight distribution provided in the embodiment of the present application, through the weight distribution of filling points and image multi-valued processing, can not only accurately and quickly identify the filling point information of objective questions without a lot of manual intervention, but also can greatly improve work efficiency and accuracy, effectively ensure the seriousness, authority and fairness of the examination, and ensure the smooth progress of the scanning work.

[0087] Reference Figure 4 , Figure 4 This is a module composition diagram of a scanning filling point recognition system based on weight distribution in some embodiments of the present application, including:

[0088] The scanning template making module 10 is used to obtain a scanned image of a blank answer sheet and to create a scanning template of the answer sheet according to the scanned image of the blank answer sheet;

[0089] An acquisition module 20 acquires a scanned image of the examinee's answer sheet;

[0090] The answer sheet information recognition module 30 is used to obtain the rectangular filling frame in the scanned image of the examinee's answer sheet according to the answer sheet scanning template;

[0091] The processing module 40 is used to assign values ​​to the pixel points in the rectangular filling frame according to a preset pixel value assignment strategy;

[0092] The segmentation module 50 is used to obtain the area of ​​each rectangular filling frame and segment the rectangular filling frame into equal areas to obtain a plurality of segmentation modules;

[0093] The statistical module 60 is used to count the pixel values ​​of the segmentation module; wherein: counting the pixel values ​​of the segmentation module includes:

[0094] Selecting the pixel value corresponding to the largest number of pixel points in the segmentation module as the pixel value of the segmentation module; or

[0095] Dividing the segmentation module into a plurality of pixel groups;

[0096] Counting the number of pixels with the same pixel value in the pixel group;

[0097] Select the pixel value corresponding to the largest number of pixels as the pixel value of the pixel group;

[0098] Counting the number of pixel groups with the same pixel value in the segmentation module;

[0099] Select the pixel value corresponding to the pixel group with the largest number as the pixel value of the segmentation module;

[0100] The calculation module 70 is used to calculate the recognition coefficient of the rectangular fill frame according to the preset segmentation module weight coefficient and the pixel value of the segmentation module, and the segmentation module weight coefficient at the center of the rectangular fill frame is the largest. When the recognition coefficient is less than the preset threshold, the rectangular fill frame is judged to be filled.

[0101] The relevant functions of the scanning and filling point recognition system based on weight distribution of the present invention can be referred to the aforementioned embodiment description of the scanning and filling point recognition method based on weight distribution, which will not be repeated here.

[0102] The scanning filling point recognition system based on weight distribution in the embodiment of the present invention automatically recognizes the answer sheet of the examinee, and then accurately calculates the filling information of the rectangular filling box of the objective question on the examinee's answer sheet, and automatically calculates the recognition coefficient according to the weight coefficient of the segmentation module to determine whether the rectangular filling box is filled, without the need for a large amount of manual intervention, thereby greatly improving work efficiency and accuracy, thereby effectively ensuring the fairness and authority of marking.

[0103] In the embodiments provided in the present application, it should be understood that the disclosed systems and methods can be implemented in other ways. The system embodiments described above are merely schematic. For example, the division of modules is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some communication interface, device or unit, which can be electrical, mechanical or other forms.

[0104] The present invention is further described above with the aid of specific embodiments. However, it should be understood that the specific description here should not be construed as limiting the essence and scope of the present invention. Various modifications made to the above embodiments by ordinary technicians in the field after reading this specification are all within the scope of protection of the present invention.

Claims

1. A scanning filling point recognition method based on weight distribution, characterized in that: The following steps are involved: Obtain a scanned image of a blank answer sheet, and create an answer sheet scan template based on the scanned image of the blank answer sheet; Obtain a scanned image of the examinee's answer sheet; According to the answer sheet scanning template, a rectangular filling frame in the scanned image of the examinee's answer sheet is obtained; Assigning values ​​to the pixel points within the rectangular filling frame according to a preset pixel value assignment strategy; Obtaining the area of ​​each rectangular filling frame, and dividing the rectangular filling frame into equal-area segments to obtain a plurality of segmentation modules; Counting the pixel values ​​of the segmentation module includes: selecting the pixel value corresponding to the largest number of pixel points in the segmentation module as the pixel value of the segmentation module; or Dividing the segmentation module into a plurality of pixel groups; Counting the number of pixels with the same pixel value in the pixel group; Select the pixel value corresponding to the largest number of pixels as the pixel value of the pixel group; Counting the number of pixel groups with the same pixel value in the segmentation module; Select the pixel value corresponding to the pixel group with the largest number as the pixel value of the segmentation module; The recognition coefficient of the rectangular fill box is calculated by weighted summing up the preset segmentation module weight coefficient and the pixel value of the segmentation module, and the segmentation module weight coefficient at the center of the rectangular fill box is the largest. When the recognition coefficient is less than the preset threshold, the rectangular fill box is judged to be filled.

