Artificial intelligence-driven intelligent identification method for thermal paper production waste

Abstract

The invention discloses an artificial intelligence-driven thermal paper production waste intelligent identification method, and relates to the technical field of thermal paper production waste identification, and the method comprises the following steps: obtaining a first normal reference value and a second normal reference value based on a first amount of normal thermal paper drawing; obtaining an initial abnormal region based on the real-time thermo-sensitive paper graph, the first normal reference value and the second normal reference value; acquiring three-dimensional point cloud charts above and below the thermo-sensitive paper in the initial abnormal region part, and marking the three-dimensional point cloud charts as real-time point cloud charts; constructing a real-time detection angle based on the real-time point cloud atlas; acquiring a historical normal angle based on the second quantity of normal thermal sensitive paper; acquiring a normal angle reference value based on the historical normal angle; based on the real-time detection angle value and a normal angle reference value, judging whether the thermo-sensitive paper is a waste material or not; the method is used for solving the problem that in an existing thermo-sensitive paper production waste recognition technology, whether interference exists or not cannot be further analyzed to cause recognition error defects, and consequently the accuracy of thermo-sensitive paper production waste recognition is low.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of thermal paper production waste identification, in particular to an artificial intelligence driven thermal paper production waste intelligent identification method. BACKGROUND

[0002] As a kind of information recording carrier widely used in retail, logistics, medical, financial and other fields, the surface quality of thermal paper directly determines the clarity, durability of printing effect and user experience; In the production process of thermal paper, due to the complexity of coating process, the fluctuation of raw material characteristics and the influence of production environment, various defects such as uneven coating, spots, scratches, wrinkles and foreign matters will inevitably occur; These defective products will flow into the market, which will cause problems such as blurred printing, unrecognized bar code and paper jam; Therefore, it is necessary to identify the thermal paper production waste. With the development of machine vision technology, the automatic optical detection system based on traditional image processing, this kind of system collects images through camera, and uses preset threshold, edge detection and other algorithms to identify defects, but the image is greatly affected by light and shielding interference, which leads to the fact that normal thermal paper is easily identified as thermal paper containing defects, for example, dust is regarded as a defect of thermal paper, such as the patent application with publication number CN120411105A, a thermal coating defect detection method and system for thermal paper are disclosed, which judges whether a defect occurs by the range of gray value in the image, this method fails to further analyze whether the error defect is caused by interference, resulting in low accuracy of defect identification, the existing thermal paper production waste identification technology fails to further analyze whether the error defect is caused by interference, resulting in low accuracy of thermal paper production waste identification. SUMMARY

[0003] The present application aims to at least solve one of the technical problems in the prior art, based on the first number of normal thermal paper drawings to obtain the first normal reference value and the second normal reference value; Based on the real-time thermal paper image, the first normal reference value and the second normal reference value, the initial abnormal area is obtained; Obtain the three-dimensional point cloud image of the thermal paper above and below the initial abnormal area part, marked as real-time point cloud image; Based on the real-time point cloud image, the real-time detection angle is constructed; Based on the second number of normal thermal paper, the historical normal angle is obtained; Based on the historical normal angle, the normal angle reference value is obtained; Based on the real-time detection angle value and the normal angle reference value, it is judged whether the thermal paper is waste, so as to solve the problem that the existing thermal paper production waste identification technology fails to further analyze whether the error defect is caused by interference, resulting in low accuracy of thermal paper production waste identification.

[0004] To achieve the above purpose, the present application provides an artificial intelligence driven thermal paper production waste intelligent identification method, which comprises the following steps: acquire a thermal paper image, marked as a real-time thermal paper image; acquire a first normal reference value and a second normal reference value based on the first number of normal thermal paper drawings; acquire an initial abnormal area based on the real-time thermal paper image, the first normal reference value, and the second normal reference value; acquire a three-dimensional point cloud image of the initial abnormal area, marked as a real-time point cloud image; construct a real-time detection angle based on the real-time point cloud image; acquire a historical normal angle based on the second number of normal thermal papers; acquire a normal angle reference value based on the historical normal angle; determine whether the thermal paper is waste based on the real-time detection angle value and the normal angle reference value.

