An electrical parameter prediction method, device, apparatus and storage medium

Abstract

The application relates to an electrical parameter prediction method, device and equipment and a storage medium. The method comprises the following steps: determining a parameter value of an optical parameter of a target chip; determining a parameter value of an electrical parameter of the target chip based on the parameter value of the optical parameter by using a target prediction model; and the electrical parameter comprises an electrical parameter obtained by testing the target chip by using electroluminescence testing. The electrical parameter prediction method provided by the application can improve the safety and testing efficiency of an LED chip.

Description

Technical Field

[0001] This application relates to the field of chip testing technology, and in particular to a method, apparatus, device and storage medium for predicting electrical parameters. Background Technology

[0002] With the rapid development of display technology, the market has placed increasingly higher demands on the manufacturing efficiency of light-emitting diode (LED) chips, and correspondingly, the testing efficiency of LED chips has also become increasingly demanding.

[0003] Electroluminescence (EL) testing has become a commonly used method for testing LED chips because it can measure their electrical performance. However, EL testing requires electrical contact with the LED chip, which may not only affect the safety of the LED chip but also lead to low testing efficiency. Summary of the Invention

[0004] Therefore, it is necessary to provide an electrical parameter prediction method, apparatus, computer equipment, computer-readable storage medium, and computer program product, aimed at improving the testing efficiency of LED chips.

[0005] In a first aspect, this application provides a method for predicting electrical parameters, including:

[0006] Determine the parameter values ​​of the optical parameters of the target chip;

[0007] The electrical parameters of the target chip are determined based on the optical parameters using a target prediction model. The electrical parameters include those obtained by testing the target chip using electroluminescence.

[0008] Secondly, this application also provides an electrical parameter prediction device, comprising:

[0009] The first determining module is used to determine the parameter values ​​of the optical parameters of the target chip;

[0010] The second determining module is used to determine the electrical parameters of the target chip based on the optical parameters using the target prediction model; the electrical parameters include those obtained by testing the target chip through electroluminescence testing.

[0011] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement some or all of the steps described in any method of the first aspect of the embodiments of this application.

[0012] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements some or all of the steps described in any method of the first aspect of the embodiments of this application.

[0013] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements some or all of the steps described in any method of the first aspect of the embodiments of this application.

[0014] The electrical parameter prediction method provided in this application only needs to determine the optical parameter values ​​of the target chip. Based on these optical parameter values, the electrical parameter values ​​of the target chip can be determined using a target prediction model. Therefore, EL testing of the target chip is no longer required, and the electrical parameters that would otherwise be obtained through EL testing can be obtained. This not only improves the safety of the LED chip by avoiding probe damage on the probe station and preventing damage or contamination of the LED chip's electrode surface, but also improves the testing efficiency of the LED chip because photoluminescence (PL) testing is several orders of magnitude faster than EL testing. Therefore, even at the wafer stage where the LED chip's electrodes are not yet fully fabricated, preliminary performance screening and yield prediction can be performed on LED chips in a semi-finished state, thereby significantly reducing the production cost of LED chips. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of an electroluminescence test.

[0016] Figure 2 This is a schematic diagram of the photoluminescence test.

[0017] Figure 3 This is a schematic diagram illustrating the application environment of an electrical parameter prediction method provided in an embodiment of this application.

[0018] Figure 4 A flowchart illustrating an electrical parameter prediction method provided in an embodiment of this application;

[0019] Figure 5 A flowchart illustrating another electrical parameter prediction method provided in an embodiment of this application;

[0020] Figure 6 A flowchart illustrating another electrical parameter prediction method provided in this application embodiment;

[0021] Figure 7A flowchart illustrating another electrical parameter prediction method provided in this application embodiment;

[0022] Figure 8 A flowchart illustrating another electrical parameter prediction method provided in this application embodiment;

[0023] Figure 9 This is a schematic diagram of the structure of an initial prediction model provided in an embodiment of this application;

[0024] Figure 10 This is a schematic diagram of the structure of another initial prediction model provided in an embodiment of this application;

[0025] Figure 11 This is a schematic diagram of the structure of an electrical parameter prediction device provided in an embodiment of this application;

[0026] Figure 12 An internal structural diagram of a computer device provided in an embodiment of this application;

[0027] Figure 13 This is an internal structural diagram of another computer device provided in an embodiment of this application. Detailed Implementation

[0028] To facilitate understanding of this application, a more complete description will be provided below with reference to the accompanying drawings. Preferred embodiments of this application are shown in the drawings. However, this application can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a thorough and complete understanding of the disclosure of this application.

