Damp-proof corrugated paper generation method and system based on multi-modal parameter fusion image transformation

Abstract

The invention discloses a moisture-proof corrugated paper generation method and system based on multi-modal parameter fusion image transformation. A damp-proof corrugated paper generation model is built by utilizing a machine learning algorithm, corrugated paper production equipment parameters, corrugated paper images and order destination information are processed, and different damp-proof material grade filling methods needed for conveying corrugated paper types needing to be produced to destinations are obtained. The filling method for different levels of moisture-proof materials is used for filling moisture-proof materials of different degrees based on a paper core cavity, and the individual requirements of customers are met. According to the method, through the built damp-proof corrugated paper generation model, the labor cost for adjusting corrugated paper production equipment is reduced, green and economical production of the damp-proof corrugated paper is achieved, intelligent and precise damp-proof material filling control is achieved for different orders, and the production quality is effectively improved.

Description

Technical Field

[0001] The invention belongs to the technical field of intelligent production of moisture-proof corrugated paper, and in particular relates to a method and system for generating moisture-proof corrugated paper by multi-modal parameter fusion image transformation. Background Art

[0002] In the corrugated paper processing industry, in order to achieve high-quality development and save raw materials, personalized moisture-proof corrugated paper production based on customer needs is the key to ensuring efficiency and quality. The traditional method of adjusting equipment production parameters mainly relies on the operator's experience and judgment, and does not take into account the situation of the order destination to fill the moisture-proof material to improve the moisture-proof performance of the corrugated paper. The manual method is not only time-consuming and labor-intensive, but it is often difficult to accurately design equipment parameters for corresponding moisture-proof material filling, resulting in waste of equipment filling materials or insufficient moisture-proof performance. In addition, with the increase in demand for corrugated paper and the complexity of production processes, the magnitude and complexity of equipment operation data have also increased significantly, and traditional methods are even more unable to cope with it.

[0003] Although the development of information technology and automation technology in recent years has provided new means for equipment monitoring and fault diagnosis, such as installing various sensors to collect equipment operation data in real time and using computer systems to monitor and analyze data, these technologies often focus on collecting and monitoring data, but there are still deficiencies in how to extract valuable information from a large amount of complex data and accurately and quickly obtain the relevant filling setting parameters of the equipment. In particular, the lack of effective data processing and analysis models makes it difficult to cope with the high-dimensional and time series characteristics of equipment operation data, resulting in waste of moisture-proof materials and insufficient timeliness in responding to customer needs.

[0004] In addition, existing technologies may not be accurate or effective in filling moisture-proof materials, thereby affecting the efficiency of equipment use and the production quality of products.

[0005] Through the above analysis, the existing technology has the following problems to be solved:

[0006] How to combine order information and image information to finely control the use of moisture-proof materials with sufficient experience, that is, how to process multimodal data, and at the same time more accurately obtain the corresponding equipment parameter setting methods in classifying the use of moisture-proof materials, and how to obtain valid data in the processing of corrugated paper images and make the model processing more accurate.

[0007] Therefore, it is necessary to design a moisture-proof corrugated paper generation method and system based on multimodal parameter fusion image transformation to address the existing problems. Summary of the invention

[0008] In order to solve the above technical problems, the present invention proposes a method and system for generating moisture-proof corrugated paper by multi-modal parameter fusion image transformation.

[0009] In a first aspect of the present invention, a method for generating moisture-proof corrugated paper by multimodal parameter fusion image transformation is provided, characterized in that:

[0010] Acquire a first corrugated paper image, and acquire a first moisture-proof material grade filling method set by a technician at this time, wherein the first corrugated paper image is not filled with moisture-proof material;

[0011] Using crawler technology to crawl the first order destination information and obtain the first corrugated paper production equipment parameters, and performing multimodal parameter fusion on the first corrugated paper image, the first order destination information and the first corrugated paper production equipment parameters to obtain a first filling feature;

[0012] Using the first filling feature and the first moisture-proof material grade filling method, a moisture-proof corrugated paper generation model is trained;

[0013] Acquire a second corrugated paper image, where the second corrugated paper image is not filled with moisture-proof material, crawl second order destination information and obtain second corrugated paper production equipment parameters, and perform multimodal parameter fusion to obtain a second filling feature;

[0014] The moisture-proof corrugated paper generation model processes the second filling feature to obtain a second moisture-proof material grade filling method that does not require manual setting by a technician, and uses the second moisture-proof material grade filling method to fill the corrugated paper with moisture-proof material.