[0005] Further, acquiring the first normal reference value and the second normal reference value based on the first number of normal thermal paper drawings includes the following sub-steps: mark the first number of normal thermal paper drawings as a first historical thermal paper image; and perform grayscale processing on the first historical thermal paper image to obtain a grayscale value, marked as a first historical grayscale image; acquire the grayscale value of the pixel points in the first historical grayscale image, marked as a historical normal grayscale value; acquire the minimum value of the historical normal grayscale value, marked as the first normal reference value; acquire the maximum value of the historical normal grayscale value, marked as the second normal reference value.

[0006] Further, acquiring the initial abnormal area based on the real-time thermal paper image, the first normal reference value, and the second normal reference value includes the following sub-steps: perform grayscale processing on the real-time thermal paper image to obtain a grayscale image, marked as a real-time thermal paper grayscale image; set the grayscale value greater than or equal to the first normal reference value and less than or equal to the second normal reference value in the real-time thermal paper grayscale image to 1; set the grayscale value less than the first normal reference value or greater than the second normal reference value to 0 to obtain a real-time thermal paper binary image; acquire an area composed of the grayscale value 0 in the real-time thermal paper binary image, marked as the initial abnormal area.

[0007] Further, constructing the real-time detection angle based on the real-time point cloud image includes the following sub-steps: establish a three-dimensional coordinate system, marked as a point cloud reference coordinate system; place the real-time point cloud image in the point cloud reference coordinate system; and make the placed thermal paper parallel to the plane formed by the X-axis and the Y-axis in the point cloud reference coordinate system; Obtain the point cloud coordinates equal to the number of Y-axis data in the X-axis of the point cloud reference coordinate system, and mark it as the vertical same coordinate. Obtain the maximum distance between the vertical same coordinates, and mark it as the thickness reference distance.

[0008] Further, the real-time detection angle based on the real-time point cloud map further includes the following sub-steps: Obtain the third number of thickness reference distances; Sort the thickness reference distances from small to large from left to right, and arrange the thickness reference distances from left to right with a serial number, marked as the thickness serial number. The thickness serial number is a positive integer starting from 1. Establish a plane rectangular coordinate system with the data of the horizontal axis as the thickness serial number and the data of the vertical axis as the thickness reference distance, and mark it as the thickness change coordinate system. Obtain the coordinate point of the horizontal coordinate data as the thickness serial number and the vertical coordinate data as the thickness reference distance corresponding to the thickness serial number, and mark it as the thickness coordinate point. Draw all the thickness coordinate points in the thickness change coordinate system to obtain a scatter plot, marked as the thickness change scatter plot. Linearly fit the thickness change scatter plot to obtain a function, marked as the thickness change straight line. Obtain the angle between the thickness change straight line and the horizontal axis of the thickness change coordinate system, and mark it as the real-time detection angle.

[0009] Further, the historical normal angle based on the second number of normal thermal papers includes the following sub-steps: Based on the three-dimensional point cloud map of the second number of normal thermal papers, mark it as the historical normal point cloud map. Obtain the real-time detection angle based on the historical normal point cloud map as the real-time point cloud map, and mark it as the historical normal angle.

[0010] Further, the normal angle reference value based on the historical normal angle includes the following sub-steps: Obtain the range of the historical normal angle; evenly divide the range of the historical normal angle into a fourth number of ranges, and mark it as the angle division range. Obtain the frequency of the historical normal angle in each angle division range, and mark it as the division range frequency.

[0011] Further, the normal angle reference value based on the historical normal angle further includes the following sub-steps: Obtain the sum of the division range frequency, and mark it as the overall range frequency. Obtain the value by dividing the overall range frequency by the fourth number, and mark it as the range mean frequency. Set the abnormal proportion; obtain the product of the range mean frequency and the abnormal proportion, and mark it as the abnormal frequency threshold. The division range frequency less than the abnormal frequency threshold is marked as an abnormal range frequency.

[0012] Further, the obtaining of the normal angle reference value based on the historical normal angle comprises the following sub-steps: The division range frequencies are sorted from left to right according to the minimum values of the corresponding angle division ranges from small to large. It is judged whether the rightmost division range frequency is an abnormal range frequency, if yes, the rightmost abnormal range frequency is deleted, and the abnormal range frequencies are continuously deleted to the left until the rightmost division range frequency is not an abnormal range frequency, and then the process is stopped; when the rightmost division range frequency is not an abnormal range frequency, the maximum value of the angle division range corresponding to the rightmost division range frequency is obtained and marked as the normal angle reference value.

