Ink-jet printing ink droplet state prediction method and device based on small sample learning

Abstract

The embodiment of the invention provides an ink-jet printing ink droplet state prediction method and device based on small sample learning, and the method comprises the steps: S10, carrying out the testing of the jet ink droplet of ink-jet printing through employing testing ink, and obtaining m groups of feature parameters corresponding to the testing ink through changing the parameter value of the driving voltage of an ink-jet printing head; s20, the printed electronic ink is used for conducting jet ink droplet testing of ink-jet printing, and n sets of characteristic parameters corresponding to the printed electronic ink are obtained by changing the parameter value of the driving voltage of an ink-jet printing head; s30, training the ink droplet state prediction model by adopting a small sample learning method based on the m groups of feature parameters corresponding to the test ink and the n groups of feature parameters corresponding to the printed electronic ink to obtain a trained ink droplet state prediction model; s40, performing ink droplet state prediction on ink droplets to be predicted by using the trained ink droplet state prediction model; the method has the beneficial effect of accurately predicting the ink droplet jetting state, and is suitable for the technical field of ink-jet printing.

Description

Technical Field

[0001] The present application relates to the technical field of inkjet printing, and in particular to a method and device for predicting ink droplet states in inkjet printing based on small sample learning. Background Art

[0002] In recent years, inkjet printing has been widely used in the manufacture of electronic devices due to its advantages, including simple processing, high material utilization, and wide substrate adaptability. The performance of electronic devices is significantly affected by the deposited morphology of the inkjet-printed material. Therefore, precise control of inkjet droplets is a prerequisite for manufacturing precision printed devices.

[0003] Currently, most printed electronic inkjet printers are based on the piezoelectric inkjet principle. This involves applying a driving voltage signal to the piezoelectric material surrounding the nozzle, causing the material to deform and squeeze the ink out of the ink cavity, forming ink droplets. Due to manufacturing tolerances, the volume of ejected ink droplets from current high-performance piezoelectric inkjet printheads, without individually adjusting the nozzle drive voltage, varies by approximately ±10%. This accuracy cannot meet the inkjet printing manufacturing requirements for precision electronic devices (for example, OLED devices require a film thickness tolerance of less than ±0.6%).

[0004] Generally, the ejection state of inkjet printing droplets is related to the printhead specifications, drive voltage waveform, ink physicochemical properties (density, viscosity, surface tension, etc.), and the printing environment (temperature, humidity, air pressure, water and oxygen, etc.). For a given inkjet printing application, the printhead specifications generally remain unchanged. However, the impact of ink physicochemical properties and printing conditions on droplet ejection during the printing process is difficult to quantify. Therefore, precise control of droplet ejection is typically achieved by adjusting the drive voltage waveform.

[0005] like Figure 1 As shown in the figure, the driving voltage signal of the inkjet print head is related to its internal structure. The driving voltage signal can be regarded as a deformation and combination of a trapezoidal wave. The key parameters of the trapezoidal wave can be defined as: voltage rise time , voltage duration , voltage fall time , voltage amplitude V. Currently, the adjustment and optimization of the drive waveform is usually based on manual operation guided by experience, namely: First, according to the printhead manual and the physical and chemical characteristics of the ink, a driving waveform is designed and loaded into the inkjet printhead to eject ink droplets. Then, the ink drop state is measured by an ink drop observation device. By comparing the measured ink drop state with the assumed ink drop state, one or several parameter values ​​of the driving waveform are manually adjusted to achieve precise control of the ink drop ejection state.

[0006] Since the mechanism by which driving voltage parameters influence the ink droplet ejection state is unclear, optimizing the driving voltage waveform manually under empirical guidance is time-consuming and labor-intensive, and it is difficult to achieve precise control. In addition, inkjet printing inks for electronic devices are usually expensive and difficult to support a large number of trial-and-error experiments. Therefore, how to achieve precise prediction and control of ink droplet states at a relatively low cost is of great significance to the development and promotion of printed electronics technology. Summary of the Invention

[0007] In order to solve one of the above-mentioned technical defects, an embodiment of the present application provides an inkjet printing ink drop state prediction method and device based on small sample learning to accurately predict the ink drop ejection state in the inkjet printing manufacturing process of electronic devices through small sample learning.