[0029] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein in the specification of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0030] When describing positional relationships, unless otherwise specified, when an element, such as a layer, film, or substrate, is referred to as being "on" another element, it may be directly on the other element or there may be intermediate elements present. Furthermore, when a layer is referred to as being "below" another layer, it may be directly below it or there may be one or more intermediate elements present. It is also understood that when a layer is referred to as being "between" two layers, it may be the only layer between the two layers, or there may be one or more intermediate elements present.

[0031] When using the terms “including,” “having,” and “comprising” as described herein, another component may be added unless explicitly qualifying terms such as “only,” “consisting of,” etc. are used. Unless otherwise stated, singular terms may include plural forms and should not be construed as having a quantity of one.

[0032] It should be understood that although the terms “first,” “second,” etc., may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used only to distinguish one element from another. For example, without departing from the scope of this application, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element.

[0033] It should also be understood that, in interpreting an element, although not explicitly described, the element is interpreted as including a range of error, which should be within the acceptable deviation range of a particular value as determined by a person skilled in the art. For example, "approximately," "about," or "substantially" can mean within one or more standard deviations, without limitation herein.

[0034] Furthermore, in the instruction manual, the phrase "planar distribution diagram" refers to the diagram when the target part is viewed from above, and the phrase "cross-sectional diagram" refers to the diagram when the target part is viewed from the side as a cross-section taken by vertically cutting the target part.

[0035] Furthermore, the accompanying drawings are not drawn to a 1:1 scale, and the relative dimensions of the components are shown in the drawings only as examples and not necessarily to actual scale.

[0036] As described in the background section, EL testing, a common test item for LED chips, such as... Figure 1 As shown, during the testing process, a current driving signal needs to be applied between the P-region electrode and the N-region electrode of the LED chip 104 through the probe 102 on the probe station. This means there is direct electrical contact between the probe 102 and the LED chip 104. This not only may damage the LED chip 104 and affect its safety, but also results in low testing efficiency because each LED chip 104 needs to be individually sampled for EL testing. Consequently, it is difficult to determine the electrical parameters of the LED chip 104 in batches. Photoluminescence (PL) testing, another commonly used test for LED chips, such as... Figure 2As shown, the main process involves emitting pulsed laser light from the laser source 202 to excite the LED chip 204 to generate electron-hole pairs and thus generate charge carriers. The charge carriers then emit light through radiative recombination. It can be seen that since there is no direct electrical contact between the laser source 202 and the LED chip 204, PL testing obviously has higher safety for the LED chip 204. However, PL testing can only reflect the characteristics of the materials used to manufacture the LED chip 204, and cannot reflect the overall performance of the LED chip 204.

[0037] Based on the aforementioned technical problems, the inventors discovered that, under normal circumstances, the electrical parameters of an LED chip are often strongly correlated with its optical parameters. If the electrical parameters of an LED chip can be predicted from its optical parameters, the EL test (EL) of the LED chip can undoubtedly be eliminated during chip testing. Based on this, the inventors further developed the technical solution of this application. Specifically, the electrical parameter prediction method provided in this application first determines the optical parameters of the target chip, and then, based on the optical parameters, determines the electrical parameters of the target chip using a target prediction model. The electrical parameters include those obtained by testing the target chip through electroluminescence (EL) testing. Using this technical solution, only the optical parameters of the LED chip need to be determined to determine the electrical parameters using the target prediction model. Therefore, since EL testing of the target chip is no longer required, the electrical parameters of the target chip, which would otherwise require EL testing, can be obtained. Thus, this technical solution not only improves the safety of LED chips but also increases the testing efficiency of LED chips.

[0038] The above is the core idea of ​​this application. The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0039] The electrical parameter prediction method provided in this application can be applied to, for example... Figure 3In the application environment shown, terminal 302 is connected to image acquisition device 304. Terminal 302 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, or other devices with data processing capabilities. Image acquisition device 304 can be, but is not limited to, a charge-coupled device (CCD) camera or other devices with image acquisition capabilities. After the laser source 306 excites the target chip 308 to emit light, the image acquisition device 304 acquires an image of the target chip 308. After obtaining the chip image from the image acquisition device 304, terminal 302 determines the optical parameter values ​​of the target chip 308. Then, based on the optical parameter values, terminal 302 determines the electrical parameter values ​​of the target chip 308 using a target prediction model.

[0040] In one exemplary embodiment, such as Figure 4 As shown, an electrical parameter prediction method is provided, which is then applied to... Figure 3 Taking the terminal in the example, the explanation includes the following steps 402 to 404. Wherein:

[0041] Step 402: Determine the parameter values ​​of the optical parameters of the target chip.

[0042] The target chip refers to one of several LED chips whose electrical parameters need to be determined.