[0013] Further, the judging of whether the thermal paper is waste based on the real-time detection angle value and the normal angle reference value comprises the following sub-steps: If the real-time detection angle is greater than the normal angle reference value, the maximum value of the normal paper thickness is obtained and marked as a thickness threshold value; The projection position of the thickness reference distance greater than the thickness threshold value on the plane formed by the X-axis and the Y-axis of the point cloud reference coordinate system is obtained and marked as an abnormal thickness coordinate point; The circle with the first distance as the radius is marked as a reference circle; The reference circle is established with the arbitrary abnormal thickness coordinate point as the center, and the reference circle is marked as a starting circle; it is judged whether the starting circle contains the abnormal thickness coordinate point, if yes, a new reference circle is established with the new abnormal thickness coordinate point as the center, and the new reference circle is continuously judged and drawn until the new reference circle does not contain the abnormal thickness coordinate point, and then the number of the abnormal thickness coordinate points searched after the process is stopped is obtained and marked as a search continuous number; An abnormal number threshold value is set; it is judged whether the search continuous number is greater than the abnormal number threshold value, if yes, the thermal paper is waste, and if not, the thermal paper needs to be further judged whether it is waste; If the real-time detection angle is less than or equal to the normal angle reference value, it is indicated that the initial abnormal area is not a defect area, and the thermal paper needs to be further judged whether it is waste.

[0014] The application has the beneficial effects that: the first normal reference value and the second normal reference value are acquired based on the first number of normal heat-sensitive paper drawings, the initial abnormal area is acquired based on the real-time heat-sensitive paper drawing, the first normal reference value and the second normal reference value, the three-dimensional point cloud graph of the heat-sensitive paper above and below the initial abnormal area is acquired, and is marked as a real-time point cloud graph; the real-time detection angle is constructed based on the real-time point cloud graph; the historical normal angle is acquired based on the second number of normal heat-sensitive paper; the normal angle reference value is acquired based on the historical normal angle; and it is judged whether the heat-sensitive paper is waste material based on the real-time detection angle value and the normal angle reference value, and the advantage is that whether the recognized error defect is caused by interference can be further analyzed, and the accuracy of heat-sensitive paper production waste material recognition is improved. The application has the beneficial effects that: the first normal reference value and the second normal reference value are acquired based on the first number of normal heat-sensitive paper drawings, the initial abnormal area is acquired based on the real-time heat-sensitive paper drawing, the first normal reference value and the second normal reference value, the three-dimensional point cloud graph of the heat-sensitive paper above and below the initial abnormal area is acquired, and is marked as a real-time point cloud graph; the real-time detection angle is constructed based on the real-time point cloud graph; the historical normal angle is acquired based on the second number of normal heat-sensitive paper; the normal angle reference value is acquired based on the historical normal angle; and it is judged whether the heat-sensitive paper is waste material based on the real-time detection angle value and the normal angle reference value, and the advantage is that whether the recognized error defect is caused by interference can be further analyzed, and the accuracy of heat-sensitive paper production waste material recognition is improved. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 The step flow chart of the method of the application is shown in the figure. Figure 2 The schematic diagram of the thickness change scatter plot of the application is shown in the figure. Figure 3 The schematic diagram of the thickness change straight line of the application is shown in the figure. Figure 4 The schematic diagram of the starting circle and the new reference circle of the application is shown in the figure. DETAILED DESCRIPTION

[0016] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the application.