[0008] According to a first aspect of an embodiment of the present application, a method for predicting ink droplet states in inkjet printing based on small sample learning is provided, comprising: S10, performing an ink droplet ejection test of inkjet printing using the test ink, and obtaining m groups of characteristic parameters corresponding to the test ink by changing the parameter value of the driving voltage of the inkjet print head; S20, performing an inkjet printing ink droplet ejection test using the printed electronic ink, and obtaining n sets of characteristic parameters corresponding to the printed electronic ink by changing a parameter value of a driving voltage of the inkjet print head; S30, training an ink drop state prediction model using a small sample learning method based on the m groups of characteristic parameters corresponding to the test ink and the n groups of characteristic parameters corresponding to the printed electronic ink to obtain a trained ink drop state prediction model; S40, using the trained ink drop state prediction model to predict the ink drop state of the ink drop to be predicted; wherein the density, surface tension and viscosity coefficient of the test ink are similar to those of the printed electronic ink; Each set of characteristic parameters corresponds to a parameter value of the driving voltage; each set of characteristic parameters includes: the volume and speed of the ejected ink droplets; The n is much smaller than m.

[0009] Preferably, before S10, the method further includes: Measure the density, surface tension and viscosity of printed electronic ink; The test ink is configured according to the density, surface tension and viscosity coefficient of the printed electronic ink.

[0010] Preferably, the driving voltage includes: a first trapezoidal voltage having a positive voltage and a second trapezoidal voltage having a negative voltage; There is a waveform interval time between the first trapezoidal voltage and the second trapezoidal voltage .

[0011] Preferably, the first trapezoidal voltage includes: a first voltage rise time , First voltage duration , First voltage drop time , the first voltage duration The first voltage amplitude within ; The second trapezoidal voltage includes: a second voltage falling time , the second voltage duration , the second voltage rise time , the second voltage duration Voltage amplitude within .

[0012] Preferably, the step S30 includes: S301, using m groups of driving voltage parameter values ​​corresponding to the m groups of characteristic parameters as input and the m groups of characteristic parameters corresponding to the test ink as output, to train an ink drop state prediction model to obtain a pre-trained model; S302, performing feature learning on the m groups of feature parameters and the m groups of driving voltage parameter values ​​corresponding to the m groups of feature parameters to obtain potential distribution features of the sample data; S303, based on the potential distribution characteristics of the sample data, using a data generation model to synthesize new sample data with a distribution similar to the sample data; S304 , tuning the parameters of the pre-trained model using the n groups of driving voltage parameter values ​​corresponding to the n groups of characteristic parameters and the synthesized new sample data to obtain a trained ink drop state prediction model.

[0013] Preferably, the pre-trained model includes s hidden layers; and S304 includes: S304-11, freezing the first sk hidden layers in the pre-trained model so that the parameters of the frozen hidden layers remain unchanged; S304-12, training and optimizing the adjustable parameters of the pre-trained model using the n sets of driving voltage parameter values ​​corresponding to the n sets of characteristic parameters and the synthesized new sample data to obtain a trained ink drop state prediction model; The adjustable parameters include: the s-k+1th to sth hidden layers, the threshold of each hidden layer, the weight between the sth hidden layer and the output layer, and the threshold of the output layer.