[0043] The optical parameters of the target chip refer to the parameters obtained through PL testing that can be used to characterize the optical properties of the target chip.

[0044] PL testing refers to the process of exciting the energy transition of an LED chip to the conduction band, forming a large number of non-equilibrium electron-hole pairs. The holes and electrons release photons through radiative recombination. Based on PL testing, the PL spectrum and optical parameters of the LED chip can be obtained.

[0045] The optical parameter values ​​of the target chip refer to the specific numerical values ​​obtained after performing PL tests on the target chip under specific PL test conditions, which can be used to quantify the optical parameters of the target chip.

[0046] Optionally, optical parameters may include brightness, wavelength, emission uniformity, etc., and may further include peak wavelength, half-width, color coordinates, etc.

[0047] The brightness of the target chip refers to the overall luminous intensity of the target chip.

[0048] The wavelength of the target chip refers to the wavelength of the light-emitting center of the target chip.

[0049] The uniformity of light emission of a target chip refers to the consistency of light emission on the surface of the target chip.

[0050] The peak wavelength of the target chip refers to the wavelength at which the intensity is the highest in the PL spectrum of the target chip after PL testing.

[0051] The half-width at half-maximum (WWHM) of a target chip refers to the spectral width in the PL spectrum of the target chip when the brightness is half of the maximum peak value.

[0052] The color coordinates of the target chip refer to the coordinate position of the emission color of the target chip in the chromaticity coordinate system.

[0053] Optionally, the optical parameters that need to be input into the target prediction model will also be different depending on the electrical parameters that the target chip needs to determine. That is to say, the optical parameters of the target chip are related to the electrical parameters that the target chip needs to determine.

[0054] Alternatively, the optical parameters of the target chip can be determined by an image acquisition device, such as a CCD camera, after performing PL testing on the target chip.

[0055] For example, the brightness of the target chip can be determined by capturing an image of the target chip using an image acquisition device to determine the brightness grayscale (0~255) of the target chip; the luminous uniformity of the target chip can be determined by acquiring multiple pixels of the luminous area of ​​the target chip (excluding the electrodes) from the chip image and overlaying them onto the target. The number of pixels can be greater than or equal to 9. The formula for calculating brightness uniformity can be expressed as follows:

[0056] (1)

[0057] Where U represents brightness uniformity, U MAX U represents the maximum brightness value among multiple pixels. MIN This represents the minimum brightness value among multiple pixels. Simply put, the closer U is to 1, the better the brightness uniformity of the target chip.

[0058] Step 404: Using the target prediction model, determine the electrical parameters of the target chip based on the optical parameter values; the electrical parameters include those obtained by testing the target chip through electroluminescence testing.

[0059] The target prediction model refers to a machine learning model that has learned the strong correlation between the optical parameters and electrical parameters of an LED chip. Based on this, the target prediction model can map the optical parameter values ​​of an LED chip to the electrical parameter values ​​of an LED chip based on the mapping relationship between the optical and electrical parameters of the LED chip.

[0060] The electrical parameters of the target chip refer to the parameters obtained through EL testing that can be used to characterize the electrical properties of the target chip.

[0061] EL testing refers to the testing of LED chips by injecting charge carriers through the application of a current driving signal, which causes electrons and holes to recombine directly and release photons. The electrical parameters of LED chips can be obtained based on EL testing.

[0062] The electrical parameter values ​​of a target chip refer to the specific numerical values ​​that can be obtained after performing EL testing on the target chip under specific EL testing conditions, and that can be used to quantify the electrical parameters of the target chip.

[0063] It should be noted that the electrical parameter values ​​of the target chip in this embodiment are predicted values, rather than actual test values ​​obtained by performing EL testing on the target chip. That is, EL testing of the target chip is not required in this embodiment.

[0064] Optionally, electrical parameters may include forward voltage, forward current utilization, reverse leakage current, forward voltage consistency, temperature drift coefficient, series resistance, turn-on voltage, etc., and may further include operating current, operating voltage, etc.

[0065] The forward voltage of the target chip refers to the voltage generated across the target chip when it is in a conducting state under a set forward current.

[0066] The forward current utilization rate of the target chip refers to the proportion of the forward current participating in the radiative recombination and photon release in the target chip to the total input forward current.

[0067] The reverse leakage current of the target chip refers to the leakage current flowing through the target chip when a specified reverse voltage is applied.

[0068] The forward voltage uniformity of the target chip refers to the dispersion and uniformity of the forward voltage of the target chip under EL test conditions.

[0069] The temperature drift coefficient of a target chip refers to the rate of change of the forward voltage of the target chip with the operating temperature.

[0070] The series resistance of the target chip refers to the equivalent internal resistance of the target chip when it is in the on state.