[0017] Embodiment 1, please refer to Figure 1 The application provides an artificial intelligence driven heat-sensitive paper production waste material intelligent recognition method, which comprises the following steps: Step S1, acquiring a heat-sensitive paper image, which is marked as a real-time heat-sensitive paper drawing; Step S2, acquiring a first normal reference value and a second normal reference value based on a first number of normal heat-sensitive paper drawings; step S2 comprises the following substeps: Step S201, a first number of normal thermal paper drawing icons are marked as first historical thermal paper drawings; the first historical thermal paper drawings are subjected to gray scale processing to obtain gray scale values, which are marked as first historical gray scale drawings; a first number of first historical thermal paper drawings are obtained to facilitate obtaining a range of gray scale values of normal thermal paper; therefore, the first number should not be too small, for example, the first number is 100; Step S202, the gray scale values of the pixel points in the first historical gray scale drawings are obtained, which are marked as historical normal gray scale values; Step S203, the minimum value of the historical normal gray scale values is obtained, which is marked as a first normal reference value; Step S204, the maximum value of the historical normal gray scale values is obtained, which is marked as a second normal reference value; the first normal reference value and the second normal reference value obtain the distribution of the gray scale values of normal thermal paper, and the gray scale values of defect or abnormal areas will be different, so that whether an abnormal area exists can be distinguished through the first normal reference value and the second normal reference value.

[0018] Step S3, an initial abnormal area is obtained based on a real-time thermal paper drawing, the first normal reference value and the second normal reference value; step S3 includes the following sub-steps: Step S301, the real-time thermal paper drawing is subjected to gray scale processing to obtain a gray scale drawing, which is marked as a real-time thermal paper gray scale drawing; Step S302, the gray scale values greater than or equal to the first normal reference value and less than or equal to the second normal reference value in the real-time thermal paper gray scale drawing are set to 1; the gray scale values less than the first normal reference value or greater than the second normal reference value are set to 0, to obtain a real-time thermal paper binary drawing; the range of the gray scale values greater than or equal to the first normal reference value and less than or equal to the second normal reference value is the range of the gray scale values of normal thermal paper; otherwise, it is an abnormal area, which may be a defect or an interference caused area; Step S303, an area composed of 0 gray scale values in the real-time thermal paper binary drawing is obtained, which is marked as an initial abnormal area.

[0019] Step S4, a three-dimensional point cloud drawing of the thermal paper above and below the initial abnormal area is obtained, which is marked as a real-time point cloud drawing; the point cloud drawing can be obtained by a laser radar or the like.

[0020] Step S5, a real-time detection angle is constructed based on the real-time point cloud drawing; step S5 includes the following sub-steps: Step S501, a three-dimensional coordinate system is established, which is marked as a point cloud reference coordinate system; the real-time point cloud drawing is in the point cloud reference coordinate system; the placed thermal paper is parallel to the plane formed by the X axis and the Y axis in the point cloud reference coordinate system; Step S502, obtain the point cloud coordinates equal to the number of X-axis data and Y-axis in the point cloud reference coordinate system, marked as vertical same coordinates; obtain the maximum distance between the vertical same coordinates, marked as thickness reference distance; obtain the thickness reference distance equivalent to the thickness of the paper through the point cloud map; Step S503, obtain a third number of thickness reference distances; the third number is to facilitate the initial abnormal area to be a defect or uneven; Step S504, sort the thickness reference distances from small to large from left to right, and mark the thickness reference distances from left to right in sequence, marked as thickness sequence number; the thickness sequence number is a positive integer starting from 1; the thickness sequence number is to facilitate data analysis; Step S505, establish a plane rectangular coordinate system with the data of the horizontal axis as the thickness sequence number and the data of the vertical axis as the thickness reference distance, marked as the thickness change coordinate system; Step S506, obtain the coordinate points of the horizontal coordinate data as the thickness sequence number and the vertical coordinate data as the thickness reference distance corresponding to the thickness sequence number, marked as the thickness coordinate points; Step S507, draw all the thickness coordinate points in the thickness change coordinate system to obtain a scatter plot, marked as the thickness change scatter plot; Step S508, linearly fit the thickness change scatter plot to obtain a function, marked as the thickness change straight line; the thickness change straight line has the beneficial effect of reducing the interference judgment of abnormal too large or abnormal too small thickness reference distances, and further accurately judging whether it is abnormal; Step S509, obtain the angle between the thickness change straight line and the horizontal axis of the thickness change coordinate system, marked as the real-time detection angle; the real-time detection angle is equivalent to the unevenness of the thickness; based on the unevenness, the abnormality of the thermal paper whether it has defects, defects and other abnormalities, the abnormal part of the thermal paper is usually uneven, so whether it is a defect can be judged by the real-time detection angle; In practical applications, please refer to Figure 2 and Figure 3 As shown, the thickness change straight line is drawn; the angle between the thickness change straight line and the horizontal axis of the thickness change coordinate system is 16°, so the real-time detection angle is 16°.