[0014] Preferably, the pre-trained model includes s hidden layers; and S304 includes: S304-21, freeze the first s hidden layers in the pre-trained model so that the parameters of the frozen hidden layers remain unchanged; S304-22, adding an additional layer between the sth hidden layer and the output layer, wherein the additional layer includes r neurons; S304-23, training and optimizing the adjustable parameters of the pre-trained model using the n sets of driving voltage parameter values ​​corresponding to the n sets of characteristic parameters and the synthesized new sample data to obtain a trained ink drop state prediction model; The adjustable parameters include: the weight between the sth hidden layer and the added layer, the threshold of the added layer, the weight between the added layer and the output layer, and the threshold of the output layer .

[0015] Preferably, the data generation model includes at least one of an interpolation model, a generative adversarial network model, and a data enhancement model.

[0016] According to a second aspect of an embodiment of the present application, there is provided an inkjet printing ink drop state prediction system based on small sample learning, comprising: a main control computer, a print controller, an inkjet print head, and an ink drop observer; The main control computer is used to send parameter values ​​for changing the driving voltage of the inkjet print head to the print controller during an ink droplet ejection test of inkjet printing using the test ink, so as to obtain m groups of characteristic parameters corresponding to the test ink; and, during an ink droplet ejection test for inkjet printing using printed electronic ink, sending a parameter value for changing a driving voltage of an inkjet print head to a print controller to obtain n sets of characteristic parameters corresponding to the printed electronic ink; and, based on the m groups of characteristic parameters corresponding to the test ink and the n groups of characteristic parameters corresponding to the printed electronic ink, using a small sample learning method to train an ink drop state prediction model to obtain a trained ink drop state prediction model; and, using the trained ink drop state prediction model to predict the ink drop state of the ink drop to be predicted; The density, surface tension and viscosity coefficient of the test ink are similar to those of the printed electronic ink; each set of characteristic parameters corresponds to a parameter value of the driving voltage; each set of characteristic parameters includes: the volume and velocity of the ejected ink droplets; n is much smaller than m; The printing controller is used to control the inkjet print head to eject ink droplets according to the parameter value of the driving voltage sent by the host computer; The ink droplet observation instrument is used to obtain ink droplet image data and send the ink droplet image data to a main control computer, so that the main control computer can obtain characteristic parameters corresponding to the ink droplet image data according to the ink droplet image data.

[0017] According to a third aspect of the embodiments of the present application, there is provided an electronic device, including: memory; processor; and computer program; The computer program is stored in the memory and configured to be executed by the processor to implement the method described above.

[0018] The inkjet printing droplet state prediction method and device based on small sample learning provided in the embodiments of the present application are used, and the inkjet printing droplet test is performed using test ink. By changing the parameter value of the driving voltage of the inkjet print head, m groups of characteristic parameters corresponding to the test ink are obtained; the inkjet printing droplet test is performed using printed electronic ink, and by changing the parameter value of the driving voltage of the inkjet print head, n groups of characteristic parameters corresponding to the printed electronic ink are obtained; this embodiment uses a small sample learning method to train the ink droplet state prediction model, and obtains a trained ink droplet state prediction model, which can reduce the waste in the trial and error process of using printed electronic ink to jet ink, improve the prediction accuracy of the ink droplet state, and is extremely practical. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings: Figure 1 Schematic diagram of driving trapezoidal wave; Figure 2 A flowchart of a method for predicting ink droplet states in inkjet printing based on small sample learning provided in an embodiment of the present application; Figure 3 Schematic diagram of the waveform of the driving voltage in the embodiment of the present application; Figure 4 A schematic diagram of the process of S30 in a method for predicting ink droplet states in inkjet printing based on small sample learning provided in an embodiment of the present application; Figure 5 is a schematic diagram of the process of S304 in a specific embodiment; Figure 6 is a schematic structural diagram of the pre-training model in a specific embodiment; Figure 7 is a schematic diagram of the process of S304 in another specific embodiment; Figure 8 is a schematic structural diagram of the pre-training model in another specific embodiment; Figure 9 A schematic diagram of the structure of an inkjet printing ink drop state prediction system based on small sample learning provided in an embodiment of the present application; In the picture: 310 is a main control computer, 320 is a printing controller, 330 is an inkjet print head, and 350 is an ink drop observation instrument; 340 is the ink drop; 3510 is the light source, and 3520 is the camera. DETAILED DESCRIPTION

[0020] In order to make the technical solutions and advantages of the embodiments of the present application more clearly understood, the exemplary embodiments of the present application are further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, and are not an exhaustive list of all the embodiments. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other unless they conflict.