[0071] The turn-on voltage of a target chip refers to the critical forward voltage at which the target chip begins to enter the conduction state and emit light.

[0072] The operating current of the target chip refers to the forward current flowing through the target chip when it is in the on state under rated operating conditions.

[0073] The operating voltage of the target chip refers to the forward voltage across the target chip when it is in the on-state under rated operating conditions.

[0074] Optionally, the optical parameters that need to be input into the target prediction model will also be different depending on the electrical parameters that the target chip needs to determine. That is to say, the optical parameters of the target chip are related to the electrical parameters that the target chip needs to determine.

[0075] Optionally, the wavelength of an LED chip can reflect its bandgap. Based on this, the forward voltage of the LED chip can be deduced from its wavelength. The shorter the wavelength, the higher the forward voltage; that is, there is a negative correlation between the wavelength and the forward voltage. For example, when the wavelength of the LED chip is 450nm (blue light wavelength), the forward voltage is 3V~3.4V.

[0076] Optionally, the brightness of an LED chip can reflect its composite efficiency and light extraction efficiency. Based on this, the forward current utilization rate and reverse leakage current of the LED chip can be deduced from its brightness. For example, when the brightness of the LED chip is low and there are no spectral anomalies, the forward current utilization rate of the LED chip is low and the reverse leakage current is high.

[0077] Optionally, the peak wavelength of an LED chip can reflect the uniformity and thermal stability of the materials used in its fabrication. Based on this, the forward voltage uniformity and temperature drift coefficient of the LED chip can be deduced from its peak wavelength. For example, if the peak wavelength of an LED chip exhibits a large degree of dispersion, then the forward voltage of the LED chip will also exhibit a large degree of dispersion, and the temperature drift coefficient will be large, indicating lower thermal stability.

[0078] Optionally, the half-width at half-maximum (WWHM) of an LED chip can reflect its spectral purity, material uniformity, and defect count. Based on this, the series resistance and reverse leakage current of the LED chip can be deduced from its WWHM. For example, a wider WWHM corresponds to a larger series resistance and a larger reverse leakage current; that is, there is a negative correlation between the WWHM and the series resistance, and a similar negative correlation between the WWHM and the reverse leakage current.

[0079] Optionally, the relationship between the number of optical parameters and the number of electrical parameters of the target chip is not limited; that is, the number of optical parameters can be greater than, equal to, or less than the number of electrical parameters. Preferably, the number of optical parameters is greater than the number of electrical parameters.

[0080] In the aforementioned electrical parameter prediction method, only the optical parameters of the target chip need to be determined. Based on these optical parameters, the electrical parameters of the target chip can be determined using a target prediction model. Therefore, EL testing of the target chip is no longer required, allowing the acquisition of electrical parameters that would otherwise be necessary. This not only improves the safety of the LED chip by avoiding probe damage on the probe station and preventing damage or contamination of the LED chip's electrode surface, but also increases testing efficiency because PL testing is orders of magnitude faster than EL testing. Consequently, even at the wafer stage where the LED chip's electrodes are not yet fully fabricated, preliminary performance screening and yield prediction can be performed on semi-finished LED chips, significantly reducing LED chip production costs.

[0081] In one exemplary embodiment, the electrical parameters of the target chip include the operating current and series resistance of the target chip;

[0082] The above-mentioned method, using a target prediction model, determines the electrical parameters of the target chip based on the optical parameter values, including:

[0083] The optical parameters are input into the target prediction model to obtain the ideality factor, the saturation current of the target chip, and the series resistance of the target chip.

[0084] In response to the selection operation for the operating voltage, a first operating voltage is determined;

[0085] Based on the first operating voltage, saturation current, series resistance, and ideality factor, the operating current of the target chip at the first operating voltage is determined using the diode current equation.

[0086] The ideal factor is a parameter used to characterize how closely the equivalent diode characteristics of the target chip approximate the characteristics of an ideal diode.

[0087] The saturation current of the target chip refers to the reverse saturation current formed by the diffusion of minority carriers when the equivalent PN junction of the target chip is in a reverse bias state.

[0088] Optionally, the optical parameters used to provide the target prediction model with the prediction ideal factor, the saturation current of the target chip, and the series resistance of the target chip can be the wavelength, brightness, and brightness uniformity of the target chip.

[0089] Optionally, the operation for selecting the operating voltage can be initiated by the user on the terminal for a first operating voltage that they wish to select.

[0090] It is easy to understand that by substituting the first operating voltage, saturation current, series resistance, and ideality factor into the diode current equation, the operating current of the target chip under the first operating voltage can be calculated. In other words, the diode current equation is a physical model.