[0021] Step S6, obtain the historical normal angle based on the second number of normal thermal papers; Step S6 includes the following sub-steps: Step S601, based on the three-dimensional point cloud map of the second number of normal thermal papers, marked as the historical normal point cloud map; obtain the real-time detection angle based on the historical normal point cloud map as the real-time point cloud map, marked as the historical normal angle; the second number is to obtain the range of the real-time detection angle of the normal thermal paper; therefore, the second number should not be too small, for example, the second number is 100; Step S7, obtaining a normal angle reference value based on the historical normal angles; step S7 includes the following sub-steps: Step S701, obtaining a range of the historical normal angles; the range of the historical normal angles is evenly divided into a fourth number of ranges, marked as angle division ranges; the fourth number is appropriately set, for example, the fourth number is set to 20; Step S702, obtaining a frequency of the historical normal angles in each angle division range, marked as a division range frequency.

[0022] Step S703, obtaining a sum of the division range frequencies, marked as an overall range frequency; Step S704, obtaining a value by dividing the overall range frequency by the fourth number, marked as a range mean frequency; Step S705, setting an abnormal proportion; obtaining a value by multiplying the range mean frequency and the abnormal proportion, marked as an abnormal frequency threshold; the abnormal frequency threshold is used to obtain a smaller division range frequency, and thus the abnormal proportion is set to be small, for example, 0.3; Step S706, marking the division range frequencies smaller than the abnormal frequency threshold as abnormal range frequencies; Step S707, sorting the division range frequencies from left to right according to the minimum values of the corresponding angle division ranges from small to large; Step S708, determining whether the rightmost division range frequency is an abnormal range frequency, if yes, deleting the rightmost abnormal range frequency, and continuously deleting the abnormal range frequencies to the left until the rightmost division range frequency is not an abnormal range frequency; when the rightmost division range frequency is not an abnormal range frequency, obtaining a maximum value of the angle division range corresponding to the rightmost division range frequency, marked as a normal angle reference value; the normal angle reference value indicates that the normal real-time detection angle cannot exceed the normal angle reference value; In actual application, the range of the historical normal angles is 0.2° to 20.2°; the range of the historical normal angles is evenly divided into 20 angle division ranges, which are 0.2° to 1.2°, 1.2° to 2.2°,..., 19.2° to 20.2°; the obtained overall range frequency is 100; the calculated range mean frequency is 100÷20=5; the calculated abnormal frequency threshold is 5×0.25=1.25; the division range frequencies smaller than 1.25 are marked as abnormal range frequencies; the division range frequencies are sorted from left to right according to the minimum values of the corresponding angle division ranges from small to large, which are 1, 2,..., 3, 1; after the abnormal range frequencies are deleted, the sorting is 1, 2,..., 3; the maximum value of the angle division range corresponding to the rightmost division range frequency 3 is 19.2°, and thus the normal angle reference value is 19.2°.