[0015] Furthermore, the moisture-proof corrugated paper generation model is constructed using machine learning or deep learning methods.

[0016] Furthermore, the first corrugated paper image or the second corrugated paper image is an image of a paperboard interface having a waterproof membrane, a face paper, a fireproof core layer and a support column, taken by a camera device installed on the top of a factory building.

[0017] Furthermore, before using the first corrugated paper image or the second corrugated paper image for multimodal parameter fusion, the first corrugated paper image or the second corrugated paper image is subjected to focusing processing of the target image and the background image to obtain the first corrugated paper image input vector or the second corrugated paper image input vector, and the image focusing model is used for image segmentation, and the image obtained after the image focusing is converted into a vector value, wherein the image focusing model realizes image focusing by extremizing the following wave function M:

[0018]

[0019] Where, I(X) is the first corrugated paper image or the second corrugated paper image to be focused, C is the evolution curve, μ, v, λ 1 ,λ 2 are all non-negative coefficients, c 1 , c 2 are the average grayscale values ​​of image I in the target and background images, in(C) is the target area, out(C) is the background area, lE is the length constraint, a is the area constraint, and c is the average grayscale value of image I in the target and background images. 1 , c 2 The specific expression is as follows:

[0020]

[0021]

[0022] in(C) and out(C) are the regional images obtained by focusing. in(C) and out(C) are binarized again and converted into vector representations, where in(C) is a corrugated paper interface image that only contains the first corrugated paper image or the second corrugated paper image.

[0023] Further, the first order destination information or the second order destination information includes the temperature, humidity and annual precipitation of the order destination;

[0024] The first corrugated paper production equipment parameter or the second corrugated paper production equipment parameter includes the operating speed and pressure of the corrugated paper production equipment.

[0025] Further, the multimodal parameter fusion is vector splicing or dimension splicing, and the vector splicing is a one-dimensional vector directly linked to the first corrugated paper image input vector, the first order destination information, and the first corrugated paper production equipment parameter, or a one-dimensional vector directly linked to the second corrugated paper image input vector, the second order destination information, and the second corrugated paper production equipment parameter;

[0026] The dimensional splicing is a three-dimensional vector obtained by dimensionally linking the first corrugated paper image input vector, the first order destination information, and the first corrugated paper production equipment parameter, or a three-dimensional vector obtained by directly linking the second corrugated paper image input vector, the second order destination information, and the second corrugated paper production equipment parameter.

[0027] Furthermore, the machine learning method adopts an improved classifier based on the Fisher criterion, and the calculation formula is as follows:

[0028]

[0029] A is the first filling feature or the second filling feature, g(S) is the output first moisture-proof material level filling method or the second moisture-proof material level filling method, W T is the normal vector perpendicular to the hyperplane, k is the south or north of the country to which the order destination belongs, k is 1 in the south and -1 in the north, S i is the contact area between the filling material and the corrugated paper, S o The area where the filling material does not contact the corrugated paper.