[0021] In the process of implementing this application, the inventors found that due to the unclear mechanism of the influence of the driving voltage parameters on the ink droplet ejection state, it is time-consuming and labor-intensive to optimize the driving voltage waveform through manual operation guided by experience, and it is difficult to achieve precise control effects. Inkjet printing inks for electronic devices are usually expensive and difficult to support a large number of trial and error experiments.

[0022] like Figure 2 As shown, an embodiment of the present application provides an inkjet printing ink drop state prediction method based on small sample learning, including: S10, performing an ink droplet ejection test of inkjet printing using the test ink, and obtaining m groups of characteristic parameters corresponding to the test ink by changing the parameter value of the driving voltage of the inkjet print head; S20, performing an inkjet printing ink droplet ejection test using the printed electronic ink, and obtaining n sets of characteristic parameters corresponding to the printed electronic ink by changing a parameter value of a driving voltage of the inkjet print head; S30, training an ink drop state prediction model using a small sample learning method based on the m groups of characteristic parameters corresponding to the test ink and the n groups of characteristic parameters corresponding to the printed electronic ink to obtain a trained ink drop state prediction model; S40, using the trained ink drop state prediction model to predict the ink drop state of the ink drop to be predicted; wherein the density, surface tension and viscosity coefficient of the test ink are similar to those of the printed electronic ink; Each set of characteristic parameters corresponds to a parameter value of the driving voltage; each set of characteristic parameters includes: the volume and speed of the ejected ink droplets; The n is much smaller than m.

[0023] Specifically, before S10, the method further includes: Measure the density, surface tension and viscosity of printed electronic ink; The test ink is configured according to the density, surface tension and viscosity coefficient of the printed electronic ink.

[0024] In this embodiment, an ink droplet test for inkjet printing is performed using test ink, and m groups of characteristic parameters corresponding to the test ink are obtained by changing the parameter value of the driving voltage of the inkjet print head; an ink droplet test for inkjet printing is performed using printed electronic ink, and n groups of characteristic parameters corresponding to the printed electronic ink are obtained by changing the parameter value of the driving voltage of the inkjet print head; this embodiment adopts a small sample learning method to train the ink droplet state prediction model, and obtains a trained ink droplet state prediction model, which can reduce the waste in the trial and error process of using printed electronic ink to jet ink, improve the prediction accuracy of the ink droplet state, and is extremely practical.

[0025] Generally speaking, the physical and chemical parameters of ink, such as density, viscosity coefficient and surface tension, are closely related to the state of inkjet printing droplets. When configuring test ink, a solvent with a density, viscosity coefficient and surface tension close to those of the target printed electronic ink should be selected. The parameter difference can be controlled within ±10%.

[0026] The flow characteristics of inkjet-printed ink droplets in the process of forming the printed pattern are determined by two dimensionless numbers, Re and We, which are expressed as follows:

[0027]

[0028] in, is the velocity of ejecting ink droplets, is the diameter of the ejected ink droplet, is the density of ink, is the surface tension of the ink, is the viscosity coefficient of the ink.

[0029] In preparation for inkjet printing, the density, viscosity coefficient and surface tension of the ink are measured to obtain the required ink droplet diameter and speed. The volume of the ink droplet can be obtained from the ink droplet diameter.

[0030] like Figure 3 As shown, specifically, the driving voltage includes: a first trapezoidal voltage with a positive voltage and a second trapezoidal voltage with a negative voltage; there is a waveform interval time between the first trapezoidal voltage and the second trapezoidal voltage. .