[0023] Step S8, judging whether the thermal paper is waste based on the real-time detection angle value and the normal angle reference value; Step S8 includes the following sub-steps: Step S801, if the real-time detection angle is greater than the normal angle reference value, obtaining the maximum value of the normal paper thickness, marked as the thickness threshold value; the thickness threshold value is the maximum value of the normal thermal paper thickness, for example, the thickness threshold value is 92μm; Step S802, obtaining the projection position of the thickness reference distance greater than the thickness threshold value on the plane composed of the X-axis and the Y-axis of the point cloud reference coordinate system, marked as the abnormal thickness coordinate point; Step S803, marking the circle with the first distance as the reference circle; the setting of the reference circle is to obtain continuous abnormal thickness coordinate points, so the first distance should not be too large; for example, the first distance is 200μm; Step S804, establishing a reference circle with any abnormal thickness coordinate point as the center, marked as the starting circle; obtaining whether the starting circle contains abnormal thickness coordinate points, if it contains, establishing a new reference circle with a new abnormal thickness coordinate point as the center, continuing to judge and draw a new reference circle, until the new reference circle does not contain abnormal thickness coordinate points, stopping, obtaining the number of abnormal thickness coordinate points searched after stopping, marked as the search continuous number; the search continuous number is used to judge whether it is a defect misjudgment caused by dust particles, because the dust particles are smaller, if it is a dust particle, the search continuous number is smaller, so the search continuous number can be used to judge whether it is a dust particle; In practical application, please refer to Figure 4 The starting circle is shown in the figure; for example, the search continuous number is 12; Step S805, setting the abnormal number threshold value; judging whether the search continuous number is greater than the abnormal number threshold value, if yes, indicating that the thermal paper is waste, if not, further judging whether the thermal paper is waste; the abnormal number threshold value is the minimum search continuous number of defects, to exclude the interference of dust particles, the abnormal number threshold value is 10; less than the abnormal number threshold value may be a misjudgment of dust particles, so the thermal paper may not be waste, which needs to be further judged; Step S806, if the real-time detection angle is less than or equal to the normal angle reference value, indicating that the initial abnormal area is not a defect area, which needs to be further judged whether the thermal paper is waste; the real-time detection angle less than or equal to the normal angle reference value indicates that the initial abnormal area is flat, which may be a misjudgment of interference; In practical application, the real-time detection angle is 16°, which is less than the normal angle reference value of 19.2°, indicating that the initial abnormal area is flat, while the thermal paper waste will appear defects and uneven thickness, here the real-time detection angle is greater than the normal angle reference value, so it may not be thermal paper waste, which needs to be further judged whether it is waste.

[0024] In some embodiments, the electronic device can include a processor, a communication interface, a memory, and a communication bus. The processor, the communication interface, and the memory can communicate with each other through the communication bus. The memory can store computer-readable instructions. The processor can invoke the instructions in the memory. When the computer-readable instructions are executed by the processor, the steps in the artificial intelligence driven thermal paper production waste intelligent identification method can be executed to achieve the following functions: obtaining a thermal paper image, marking it as a real-time thermal paper image; obtaining a first normal reference value and a second normal reference value based on a first number of normal thermal paper drawings; obtaining an initial abnormal area based on the real-time thermal paper image, the first normal reference value, and the second normal reference value; obtaining a three-dimensional point cloud image of the thermal paper above and below the initial abnormal area, marking it as a real-time point cloud image; constructing a real-time detection angle based on the real-time point cloud image; obtaining a historical normal angle based on a second number of normal thermal paper; obtaining a normal angle reference value based on the historical normal angle; and determining whether the thermal paper is waste based on the real-time detection angle value and the normal angle reference value.

[0025] In addition, the logical instructions in the memory described above can be implemented in the form of a software function unit and sold or used as an independent product. When used, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various program code storage media.

[0026] In embodiment 3, the present application also provides a computer program product, which comprises a computer program stored on a computer readable storage medium, and the computer program comprises program instructions, when the program instructions are executed by a computer, the computer can execute the artificial intelligence driven thermal paper production waste intelligent identification method provided by each method, and the method comprises the following steps: acquiring a thermal paper image, and marking the thermal paper image as a real-time thermal paper image; acquiring a first normal reference value and a second normal reference value based on a first number of normal thermal paper drawing maps; acquiring an initial abnormal area based on the real-time thermal paper image, the first normal reference value and the second normal reference value; acquiring a three-dimensional point cloud map of the thermal paper above and below the initial abnormal area, and marking the three-dimensional point cloud map as a real-time point cloud map; constructing a real-time detection angle based on the real-time point cloud map; acquiring a historical normal angle based on a second number of normal thermal paper; acquiring a normal angle reference value based on the historical normal angle; and determining whether the thermal paper is waste based on the real-time detection angle value and the normal angle reference value.

[0027] In embodiment 4, the present application also provides a computer readable storage medium, and the present application provides a storage medium, which stores a computer program, and the computer program is executed by a processor to run the steps in the artificial intelligence driven thermal paper production waste intelligent identification method to realize the following functions: acquiring a thermal paper image, and marking the thermal paper image as a real-time thermal paper image; acquiring a first normal reference value and a second normal reference value based on a first number of normal thermal paper drawing maps; acquiring an initial abnormal area based on the real-time thermal paper image, the first normal reference value and the second normal reference value; acquiring a three-dimensional point cloud map of the thermal paper above and below the initial abnormal area, and marking the three-dimensional point cloud map as a real-time point cloud map; constructing a real-time detection angle based on the real-time point cloud map; acquiring a historical normal angle based on a second number of normal thermal paper; acquiring a normal angle reference value based on the historical normal angle; and determining whether the thermal paper is waste based on the real-time detection angle value and the normal angle reference value.