[0030] A moisture-proof corrugated paper generation system based on multi-modal parameter fusion image transformation is also provided, characterized in that the system comprises:

[0031] A corrugated paper filling image acquisition module is used to acquire a first corrugated paper image, where the first corrugated paper image is not filled with a moisture-proof material, and to acquire a first moisture-proof material grade filling method set by a technician at this time;

[0032] The corrugated paper filling processing module crawls the first order destination information using crawler technology and obtains the first corrugated paper production equipment parameters. The corrugated paper filling processing module is also connected to the corrugated paper filling image acquisition module to obtain the first corrugated paper image, and performs multimodal parameter fusion on the first corrugated paper image, the first order destination information and the first corrugated paper production equipment parameters to obtain a first filling feature; the corrugated paper filling processing module also receives the second corrugated paper image obtained by the corrugated paper filling image acquisition module, the second corrugated paper image is not filled with moisture-proof material, crawls the second order destination information and obtains the second corrugated paper production equipment parameters, and performs multimodal parameter fusion on them to obtain a second filling feature;

[0033] The moisture-proof corrugated paper generation model module receives the first filling feature and the first moisture-proof material grade filling method, and trains to obtain a moisture-proof corrugated paper generation model, wherein the moisture-proof corrugated paper generation model is a machine learning model; the moisture-proof corrugated paper generation model module also receives the second filling feature, and processes to obtain a second moisture-proof material grade filling method that does not require manual setting by a technician;

[0034] The corrugated paper filling control module receives the second moisture-proof material grade filling method and fills the corrugated paper with the moisture-proof material.

[0035] Furthermore, the moisture-proof corrugated paper generation model is constructed using a machine learning or deep learning method;

[0036] The machine learning method adopts an improved classifier based on the Fisher criterion, and the calculation formula is as follows:

[0037]

[0038] A is the first filling feature or the second filling feature, g(S) is the output first moisture-proof material level filling method or the second moisture-proof material level filling method, W T is the normal vector perpendicular to the hyperplane, k is the south or north of the country to which the order destination belongs, k is 1 in the south and -1 in the north, S i is the contact area between the filling material and the corrugated paper, S o The area where the filling material does not contact the corrugated paper.

[0039] Furthermore, before using the first corrugated paper image or the second corrugated paper image for multimodal parameter fusion, the first corrugated paper image or the second corrugated paper image is subjected to focusing processing of the target image and the background image to obtain the first corrugated paper image input vector or the second corrugated paper image input vector, and the image focusing model is used for image segmentation, and the image obtained after the image focusing is converted into a vector value, wherein the image focusing model realizes image focusing by extremizing the following wave function M:

[0040]

[0041] Where, I(X) is the first corrugated paper image or the second corrugated paper image to be focused, C is the evolution curve, μ, v, λ 1 ,λ 2 are all non-negative coefficients, c 1 , c 2 are the average grayscale values ​​of image I in the target and background images, in(C) is the target area, out(C) is the background area, lE is the length constraint, a is the area constraint, and c is the average grayscale value of image I in the target and background images. 1 , c 2 The specific expression is as follows:

[0042]

[0043] in(C) and out(C) are the regional images obtained by focusing. in(C) and out(C) are binarized again and converted into vector representations, where in(C) is a corrugated paper interface image that only contains the first corrugated paper image or the second corrugated paper image.

[0044] The present invention discloses a method and system for generating moisture-proof corrugated paper by multi-modal parameter fusion image transformation. By combining order information, processed transformed image information and corrugated paper equipment parameter data, the method of using moisture-proof materials with sufficient experience is finely controlled, and data processing of corrugated paper filling materials with multiple modal data is performed. The method of using moisture-proof filling materials is obtained more accurately in the classification method, and the order information is fully utilized to perform corresponding production control of moisture-proof corrugated paper, which can well save the cost of using moisture-proof materials, improve customer satisfaction and reduce the cost of enterprises.

[0045] More embodiments and improved effects of the present invention will be further introduced in conjunction with the drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a flow chart of a moisture-proof corrugated paper generation method of multi-modal parameter fusion image transformation of the present invention;

[0047] Figure 2 It is a schematic diagram of a moisture-proof corrugated paper generation system of a multi-modal parameter fusion image transformation of the present invention;

[0048] Figure 3 is a schematic diagram of a target image portion of a moisture-proof material of moisture-proof corrugated paper in an embodiment of the present invention;

[0049] Numbers in the figure: 1. Waterproof membrane; 2. Face paper; 3. Fireproof core layer; 4. Honeycomb paper core; 5. Cavity; 6. Support column.