[0031] The driving voltage may also be Figure 1 A more complex combination of ladder voltages is shown.

[0032] Furthermore, the first trapezoidal voltage includes: a first voltage rise time , First voltage duration , First voltage drop time , the first voltage duration The first voltage amplitude within ; The second trapezoidal voltage includes: a second voltage falling time , the second voltage duration , the second voltage rise time , the second voltage duration Voltage amplitude within .

[0033] In this embodiment, changing the parameter value of the driving voltage of the inkjet print head includes: On the basis of stable inkjet, fine-tune the first voltage rise time , First voltage duration , First voltage drop time , the first voltage duration The first voltage amplitude within , and the second voltage fall time , the second voltage duration , the second voltage rise time , the second voltage duration Voltage amplitude within , and waveform interval time The specific parameter values ​​of the driving voltage are used to obtain the ejection state of the ink droplets under different waveforms, thereby obtaining m groups of characteristic parameters corresponding to the test ink and n groups of characteristic parameters corresponding to the printed electronic ink.

[0034] like Figure 4 As shown, in this embodiment, the S30 includes: S301, using m groups of driving voltage parameter values ​​corresponding to the m groups of characteristic parameters as input and the m groups of characteristic parameters corresponding to the test ink as output, to train an ink drop state prediction model to obtain a pre-trained model; S302, performing feature learning on the m groups of feature parameters and the m groups of driving voltage parameter values ​​corresponding to the m groups of feature parameters to obtain potential distribution features of the sample data; S303, based on the potential distribution characteristics of the sample data, using a data generation model to synthesize new sample data with a distribution similar to the sample data; S304 , tuning the parameters of the pre-trained model using the n groups of driving voltage parameter values ​​corresponding to the n groups of characteristic parameters and the synthesized new sample data to obtain a trained ink drop state prediction model.

[0035] In this embodiment, the pre-trained model includes s hidden layers, each of which may include q nodes; the i-th neuron in the input layer The weight of the hth neuron in the first hidden layer is , the threshold of the hth neuron in the first hidden layer is ; The input of the hth neuron in the first hidden layer is: ; The output is: ;in is the activation function.

[0036] Therefore, the output value of the sth hidden layer can be calculated from the output of the s-1th hidden layer, the weights between the neurons in the s-1th layer and the sth layer, and the threshold of the neurons in the sth layer; the tth parameter of the output layer can be calculated from the output of the sth hidden layer Calculated by the following formula:

[0037] in: is the threshold of the t-th neuron in the output layer.

[0038] In this embodiment, in S304, the parameters of the pre-trained model are tuned, which may include multiple implementation methods.

[0039] In a specific embodiment, Figure 5 、 Figure 6 As shown, the S304 includes: S304-11, freezing the first sk hidden layers in the pre-trained model so that the parameters of the frozen hidden layers remain unchanged; S304-12, training and optimizing the adjustable parameters of the pre-trained model using the n sets of driving voltage parameter values ​​corresponding to the n sets of characteristic parameters and the synthesized new sample data to obtain a trained ink drop state prediction model; The adjustable parameters include: the s-k+1th to sth hidden layers, the threshold of each hidden layer, the weight between the sth hidden layer and the output layer, and the threshold of the output layer.

[0040] In another specific embodiment, Figure 7 、 Figure 8 As shown, the S304 may include: S304-21, freeze the first s hidden layers in the pre-trained model so that the parameters of the frozen hidden layers remain unchanged; S304-22, adding an additional layer between the sth hidden layer and the output layer, wherein the additional layer includes r neurons; S304-23, training and optimizing the adjustable parameters of the pre-trained model using the n sets of driving voltage parameter values ​​corresponding to the n sets of characteristic parameters and the synthesized new sample data to obtain a trained ink drop state prediction model; The adjustable parameters include: the weight between the sth hidden layer and the added layer, the threshold of the added layer, the weight between the added layer and the output layer, and the threshold of the output layer .