[0091] Optionally, after determining the operating current of the target chip at the first operating voltage using the diode current equation, the IV curve of the target chip can be further output based on the operating current of the target chip at each of the multiple first operating voltages. In the IV curve of the target chip, each data point represents the operating current of the target chip at the corresponding first operating voltage.

[0092] In an exemplary embodiment, the formula for calculating the diode current equation is expressed as follows:

[0093] (2)

[0094] Where I represents the operating current, V0 represents the first operating voltage, I0 represents the saturation current, and R... s Let n represent the series resistance, n represent the ideality factor, q represent the charge, k represent the Boltzmann constant, and T represent the temperature.

[0095] In the diode current equation, the operating current is the value to be calculated, the first operating voltage is the voltage value actively set by the user (for example, the first operating voltage can be 0, 0.5V, 1V, etc.), the saturation current, series resistance and ideality factor are the predicted values ​​output by the target prediction model, and the charge, Boltzmann constant and temperature are physical constants.

[0096] For example, assume that the target prediction model predicts a saturation current I0 = 1.2 * 10⁻¹² A and a series resistance R. sGiven an Ω resistance of 5Ω and an ideality factor n = 1.2, by substituting the first operating voltages V0 = 0, V0 = 0.6V, and V0 = 1.2V into the diode current equation, the corresponding operating current I can be obtained for each of these first operating voltages. Furthermore, by analyzing the different operating currents of the target chip under different first operating voltages, the terminal can output the IV curve of the target chip. Based on this, this embodiment can achieve high-precision IV curve prediction of the LED chip in a non-contact and high-speed manner using a target prediction model.

[0097] For example, such as Figure 5 As shown, the optical parameters of the target chip include its wavelength, brightness, and brightness uniformity. The wavelength, brightness, and brightness uniformity of the target chip are input into the target prediction model, which outputs the ideality factor, the saturation current of the target chip, and the series resistance of the target chip. Then, the ideality factor, the saturation current of the target chip, and the series resistance of the target chip are substituted into the calculation formula of the diode current equation. Furthermore, by selecting different first operating voltages and substituting them into the calculation formula of the diode current equation, the IV curve of the target chip can be obtained.

[0098] In this embodiment, the target prediction model can predict the ideality factor, the saturation current of the target chip, and the series resistance of the target chip using only the parameter values ​​of the optical parameters of the target chip. Furthermore, it can obtain the operating current of the target chip at the first operating voltage using the diode current equation and the selected first operating voltage. Thus, by performing PL testing on the target chip, the operating current of the target chip at the first operating voltage, which originally required EL testing, can be obtained. Based on this, both the safety and testing efficiency of the target chip can be improved.

[0099] In one exemplary embodiment, such as Figure 6 As shown, the optical parameters of the target chip include the wavelength, brightness, and brightness uniformity of the target chip; the electrical parameters of the target chip include the operating voltage and operating current of the target chip.

[0100] The above-mentioned method, using a target prediction model, determines the electrical parameters of the target chip based on the optical parameter values, including:

[0101] The wavelength, brightness, and brightness uniformity of the target chip are input into the target prediction model to obtain the operating voltage and operating current of the target chip.

[0102] Among them, for the wavelength, brightness and brightness uniformity of the target chip that are simultaneously input into the target prediction model, there is a corresponding relationship between the working voltage and working current output by the target prediction model. That is to say, if the target chip has the optical characteristics of corresponding wavelength, corresponding brightness and corresponding brightness uniformity, the target chip will be in normal working state based on the working voltage, and in this case, the target chip has the corresponding working current.

[0103] In this embodiment, the target prediction model can predict the operating voltage and current of the target chip simply by using the wavelength, brightness, and brightness uniformity of the target chip. Thus, by performing PL testing on the target chip, the operating current of the target chip under different operating voltages, which originally required EL testing, can be obtained. Based on this, both the safety and testing efficiency of the target chip can be improved.

[0104] In one exemplary embodiment, such as Figure 7 As shown, the above method further includes steps 702 to 708:

[0105] Step 702: Obtain training data including at least two parameter sets. Each parameter set includes the parameter values ​​of optical parameters and the measured parameter values ​​of electrical parameters of multiple chips. The parameter values ​​of optical parameters of multiple chips are obtained by photoluminescence testing, and the measured parameter values ​​of electrical parameters of multiple chips are obtained by electroluminescence testing.

[0106] Step 704: Input the optical parameter values ​​of multiple chips in the first parameter set into the initial prediction model to obtain the predicted parameter values ​​of the electrical parameters of multiple chips in the first parameter set, wherein at least two parameter sets include the first parameter set;

[0107] Step 706: Based on the measured parameter values ​​and predicted parameter values ​​of the electrical parameters of multiple chips in the first parameter set, determine the loss value corresponding to the first parameter set;

[0108] Step 708: Optimize the parameters of the initial prediction model based on the loss value corresponding to the first parameter set until the training termination condition is met, and obtain the target prediction model.