[0050] Figure 4 Schematic diagram of the opening surface of the corrugated paper without adding moisture-proof material in the embodiment of the present invention;

[0051] Figure 5 It is a schematic diagram of the structure of an electronic device for implementing the method of the present invention according to an embodiment of the present invention. DETAILED DESCRIPTION

[0052] The invention is further described below in conjunction with the accompanying drawings and specific implementation methods.

[0053] In order to solve the above technical problems, the present invention proposes a method and system for generating moisture-proof corrugated paper by multi-modal parameter fusion image transformation.

[0054] In a first aspect of the present invention, a method for generating moisture-proof corrugated paper by multimodal parameter fusion image transformation is provided, characterized in that:

[0055] Acquire a first corrugated paper image, and acquire a first moisture-proof material grade filling method set by a technician at this time, wherein the first corrugated paper image is not filled with moisture-proof material;

[0056] The first moisture-proof material set in this embodiment is water-absorbent pearl cotton, and the thickness of the washable pearl cotton is 1mm-10mm.

[0057] Using crawler technology to crawl the first order destination information and obtain the first corrugated paper production equipment parameters, and performing multimodal parameter fusion on the first corrugated paper image, the first order destination information and the first corrugated paper production equipment parameters to obtain a first filling feature;

[0058] Using the first filling feature and the first moisture-proof material grade filling method, a moisture-proof corrugated paper generation model is trained;

[0059] Acquire a second corrugated paper image, where the second corrugated paper image is not filled with moisture-proof material, crawl second order destination information and obtain second corrugated paper production equipment parameters, and perform multimodal parameter fusion to obtain a second filling feature;

[0060] The moisture-proof corrugated paper generation model processes the second filling feature to obtain a second moisture-proof material grade filling method that does not require manual setting by a technician, and uses the second moisture-proof material grade filling method to fill the corrugated paper with moisture-proof material.

[0061] In this embodiment, the first moisture-proof material grade filling method or the second moisture-proof material grade filling method is an artificially set pearl cotton filling grade, and the pearl cotton filling grade includes the thickness of the pearl cotton. Specifically, the pearl cotton filling with a thickness of 1mm-3mm is set as the first grade, the pearl cotton filling with a thickness of 4mm-6mm is set as the second grade, and the pearl cotton filling with a thickness of 7mm-10mm is set as the third grade.

[0062] Furthermore, the moisture-proof corrugated paper generation model is constructed using machine learning or deep learning methods.

[0063] Furthermore, the first corrugated paper image or the second corrugated paper image is an image of a paperboard interface having a waterproof film, a face paper, a fireproof core layer and a support column taken by a camera device installed on the top of a factory building. Figure 3 shown.

[0064] In this embodiment, since the corrugated paper production line is long and occupies a huge area, a camera is only installed on the ceiling of the factory at the entrance of the pearl cotton filling to capture the first corrugated paper image or the second corrugated paper image and transmit the data to the computer processing terminal using the factory's Internet.

[0065] Furthermore, before using the first corrugated paper image or the second corrugated paper image for multimodal parameter fusion, the first corrugated paper image or the second corrugated paper image is subjected to focusing processing of the target image and the background image to obtain the first corrugated paper image input vector or the second corrugated paper image input vector, and the image focusing model is used for image segmentation, and the image obtained after the image focusing is converted into a vector value, wherein the image focusing model realizes image focusing by extremizing the following wave function M:

[0066]

[0067] Where, I(X) is the first corrugated paper image or the second corrugated paper image to be focused, C is the evolution curve, μ, v, λ 1 ,λ 2 are all non-negative coefficients, c 1 , c 2 are the average grayscale values ​​of image I in the target and background images, in(C) is the target area, out(C) is the background area, lE is the length constraint, a is the area constraint, and c is the average grayscale value of image I in the target and background images. 1 , c 2 The specific expression is as follows:

[0068]

[0069] in(C) and out(C) are the regional images obtained by focusing. in(C) and out(C) are binarized again and converted into vector representations, where in(C) is a corrugated paper interface image that only contains the first corrugated paper image or the second corrugated paper image.