[0041] Specifically, the additional layer may include r neurons, and the output of the jth neuron in the additional layer is It can be calculated by the following formula:

[0042] The output layer can be calculated by the following formula:

[0043] in, is the weight between the sth hidden layer and the added layer, is the weight between the addition layer and the output layer, is the threshold of the j-th neuron in the added layer, is the threshold of the output layer.

[0044] In addition, tuning the parameters of the pre-trained model can also include a combination of the above two schemes, namely: retaining one or several hidden layers adjacent to the output layer in the pre-trained model, freezing the remaining hidden layers, and adding one or more additional layers (custom networks) between the hidden layer and the output layer.

[0045] In this embodiment, the data generation model includes at least one of an interpolation model, a generative adversarial network model, and a data enhancement model.

[0046] When using the interpolation model, the distribution characteristic values ​​of the m groups of characteristic parameters corresponding to the test ink can be used to estimate the missing values ​​of the n groups of characteristic parameters corresponding to the printed electronic ink, and interpolation techniques such as linear interpolation, polynomial interpolation or spline interpolation can be used to insert new data into the n groups of characteristic parameters.

[0047] When using a generative adversarial network model, new sample data can be generated by learning the distribution characteristics of n groups of feature parameters, and the generated new sample data can be discriminated using the n groups of feature parameters, thereby generating new sample data that is similar to the ink droplet ejection data characteristics of electronic ink.

[0048] When using a data augmentation model, new samples that are similar to but slightly changed from the small sample data can be generated by applying transformations and perturbations such as rotation, translation, scaling, and flipping to m groups of feature parameters.

[0049] like Figure 9 As shown, the present application also provides an inkjet printing ink drop state prediction system based on small sample learning, including: a main control computer 310, a printing controller 320, an inkjet print head 330 and an ink drop observer 350; The main control computer 310 is used to send parameter values ​​for changing the driving voltage of the inkjet print head to the print controller 320 during the ink droplet ejection test of inkjet printing using the test ink, so as to obtain m sets of characteristic parameters corresponding to the test ink; and, during an ink droplet ejection test for inkjet printing using printed electronic ink, sending a parameter value for changing a driving voltage of an inkjet print head to the print controller 320 to obtain n sets of characteristic parameters corresponding to the printed electronic ink; and, based on the m groups of characteristic parameters corresponding to the test ink and the n groups of characteristic parameters corresponding to the printed electronic ink, using a small sample learning method to train an ink drop state prediction model to obtain a trained ink drop state prediction model; and, using the trained ink drop state prediction model to predict the ink drop state of the ink drop to be predicted; The density, surface tension and viscosity coefficient of the test ink are similar to those of the printed electronic ink; each set of characteristic parameters corresponds to a parameter value of the driving voltage; each set of characteristic parameters includes: the volume and velocity of the ejected ink droplets; n is much smaller than m; The printing controller 320 is used to control the inkjet print head 330 to eject ink droplets according to the parameter value of the driving voltage sent by the host computer 310; The ink droplet imager 350 is used to obtain ink droplet image data and send the ink droplet image data to the main control computer 310 so that the main control computer 310 can obtain characteristic parameters corresponding to the ink droplet image data according to the ink droplet image data.

[0050] In this embodiment, the ink droplet imager 350 may include a light source 3510 and a camera 3520. When the print controller 320 controls the inkjet print head 330 to eject ink droplets 340, control signals from the print controller 320 may be connected to the light source 3510 and the camera 3520 of the ink droplet imager 350, respectively, to ensure that images are captured simultaneously with ink droplet ejection. The ink droplet imager 350 then transmits the captured ink droplet images back to the host computer 310.

[0051] In this embodiment, the waveform of the driving voltage can be designed according to the inkjet print head manual.