[0028] Based on the description of the above embodiments, the embodiments of the present application can be provided as a method, a system or a computer program product. Based on such understanding, the above technical solutions can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the method described in each embodiment or some parts of the embodiment.

[0029] In the embodiments of the present application, it should be understood that the disclosed system or method can be implemented in other ways. The embodiments described above are only illustrative, for example, the division of modules or units is only a logical function division, and other division manners can be used in actual implementation, for example, a plurality of modules or units can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed each other can be through some communication interface, the indirect coupling or communication connection between the system, the module and the unit can be electrical, mechanical or other forms.

[0030] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An artificial intelligence driven thermal paper production waste intelligent identification method, characterized in that, The method comprises the following steps: acquiring a thermal paper image, marked as a real-time thermal paper image; acquiring a first normal reference value and a second normal reference value based on a first number of normal thermal paper drawings; acquiring an initial abnormal area based on the real-time thermal paper image, the first normal reference value, and the second normal reference value; acquiring a three-dimensional point cloud image of the thermal paper above and below the initial abnormal area, marked as a real-time point cloud image; constructing a real-time detection angle based on the real-time point cloud image; acquiring a historical normal angle based on a second number of normal thermal paper; acquiring a normal angle reference value based on the historical normal angle; judging whether the thermal paper is waste based on the real-time detection angle value and the normal angle reference value. 2.The artificial intelligence driven thermal paper production waste intelligent identification method of claim 1, wherein, The method of acquiring a first normal reference value and a second normal reference value based on a first number of normal thermal paper drawings comprises the following sub-steps: marking the first number of normal thermal paper drawings as a first historical thermal paper image; performing grayscale processing on the first historical thermal paper image to obtain a grayscale value, marked as a first historical grayscale image; acquiring a grayscale value of a pixel point in the first historical grayscale image, marked as a historical normal grayscale value; acquiring a minimum value of the historical normal grayscale value, marked as the first normal reference value; acquiring a maximum value of the historical normal grayscale value, marked as the second normal reference value. 3.The artificial intelligence driven thermal paper production waste intelligent identification method of claim 2, wherein, The method of acquiring an initial abnormal area based on the real-time thermal paper image, the first normal reference value, and the second normal reference value comprises the following sub-steps: performing grayscale processing on the real-time thermal paper image to obtain a grayscale image, marked as a real-time thermal paper grayscale image; setting a grayscale value greater than or equal to the first normal reference value and less than or equal to the second normal reference value in the real-time thermal paper grayscale image to 1; setting a grayscale value less than the first normal reference value or greater than the second normal reference value to 0 to obtain a real-time thermal paper binary image; acquiring an area composed of a grayscale value of 0 in the real-time thermal paper binary image, marked as the initial abnormal area. 4.The artificial intelligence driven thermal paper production waste intelligent identification method of claim 3, wherein, The method of constructing a real-time detection angle based on the real-time point cloud image comprises the following sub-steps: establishing a three-dimensional coordinate system, marked as a point cloud reference coordinate system; placing the real-time point cloud image in the point cloud reference coordinate system; and making the placed thermal paper parallel to a plane formed by an X-axis and a Y-axis in the point cloud reference coordinate system; acquiring a point cloud coordinate with equal X-axis data and Y-axis quantity in the point cloud reference coordinate system, marked as a vertical same coordinate; and acquiring a maximum distance between the vertical same coordinates, marked as a thickness reference distance. 