[0070] In this embodiment, since the image of corrugated paper filled with moisture-proof material taken from a high altitude occupies too much of the background image except for the corrugated paper, this method is used to extract the image of the target area of ​​the corrugated paper. In this embodiment, the image obtained completely is represented by the average grayscale value of (125, 136, 180), and the average grayscale value of the target corrugated paper image obtained after focusing processing is (180, 215, 243).

[0071] Further, the first order destination information or the second order destination information includes the temperature, humidity and annual precipitation of the order destination;

[0072] In this embodiment, if the destination information of the first order is Hengyang, Hunan, the temperature, humidity and annual precipitation vector of the destination is (26, 85%, 1000), indicating that the temperature at the destination is 26°C, the average humidity is 85%, and the annual precipitation is 1000 mm when the order is delivered.

[0073] The first corrugated paper production equipment parameter or the second corrugated paper production equipment parameter includes the operating speed and pressure of the corrugated paper production equipment.

[0074] The operating speed and pressure of the corrugated paper production equipment in this embodiment are 100m / min and 0.5Mpa.

[0075] Further, the multimodal parameter fusion is vector splicing or dimension splicing, and the vector splicing is a one-dimensional vector directly linked to the first corrugated paper image input vector, the first order destination information, and the first corrugated paper production equipment parameter, or a one-dimensional vector directly linked to the second corrugated paper image input vector, the second order destination information, and the second corrugated paper production equipment parameter;

[0076] In this embodiment, the one-dimensional vector is (180, 215, 243, 26, 85%, 1000, 100, 0.5).

[0077] The dimensional splicing is a three-dimensional vector obtained by dimensionally linking the first corrugated paper image input vector, the first order destination information, and the first corrugated paper production equipment parameter, or a three-dimensional vector obtained by directly linking the second corrugated paper image input vector, the second order destination information, and the second corrugated paper production equipment parameter and performing unit vector value complementation.

[0078] In this embodiment, the three-dimensional vector is

[0079] Furthermore, the machine learning method adopts an improved classifier based on the Fisher criterion, and the calculation formula is as follows:

[0080]

[0081] A is the first filling feature or the second filling feature, g(S) is the output first moisture-proof material level filling method or the second moisture-proof material level filling method, W T is the normal vector perpendicular to the hyperplane, k is the south or north of the country to which the order destination belongs, k is 1 in the south and -1 in the north, S i is the contact area between the filling material and the corrugated paper, S o The area where the filling material does not contact the corrugated paper.

[0082] A moisture-proof corrugated paper generation system based on multi-modal parameter fusion image transformation is also provided, characterized in that the system comprises:

[0083] A corrugated paper filling image acquisition module is used to acquire a first corrugated paper image, where the first corrugated paper image is not filled with a moisture-proof material, and to acquire a first moisture-proof material grade filling method set by a technician at this time;

[0084] The corrugated paper filling processing module crawls the first order destination information using crawler technology and obtains the first corrugated paper production equipment parameters. The corrugated paper filling processing module is also connected to the corrugated paper filling image acquisition module to obtain the first corrugated paper image, and performs multimodal parameter fusion on the first corrugated paper image, the first order destination information and the first corrugated paper production equipment parameters to obtain a first filling feature; the corrugated paper filling processing module also receives the second corrugated paper image obtained by the corrugated paper filling image acquisition module, the second corrugated paper image is not filled with moisture-proof material, crawls the second order destination information and obtains the second corrugated paper production equipment parameters, and performs multimodal parameter fusion on them to obtain a second filling feature;

[0085] The moisture-proof corrugated paper generation model module receives the first filling feature and the first moisture-proof material grade filling method, and trains to obtain a moisture-proof corrugated paper generation model, wherein the moisture-proof corrugated paper generation model is a machine learning model; the moisture-proof corrugated paper generation model module also receives the second filling feature, and processes to obtain a second moisture-proof material grade filling method that does not require manual setting by a technician;

[0086] The corrugated paper filling control module receives the second moisture-proof material grade filling method and fills the corrugated paper with the moisture-proof material.