[0052] An embodiment of the present application further provides an electronic device, including: memory; processor; and computer program; The computer program is stored in the memory and configured to be executed by the processor to implement the method described above.

[0053] An embodiment of the present application further provides a computer-readable storage medium on which a computer program is stored; the computer program is executed by a processor to implement the method described above.

[0054] In the embodiments of the present application, the method, system, electronic device and computer-readable storage medium are based on the same inventive concept. Since the principles of solving problems by the method, system, electronic device and computer-readable storage medium are similar, the implementation of the method, system, electronic device and computer-readable storage medium can refer to each other, and the repeated parts will not be repeated.

[0055] In summary, this application uses a small sample learning method to accurately predict the state of expensive printed electronic ink inkjet printing droplets, which can achieve accurate prediction and control of the state of inkjet printing droplets with less time and economic cost, and is of great significance to the development and promotion of printed electronic technology.

[0056] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of the present application may be implemented in various computer languages, such as C, VHDL, Verilog, object-oriented programming language Java, and interpreted scripting language JavaScript.

[0057] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0058] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0059] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0060] In this application, unless otherwise expressly specified or limited, terms such as "mounted," "connected," "connect," and "fixed" should be understood broadly. For example, they may refer to fixed connections, detachable connections, or integration; mechanical connections, electrical connections, or communication; direct connections or indirect connections through an intermediate medium; and internal communication between two components or interaction between two components. Those skilled in the art will understand the specific meanings of these terms in this application based on specific circumstances.

[0061] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present application.

[0062] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. A method for predicting ink droplet states in inkjet printing based on small sample learning, characterized in that: include: S10, performing an ink droplet ejection test of inkjet printing using the test ink, and obtaining m groups of characteristic parameters corresponding to the test ink by changing the parameter value of the driving voltage of the inkjet print head; S20, performing an inkjet printing ink droplet ejection test using the printed electronic ink, and obtaining n sets of characteristic parameters corresponding to the printed electronic ink by changing a parameter value of a driving voltage of the inkjet print head; S30, training an ink drop state prediction model using a small sample learning method based on the m groups of characteristic parameters corresponding to the test ink and the n groups of characteristic parameters corresponding to the printed electronic ink to obtain a trained ink drop state prediction model; S40, using the trained ink drop state prediction model to predict the ink drop state of the ink drop to be predicted; wherein the density, surface tension and viscosity coefficient of the test ink are similar to those of the printed electronic ink; Each set of characteristic parameters corresponds to a parameter value of the driving voltage; each set of characteristic parameters includes: the volume and speed of the ejected ink droplets; The n is much smaller than m.

2. The inkjet printing ink drop state prediction method based on small sample learning according to claim 1 is characterized in that: Before S10, the method further includes: Measure the density, surface tension and viscosity of printed electronic ink; The test ink is configured according to the density, surface tension and viscosity coefficient of the printed electronic ink.

3. The inkjet printing ink drop state prediction method based on small sample learning according to claim 1, characterized in that: The driving voltage includes: a first trapezoidal voltage having a positive voltage and a second trapezoidal voltage having a negative voltage; There is a waveform interval time between the first trapezoidal voltage and the second trapezoidal voltage .

4. The inkjet printing ink drop state prediction method based on small sample learning according to claim 3, characterized in that: The first trapezoidal voltage includes: a first voltage rise time , First voltage duration , First voltage drop time , the first voltage duration The first voltage amplitude within ; The second trapezoidal voltage includes: a second voltage falling time , the second voltage duration , the second voltage rise time , the second voltage duration Voltage amplitude within .