5.The artificial intelligence driven thermal paper production waste intelligent identification method of claim 4, wherein, The method of constructing a real-time detection angle based on the real-time point cloud image further comprises the following sub-steps: acquiring a third number of thickness reference distances; sorting the thickness reference distances from left to right according to the thickness reference distances from small to large, and marking and arranging serial numbers of the thickness reference distances from left to right, marked as a thickness serial number; the thickness serial number is a positive integer starting from 1; establishing a plane rectangular coordinate system with data of a horizontal axis as the thickness serial number and data of a vertical axis as the thickness reference distance, marked as a thickness change coordinate system; acquiring a coordinate point with data of a horizontal coordinate as the thickness serial number and data of a vertical coordinate as the thickness reference distance corresponding to the thickness serial number, marked as a thickness coordinate point; drawing all the thickness coordinate points in the thickness change coordinate system to obtain a scatter plot, marked as a thickness change scatter plot; performing linear fitting on the thickness change scatter plot to obtain a function, marked as a thickness change straight line; An angle between the straight line of thickness variation and the horizontal axis of the thickness variation coordinate system is obtained, and is marked as a real-time detection angle. 6.The artificial intelligence driven thermal paper production waste intelligent identification method of claim 5, wherein, The historical normal angle based on the second number of normal thermal paper includes the following sub-steps: A three-dimensional point cloud diagram based on the second number of normal thermal paper is obtained, and is marked as a historical normal point cloud diagram; the historical normal point cloud diagram is regarded as a real-time point cloud diagram to obtain the real-time detection angle, and is marked as a historical normal angle. 7.The artificial intelligence driven thermal paper production waste intelligent identification method of claim 6, wherein, The normal angle reference value based on the historical normal angle includes the following sub-steps: A range of the historical normal angle is obtained; the range of the historical normal angle is evenly divided into a fourth number of ranges, and is marked as an angle division range; A frequency of the historical normal angle in each angle division range is obtained, and is marked as a division range frequency. 8.The artificial intelligence driven thermal paper production waste intelligent identification method of claim 7, wherein, The normal angle reference value based on the historical normal angle further includes the following sub-steps: A sum of the division range frequencies is obtained, and is marked as an overall range frequency; A value obtained by dividing the overall range frequency by the fourth number is marked as a range mean frequency; An abnormal proportion is set; a product of the range mean frequency and the abnormal proportion is obtained, and is marked as an abnormal frequency threshold; The division range frequency less than the abnormal frequency threshold is marked as an abnormal range frequency. 9.The artificial intelligence driven thermal paper production waste intelligent identification method of claim 8, wherein, The normal angle reference value based on the historical normal angle further includes the following sub-steps: The division range frequencies are sorted from left to right according to the minimum values of the corresponding angle division ranges from small to large; It is judged whether the rightmost division range frequency is an abnormal range frequency, if yes, the rightmost abnormal range frequency is deleted, and the abnormal range frequency is continuously deleted to the left until the rightmost division range frequency is not an abnormal range frequency; when the rightmost division range frequency is not an abnormal range frequency, a maximum value of the angle division range corresponding to the rightmost division range frequency is obtained, and is marked as a normal angle reference value. 10.The artificial intelligence driven thermal paper production waste intelligent identification method of claim 9, wherein, The judgment of whether the thermal paper is waste based on the real-time detection angle value and the normal angle reference value includes the following sub-steps: If the real-time detection angle is greater than the normal angle reference value, a maximum value of the normal paper thickness is obtained, and is marked as a thickness threshold; A projection position of the thickness reference distance greater than the thickness threshold on a plane formed by the X-axis and the Y-axis of the point cloud reference coordinate system is obtained, and is marked as an abnormal thickness coordinate point; A circle with a first distance as a radius is marked as a reference circle; A reference circle is established with an arbitrary abnormal thickness coordinate point as a center, and is marked as a starting circle; it is judged whether the starting circle contains an abnormal thickness coordinate point, if yes, a new reference circle is established with a new abnormal thickness coordinate point as a center, and a new reference circle is continuously judged and drawn until the new reference circle does not contain an abnormal thickness coordinate point; a number of abnormal thickness coordinate points searched after stopping is obtained, and is marked as a search continuous number; An abnormal number threshold is set; it is judged whether the search continuous number is greater than the abnormal number threshold, if yes, the thermal paper is waste, and if not, the thermal paper needs to be further judged whether it is waste. If the real-time detection angle is less than or equal to the normal angle reference value, it is indicated that the initial abnormal area is not a defect area, and the thermal paper needs to be further judged whether it is waste.

Images