2. The scanning filling point recognition method based on weight distribution according to claim 1 is characterized in that: The pixel value corresponding to the largest number of pixels is selected as the pixel value of the pixel group, including: There are at least two groups of pixel points with the same and maximum number, and the smallest pixel value among the multiple groups is selected as the pixel value of the pixel group.

3. The scanning filling point recognition method based on weight distribution according to claim 1 is characterized in that: The pixel values ​​corresponding to the pixel group with the largest number are selected as the pixel values ​​of the segmentation module, including: There are at least two groups of pixel groups with the same and largest number, and the pixel value of the smallest pixel group among the multiple groups is selected as the pixel value of the segmentation module.

4. The scanning filling point recognition method based on weight distribution according to claim 1 is characterized in that: Obtaining a scanned image of a blank answer sheet, and establishing an answer sheet scan template according to the scanned image of the blank answer sheet, including: On the scanned image of the blank answer sheet, select the positioning mark point and the rectangular filling frame, mark the coordinate position of the selected rectangular filling frame, and generate an XML file answer sheet scanning template.

5. The scanning filling point recognition method based on weight distribution according to claim 4 is characterized in that: According to the answer sheet scan template, obtaining a rectangular filling frame in the scanned image of the examinee's answer sheet includes: Obtaining the answer sheet scan template and the examinee's answer sheet scan image; According to the positioning mark points, the scanned image of the examinee's answer sheet and the scanned image of the blank answer sheet are placed in the same coordinate system; According to the coordinate position of the rectangular filling frame of the answer sheet scanning template, the position information of the rectangular filling frame in the scanned image of the examinee's answer sheet is obtained.

6. The scanning filling point recognition method based on weight distribution according to claim 1 is characterized in that: Assigning values ​​to the pixel points within the rectangular filling frame according to a preset pixel value assignment strategy includes: Divide the grayscale value range of the rectangular fill frame into N equal parts. Divide the grayscale value range 0 to n of the rectangular filling frame into N grayscale value intervals of equal length; With a step size of (n+1) / N, select N grayscale values ​​in the grayscale value range of 0 to n, and make the N grayscale values ​​correspond to the N grayscale value intervals one by one; The pixel values ​​of the pixels in the rectangular filling frame are revalued so that the pixel values ​​of the pixels are equal to the grayscale values ​​corresponding to the grayscale value interval.

7. The scanning filling point recognition method based on weight distribution according to claim 1 is characterized in that: Calculating the recognition coefficient of the rectangular filling frame according to the preset segmentation module weight coefficient includes: The pixel values ​​of several segmentation modules are counted as A1, A2, A3...A n ; The weight coefficients of several segmentation modules are B1, B2, B3...B n ; Define the recognition coefficient of the rectangular filling frame as R; Then R = A1*B1+A2*B2+...+A n *B n .

8. A scanning filling point recognition system based on weight distribution, characterized in that: include: A scanning template making module is used to obtain a scanned image of a blank answer sheet and to create a scanning template for the answer sheet according to the scanned image of the blank answer sheet; An acquisition module obtains a scanned image of the examinee's answer sheet; An answer sheet information recognition module is used to obtain a rectangular filling frame in the scanned image of the examinee's answer sheet according to the answer sheet scanning template; A processing module, used for assigning values ​​to the pixel points in the rectangular filling frame according to a preset pixel value assignment strategy; A segmentation module, used for obtaining the area of ​​each rectangular filling frame, and performing equal-area segmentation on the rectangular filling frame to obtain a plurality of segmentation modules; A statistics module is used to count the pixel values ​​of the segmentation module, wherein: counting the pixel values ​​of the segmentation module includes: Selecting the pixel value corresponding to the largest number of pixel points in the segmentation module as the pixel value of the segmentation module; or Dividing the segmentation module into a plurality of pixel groups; Counting the number of pixels with the same pixel value in the pixel group; Select the pixel value corresponding to the largest number of pixels as the pixel value of the pixel group; Counting the number of pixel groups with the same pixel value in the segmentation module; Select the pixel value corresponding to the pixel group with the largest number as the pixel value of the segmentation module; The calculation module is used to calculate the recognition coefficient of the rectangular fill box by weighted summing up the preset segmentation module weight coefficient and the pixel value of the segmentation module, and the segmentation module weight coefficient at the center of the rectangular fill box is the largest. When the recognition coefficient is less than a preset threshold, the rectangular fill box is judged to be filled.

Images