[0109] Among them, "multiple chips" refers to multiple training LED chips used to train the initial prediction model.

[0110] The measured values ​​of electrical parameters are actual values ​​obtained after performing actual EL tests on multiple chips.

[0111] The photoluminescence test conditions for obtaining the optical parameter values ​​of multiple chips can be: temperature of 23±1℃, humidity of (55±5)%RH, and PL test of multiple chips in a black cavity. The black cavity is to reduce the interference of stray light and other factors on the PL test.

[0112] The test conditions for electroluminescence testing to obtain the electrical parameters of multiple chips should be the same as the test conditions for photoluminescence testing to obtain the optical parameters of multiple chips.

[0113] Both PL and EL tests used to acquire training data require ensuring that the relevant test equipment has sufficient stability. For example, the excitation light power in the test conditions needs to have sufficient stability to reduce interference from other factors.

[0114] During PL testing of multiple chips, short pulse energy can be used to excite the region between the PN junctions of each chip to ensure sufficient excitation of each chip. Furthermore, the self-emission spectra generated by each chip can be collected by a spectrometer, and the optical parameters such as brightness, wavelength, and color coordinates of each chip can be obtained through calculation and analysis. In addition, during PL testing of multiple chips, it is necessary to ensure that the self-emission spectra of multiple chips can cover different wavelength bands to ensure that the parameter set has higher data comprehensiveness, thereby improving the prediction accuracy of the trained target prediction model.

[0115] During EL testing of multiple chips, a forward bias voltage can be applied to the region between the PN junctions of the chips, with the P-region of each chip connected to the positive terminal and the N-region connected to the negative terminal, driving each chip to emit light. The emitted light images of each chip can then be captured by an image acquisition device, and the light intensity or other optical parameters of each chip can be collected. During EL testing of multiple chips, relevant data can be simultaneously acquired with the operating voltage as the x-axis and the operating current as the y-axis to output the IV curve, turn-on voltage, reverse leakage current, and other electrical parameters of each chip.

[0116] Optionally, in each parameter set, the ratio of the number of samples between the optical parameter values ​​and the measured electrical parameter values ​​of multiple chips can be 8:2 to ensure that the trained target prediction model has sufficient generalization ability.

[0117] Optionally, the training termination condition can be that the loss value corresponding to the first parameter set is less than or equal to the loss value threshold.

[0118] In this embodiment, by using the optical parameter values ​​and electrical parameter measurement values ​​of multiple chips as training data for the initial prediction model, a target prediction model can be trained that can predict the electrical parameter values ​​of the target chip simply by inputting the optical parameter values ​​of the target chip. After the target prediction model is trained, there is no need to perform EL testing on batches of LED chips, which can improve both the safety and testing efficiency of batches of LED chips.

[0119] In an exemplary embodiment, obtaining training data comprising at least two parameter sets includes:

[0120] Acquire first initial training data including at least two first initial parameter sets; each first initial parameter set includes initial parameter values ​​of optical parameters and initial measurement parameter values ​​of electrical parameters of multiple chips;

[0121] Based on the parameter value thresholds of optical parameters and the measurement parameter thresholds of electrical parameters, abnormal data in the first initial training data are removed to obtain training data including at least two parameter sets.

[0122] This can be achieved by using box plots or image recognition methods to identify abnormal data in the first initial training data that exceed the threshold values ​​of optical parameters and the measurement threshold values ​​of electrical parameters, respectively, in order to remove abnormal data from the training data.

[0123] For example, the source chip of the abnormal data in the first initial training data may be a chip with foreign objects on its surface, a chip with poor appearance due to broken shape, tilted shape, or abnormal shape, a chip with excessively dim or bright light emission, a chip with an operating voltage outside the range of 2~5V, or a chip with an operating current outside the normal operating range.

[0124] In another exemplary embodiment, the optical parameters of the multiple chips include the wavelengths and brightness of the multiple chips. The aforementioned threshold values ​​for optical parameters and measurement threshold values ​​for electrical parameters are used to remove outlier data from the first initial training data, resulting in training data comprising at least two parameter sets, including:

[0125] Based on the parameter value thresholds of optical parameters and the measurement parameter thresholds of electrical parameters, abnormal data in the first initial training data are removed to obtain second initial training data including at least two second initial parameter sets.

[0126] Based on the wavelength and brightness of multiple chips, multiple wavelength ranges and multiple brightness ranges of multiple chips are obtained.