[0087] Furthermore, the moisture-proof corrugated paper generation model is constructed using a machine learning or deep learning method;

[0088] The machine learning method adopts an improved classifier based on the Fisher criterion, and the calculation formula is as follows:

[0089]

[0090] A is the first filling feature or the second filling feature, g(S) is the output first moisture-proof material level filling method or the second moisture-proof material level filling method, W T is the normal vector perpendicular to the hyperplane, k is the south or north of the country to which the order destination belongs, k is 1 in the south and -1 in the north, S i is the contact area between the filling material and the corrugated paper, S o The area where the filling material does not contact the corrugated paper.

[0091] In this embodiment, if Hengyang, Hunan is selected as the order destination, k is 1, and if g(S) is less than 0, the first level is output, if g(S) is equal to 0, the second level is output, and if g(S) is greater than 0, the third level is output. The corrugated paper filling control module fills the moisture-proof material with the corresponding thickness according to the output level.

[0092] Furthermore, before using the first corrugated paper image or the second corrugated paper image for multimodal parameter fusion, the first corrugated paper image or the second corrugated paper image is subjected to focusing processing of the target image and the background image to obtain the first corrugated paper image input vector or the second corrugated paper image input vector, and the image focusing model is used for image segmentation, and the image obtained after the image focusing is converted into a vector value, wherein the image focusing model realizes image focusing by extremizing the following wave function M:

[0093]

[0094] Where, I(X) is the first corrugated paper image or the second corrugated paper image to be focused, C is the evolution curve, μ, v, λ 1 ,λ 2 are all non-negative coefficients, c 1 , c 2 are the average grayscale values ​​of image I in the target and background images, in(C) is the target area, out(C) is the background area, lE is the length constraint, a is the area constraint, and c is the average grayscale value of image I in the target and background images. 1 , c 2 The specific expression is as follows:

[0095]

[0096] in(C) and out(C) are the regional images obtained by focusing. in(C) and out(C) are binarized again and converted into vector representations, where in(C) is a corrugated paper interface image that only contains the first corrugated paper image or the second corrugated paper image.

[0097] The present invention discloses a method and system for generating moisture-proof corrugated paper by multi-modal parameter fusion image transformation. By combining order information, processed transformed image information and corrugated paper equipment parameter data, the method of using moisture-proof materials with sufficient experience is finely controlled, and data processing of corrugated paper filling materials with multiple modal data is performed. The method of using moisture-proof filling materials is obtained more accurately in the classification method, and the order information is fully utilized to perform corresponding production control of moisture-proof corrugated paper, which can well save the cost of using moisture-proof materials, improve customer satisfaction and reduce the cost of enterprises.

[0098] Of course, it can be understood that each embodiment of the present invention can achieve one of the effects alone, and a combination of multiple embodiments of the present invention can achieve all of the above effects, but it is not required that every embodiment of the present invention achieve all of the above advantages and effects, because each embodiment of the present invention can constitute a separate technical solution and make one or more contributions to the prior art.

[0099] The present invention does not particularly specify the structure of some modules, which shall be subject to the contents recorded in the prior art. The prior art mentioned in the above background technology section and the specific embodiment section of the present invention can be used as part of the present invention to understand the meaning of some technical features or parameters. The scope of protection of the present invention shall be subject to the contents actually recorded in the claims.