5. The inkjet printing ink drop state prediction method based on small sample learning according to claim 1, characterized in that: The S30 includes: S301, using m groups of driving voltage parameter values ​​corresponding to the m groups of characteristic parameters as input and the m groups of characteristic parameters corresponding to the test ink as output, to train an ink drop state prediction model to obtain a pre-trained model; S302, performing feature learning on the m groups of feature parameters and the m groups of driving voltage parameter values ​​corresponding to the m groups of feature parameters to obtain potential distribution features of the sample data; S303, based on the potential distribution characteristics of the sample data, using a data generation model to synthesize new sample data with a distribution similar to the sample data; S304 , tuning the parameters of the pre-trained model using the n groups of driving voltage parameter values ​​corresponding to the n groups of characteristic parameters and the synthesized new sample data to obtain a trained ink drop state prediction model.

6. The inkjet printing ink drop state prediction method based on small sample learning according to claim 5, characterized in that: The pre-trained model includes s hidden layers; S304 includes: S304-11, freezing the first sk hidden layers in the pre-trained model so that the parameters of the frozen hidden layers remain unchanged; S304-12, training and optimizing the adjustable parameters of the pre-trained model using the n sets of driving voltage parameter values ​​corresponding to the n sets of characteristic parameters and the synthesized new sample data to obtain a trained ink drop state prediction model; The adjustable parameters include: the s-k+1th to sth hidden layers, the threshold of each hidden layer, the weight between the sth hidden layer and the output layer, and the threshold of the output layer.

7. The inkjet printing ink drop state prediction method based on small sample learning according to claim 5, characterized in that: The pre-trained model includes s hidden layers; S304 includes: S304-21, freeze the first s hidden layers in the pre-trained model so that the parameters of the frozen hidden layers remain unchanged; S304-22, adding an additional layer between the sth hidden layer and the output layer, wherein the additional layer includes r neurons; S304-23, training and optimizing the adjustable parameters of the pre-trained model using the n sets of driving voltage parameter values ​​corresponding to the n sets of characteristic parameters and the synthesized new sample data to obtain a trained ink drop state prediction model; The adjustable parameters include: the weight between the sth hidden layer and the added layer, the threshold of the added layer, the weight between the added layer and the output layer, and the threshold of the output layer .

8. The inkjet printing ink drop state prediction method based on small sample learning according to claim 5, characterized in that: The data generation model includes at least one of an interpolation model, a generative adversarial network model, and a data enhancement model.

9. An inkjet printing ink drop state prediction system based on small sample learning, characterized in that: include: A main control computer (310), a printing controller (320), an inkjet print head (330) and an ink drop observation instrument (350); The main control computer (310) is used to send parameter values ​​for changing the driving voltage of the inkjet print head to the printing controller (320) during an ink droplet ejection test using the test ink to obtain m groups of characteristic parameters corresponding to the test ink; and, during an ink droplet ejection test process of inkjet printing using printed electronic ink, sending a parameter value for changing a driving voltage of an inkjet print head to a printing controller (320) to obtain n sets of characteristic parameters corresponding to the printed electronic ink; and, based on the m groups of characteristic parameters corresponding to the test ink and the n groups of characteristic parameters corresponding to the printed electronic ink, using a small sample learning method to train an ink drop state prediction model to obtain a trained ink drop state prediction model; and, using the trained ink drop state prediction model to predict the ink drop state of the ink drop to be predicted; The density, surface tension and viscosity coefficient of the test ink are similar to those of the printed electronic ink; each set of characteristic parameters corresponds to a parameter value of the driving voltage; each set of characteristic parameters includes: the volume and velocity of the ejected ink droplets; n is much smaller than m; The printing controller (320) is used to control the inkjet print head (330) to eject ink droplets according to the parameter value of the driving voltage sent by the main control computer (310); The ink droplet observation instrument (350) is used to obtain ink droplet image data and send the ink droplet image data to the main control computer (310), so that the main control computer (310) obtains characteristic parameters corresponding to the ink droplet image data based on the ink droplet image data.

10. An electronic device, characterized in that: include: Memory; processor; as well as computer programs; The computer program is stored in the memory and configured to be executed by the processor to implement the method according to any one of claims 1 to 8.

Images