[0127] Determine the number of first samples in each wavelength range and the number of second samples in each brightness range for multiple chips;

[0128] If the variance among multiple first sample quantities is greater than or equal to a first variance threshold, sampling is performed on the corresponding wavelength ranges that do not reach the average of the multiple first sample quantities based on the average of the multiple first sample quantities, so that the variance among multiple first sample quantities is less than the first variance threshold; and / or, if the variance among multiple second sample quantities is greater than or equal to a second variance threshold, sampling is performed on the corresponding brightness ranges that do not reach the average of the multiple second sample quantities based on the average of the multiple second sample quantities, so that the variance among multiple second sample quantities is less than the second variance threshold;

[0129] The training data consists of at least two parameter sets in which the variance between multiple first sample sizes is less than a first variance threshold and / or the variance between multiple second sample sizes is less than a second variance threshold.

[0130] The first variance threshold can be the same as the second variance threshold.

[0131] The purpose of ensuring that the variance among multiple first sample sizes is less than a first variance threshold, and the purpose of ensuring that the variance among multiple second sample sizes is less than a second variance threshold, is to avoid bias in the training data towards the majority class. In this way, the trained target prediction model can be prevented from being biased towards the majority class during the prediction process. That is to say, the prediction accuracy of the target prediction model can be improved by increasing the comprehensiveness and objectivity of the training data.

[0132] In one exemplary embodiment, the optical parameters of the multiple chips include the wavelengths of the multiple chips; such as Figure 8 As shown, Figure 7 The method shown also includes steps 802 to 804:

[0133] Step 802: Group the multiple chips based on the wavelengths of the multiple chips in the first parameter set to obtain at least two chip groups in the first parameter set;

[0134] Step 804: Determine the weight of each chipset based on the number of chips included in at least two chipsets; the weight of a chipset is negatively correlated with the number of chips it includes.

[0135] The above-mentioned loss value corresponding to the first parameter set is determined based on the measured parameter values ​​and predicted parameter values ​​of the electrical parameters of multiple chips in the first parameter set. That is, step 706 above includes the following sub-steps 7062 to 7064:

[0136] Sub-step 7062: Based on the measured parameter values ​​and predicted parameter values ​​of the electrical parameters of each chip in the first parameter set, determine the loss value corresponding to each chip in the first parameter set;

[0137] Sub-step 7064: Based on the weights of the chipsets corresponding to each chip in the first parameter set, perform a weighted summation of the loss values ​​corresponding to each chip in the first parameter set to obtain the loss value corresponding to the first parameter set.

[0138] The process of grouping multiple chips based on the wavelengths of multiple chips in the first parameter set can be done by grouping multiple chips into intervals with a wavelength difference of 10nm, or optionally by grouping multiple chips into intervals with wavelength differences of 5nm and 15nm. This application does not limit the specific grouping.

[0139] To avoid the undesirable situation where the initial prediction model is biased towards the majority class data due to insufficient sample numbers in certain wavelength ranges in the training data, it is necessary to increase the weight of minority class samples during the calculation of the loss value of each chip to ensure the prediction accuracy of the target prediction model obtained in the final training. Therefore, the weight of the chip group is negatively correlated with the number of chips included.

[0140] The weights of each chipset should be such that each chipset contributes equally to the calculation of the loss value. For example, suppose there are 50 chips with a wavelength range of 450-460nm (blue light) and 500 chips with a wavelength range of 550-560nm (green light). Without considering chips in other wavelength ranges, since the ratio of the former to the latter is 1:10, the weight of the chips with a wavelength range of 450-460nm (blue light) can be 10 and the weight of the chips with a wavelength range of 550-560nm (green light) can be 1 in the calculation of the loss value. Based on this, the influence of the difference in the number of chips can be eliminated, and both chips can contribute equally to the calculation of the loss value.

[0141] Optionally, the weight of each chipset is determined based on the number of chips included in at least two chipsets. This can be done by determining the average number of chips based on the number of chips included in at least two chipsets. In this case, the weight of chipsets with a number of chips greater than or equal to the average number of chips is determined to be less than or equal to 1, while the weight of chipsets with a number of chips less than the average number of chips is determined to be greater than 1.

[0142] In an exemplary embodiment, the formula for calculating the loss value corresponding to the first parameter set is expressed as:

[0143] (3)

[0144] Where Loss represents the loss value corresponding to the first parameter set, n represents the number of chips, and w i y represents the weight of the chipset corresponding to the i-th chip. i y' represents the measured value of the electrical parameters of the i-th chip. i This represents the predicted parameter value for the i-th chip.