Claims

1. A method for generating moisture-proof corrugated paper by multimodal parameter fusion image transformation, characterized in that: Acquire a first corrugated paper image, and acquire a first moisture-proof material grade filling method set by a technician at this time, wherein the first corrugated paper image is not filled with moisture-proof material; Using crawler technology to crawl the first order destination information and obtain the first corrugated paper production equipment parameters, and performing multimodal parameter fusion on the first corrugated paper image, the first order destination information and the first corrugated paper production equipment parameters to obtain a first filling feature; Using the first filling feature and the first moisture-proof material grade filling method, a moisture-proof corrugated paper generation model is trained; Acquire a second corrugated paper image, where the second corrugated paper image is not filled with moisture-proof material, crawl second order destination information and obtain second corrugated paper production equipment parameters, and perform multimodal parameter fusion to obtain a second filling feature; The moisture-proof corrugated paper generation model processes the second filling feature to obtain a second moisture-proof material grade filling method that does not require manual setting by a technician, and uses the second moisture-proof material grade filling method to fill the corrugated paper with moisture-proof material.

2. The method for generating moisture-proof corrugated paper by multimodal parameter fusion image transformation as claimed in claim 1, characterized in that: The moisture-proof corrugated paper generation model is constructed using a machine learning or deep learning method.

3. The method for generating moisture-proof corrugated paper by multimodal parameter fusion image transformation as claimed in claim 2, characterized in that: The first corrugated paper image or the second corrugated paper image is an image of a paperboard interface having a waterproof membrane, a face paper, a fireproof core layer and a support column, taken by a camera device installed on the top of a factory building.

4. The method for generating moisture-proof corrugated paper by multimodal parameter fusion image transformation as claimed in claim 1, characterized in that: Before using the first corrugated paper image or the second corrugated paper image for multimodal parameter fusion, the first corrugated paper image or the second corrugated paper image is subjected to focusing processing of the target image and the background image to obtain the first corrugated paper image input vector or the second corrugated paper image input vector, and the image focusing model is used to perform image segmentation, and the image obtained after the image focusing is converted into a vector value, wherein the image focusing model realizes image focusing by extremizing the following wave function M: M(c1, c2, C) =μ·lE(C)+v·a(in(C))+λ1∫ in(C) ||I(x)-c1||dx+λ2∫ out(C) ||I(x)-c2||dx Wherein, I(X) is the first corrugated paper image or the second corrugated paper image to be focused, C is the evolution curve, μ, v, λ1, λ2 are all non-negative coefficients, c1, c2 are the average grayscale values ​​of image I in the target and background images, in(C) is the target area, out(C) is the background area, lE is the length constraint, a is the area constraint, and the specific expressions of the average grayscale values ​​c1, c2 of image I in the target and background images are as follows: in(C) and out(C) are the regional images obtained by focusing. in(C) and out(C) are binarized again and converted into vector representations, where in(C) is a corrugated paper interface image that only contains the first corrugated paper image or the second corrugated paper image.

5. The method for generating moisture-proof corrugated paper by multimodal parameter fusion image transformation as claimed in claim 4, characterized in that: The first order destination information or the second order destination information includes the temperature, humidity and annual precipitation of the order destination; The first corrugated paper production equipment parameter or the second corrugated paper production equipment parameter includes the operating speed and pressure of the corrugated paper production equipment.

6. The method for generating moisture-proof corrugated paper by multimodal parameter fusion image transformation as claimed in claim 5, characterized in that: The multimodal parameter fusion is vector splicing or dimension splicing, and the vector splicing is a one-dimensional vector directly linked to the first corrugated paper image input vector, the first order destination information, and the first corrugated paper production equipment parameter, or a one-dimensional vector directly linked to the second corrugated paper image input vector, the second order destination information, and the second corrugated paper production equipment parameter; The dimensional splicing is a three-dimensional vector obtained by dimensionally linking the first corrugated paper image input vector, the first order destination information, and the first corrugated paper production equipment parameter, or a three-dimensional vector obtained by directly linking the second corrugated paper image input vector, the second order destination information, and the second corrugated paper production equipment parameter.

7. The method for generating moisture-proof corrugated paper by multimodal parameter fusion image transformation as claimed in claim 6, characterized in that: The machine learning method adopts an improved classifier based on the Fisher criterion, and the calculation formula is as follows: A is the first filling feature or the second filling feature, g(S) is the output first moisture-proof material level filling method or the second moisture-proof material level filling method, W T is the normal vector perpendicular to the hyperplane, k is the south or north of the country to which the order destination belongs, k is 1 in the south and -1 in the north, S i is the contact area between the filling material and the corrugated paper, S o The area where the filling material does not contact the corrugated paper.

8. A moisture-proof corrugated paper generation system based on multimodal parameter fusion image transformation, characterized in that: The system comprises: A corrugated paper filling image acquisition module is used to acquire a first corrugated paper image, where the first corrugated paper image is not filled with a moisture-proof material, and to acquire a first moisture-proof material grade filling method set by a technician at this time; The corrugated paper filling processing module crawls the first order destination information using crawler technology and obtains the first corrugated paper production equipment parameters. The corrugated paper filling processing module is also connected to the corrugated paper filling image acquisition module to obtain the first corrugated paper image, and performs multimodal parameter fusion on the first corrugated paper image, the first order destination information and the first corrugated paper production equipment parameters to obtain a first filling feature; the corrugated paper filling processing module also receives the second corrugated paper image obtained by the corrugated paper filling image acquisition module, the second corrugated paper image is not filled with moisture-proof material, crawls the second order destination information and obtains the second corrugated paper production equipment parameters, and performs multimodal parameter fusion on them to obtain a second filling feature; The moisture-proof corrugated paper generation model module receives the first filling feature and the first moisture-proof material grade filling method, and trains to obtain a moisture-proof corrugated paper generation model, wherein the moisture-proof corrugated paper generation model is a machine learning model; the moisture-proof corrugated paper generation model module also receives the second filling feature, and processes to obtain a second moisture-proof material grade filling method that does not require manual setting by a technician; The corrugated paper filling control module receives the second moisture-proof material grade filling method and fills the corrugated paper with the moisture-proof material.

9. The moisture-proof corrugated paper generation system of multimodal parameter fusion image transformation as claimed in claim 8, characterized in that: The moisture-proof corrugated paper generation model is constructed using a machine learning or deep learning method; The machine learning method adopts an improved classifier based on the Fisher criterion, and the calculation formula is as follows: A is the first filling feature or the second filling feature, g(S) is the output first moisture-proof material level filling method or the second moisture-proof material level filling method, W T is the normal vector perpendicular to the hyperplane, k is the south or north of the country to which the order destination belongs, k is 1 in the south and -1 in the north, S i is the contact area between the filling material and the corrugated paper, S o The area where the filling material does not contact the corrugated paper.

10. The moisture-proof corrugated paper generation system of multi-modal parameter fusion image transformation as claimed in claim 9, characterized in that: Before using the first corrugated paper image or the second corrugated paper image for multimodal parameter fusion, the first corrugated paper image or the second corrugated paper image is subjected to focusing processing of the target image and the background image to obtain the first corrugated paper image input vector or the second corrugated paper image input vector, and the image focusing model is used to perform image segmentation, and the image obtained after the image focusing is converted into a vector value, wherein the image focusing model realizes image focusing by extremizing the following wave function M: M(c1, c2, C) =μ·lE(C)+v·a(in(C))+λ1∫ in(C) ||I(x)-c1||dx+λ2∫ out(C) ||I(x)-c2||dx Wherein, I(X) is the first corrugated paper image or the second corrugated paper image to be focused, C is the evolution curve, μ, v, λ1, λ2 are all non-negative coefficients, c1, c2 are the average grayscale values ​​of image I in the target and background images, in(C) is the target area, out(C) is the background area, lE is the length constraint, a is the area constraint, and the specific expressions of the average grayscale values ​​c1, c2 of image I in the target and background images are as follows: in(C) and out(C) are the regional images obtained by focusing. in(C) and out(C) are binarized again and converted into vector representations, where in(C) is a corrugated paper interface image that only contains the first corrugated paper image or the second corrugated paper image.

Images