[0145] Optionally, after calculating the loss value corresponding to the first parameter set, the prediction error of the initial prediction model can be evaluated using the root mean square error (RMSE) calculation formula and the normalized root mean square error (NRMSE) calculation formula. That is, the training termination condition can be that the prediction error obtained by the NRMSE calculation formula is less than or equal to the error threshold.

[0146] In this embodiment, during the process of weighted summation of the loss values ​​corresponding to each chip in the first parameter set to obtain the loss value corresponding to the first parameter set, each chip in the first parameter set corresponds to a weight of its respective chip group, and the weight of each chip group is negatively correlated with the number of chips it includes. Thus, during the training process of the initial prediction model, it is ensured that chips of different wavelengths can contribute to the training process of the initial prediction model to the same degree, avoiding the undesirable situation that the initial prediction model tends to favor the majority class samples due to different numbers of chips of different wavelengths. Consequently, the prediction accuracy of the target prediction model finally trained is significantly improved.

[0147] In one exemplary embodiment, such as Figure 9 As shown, the initial prediction module 90 includes a first hidden layer 902, a second hidden layer 904, and a third hidden layer 906; the above-mentioned inputting the optical parameter values ​​of multiple chips in the first parameter set into the initial prediction module 90 to obtain the predicted parameter values ​​of the electrical parameters of multiple chips in the first parameter set includes:

[0148] The parameter values ​​of the optical parameters of multiple chips in the first parameter set are input into the first hidden layer 902 to obtain the derived features of multiple chips in the first parameter set.

[0149] The derived features of multiple chips in the first parameter set are input into the second hidden layer 904 to obtain the correlation between the electrical parameters and optical parameters of multiple chips in the first parameter set.

[0150] The optical parameter values ​​of multiple chips in the first parameter set and the correlation between the electrical parameters and the optical parameters are input into the third hidden layer 906 to obtain the predicted parameter values ​​of the electrical parameters of multiple chips in the first parameter set.

[0151] Here, the derived features refer to the features derived from the optical parameters of multiple chips in the first parameter set after mathematical operations. Based on the derived features, the second hidden layer 904 can uncover the correlation between the electrical and optical parameters of multiple chips.

[0152] Optionally, derived features may include features such as the ratio between brightness and wavelength, the square of wavelength, and the product between wavelength and brightness uniformity.

[0153] Optionally, the initial prediction module 90 can be a neural network, based on which the number of neurons in the second hidden layer 904 can be greater than the number of neurons in the first hidden layer 902 and the number of neurons in the third hidden layer 906, respectively.

[0154] Optionally, each hidden layer may contain 64 to 256 neurons. The first hidden layer 902 may contain 64 neurons, the second hidden layer 904 may contain 128 neurons, and the third hidden layer 906 may contain 64 neurons. It is readily understood that the greater the number of optical parameters across the multiple chips in the first parameter set, the greater the number of neurons in each hidden layer; that is, the number of optical parameters across the multiple chips in the first parameter set is positively correlated with the number of neurons in each hidden layer.

[0155] Optionally, the ratio between the number of neurons in the second hidden layer 904 and the number of neurons in the first hidden layer 902 can be 2.

[0156] For example, the relationship between electrical parameters and optical parameters can be the influence of brightness on operating current, or the influence of wavelength on the slope of the IV curve, etc.

[0157] Optionally, the initial prediction module 90 may further include a batch normalization layer (or BatchNorm layer) and a random deactivation layer (or Dropout layer). The batch normalization layer is used to standardize the brightness parameter values ​​of multiple chips (mean=0, variance=1, normal distribution) when the brightness parameter values ​​of multiple chips input to each layer fluctuate greatly, so that the brightness parameter values ​​of multiple chips input to the initial prediction module 90 have high stability in distribution. The random deactivation layer is used to give each hidden layer a certain probability of inactivation during the training process of the initial prediction module 90, so as to avoid the initial prediction module 90 from being specifically trained for parameter values ​​with special optical parameters (for example, special chips with abnormalities in a few wavelength ranges), thereby improving the generalization ability of the final trained target prediction model by avoiding overfitting.

[0158] Optionally, the inactivation probability of the random inactivation layer can be 0.2~0.3 or other probability ranges, or it can be 0.2, 0.25, 0.3 or other probability values. Further optionally, the inactivation probability of the random inactivation layer is positively correlated with the number of optical parameters of the multiple chips in the first parameter set.

[0159] In the initial prediction module 90, the learning process of the optical parameters of multiple chips in the first parameter set in the first hidden layer 902, the second hidden layer 904 and the third hidden layer 906 is a nonlinear mapping learning process.

[0160] Optionally, in the process of inputting the derived features of multiple chips in the first parameter set into the second hidden layer 904 to obtain the correlation between the electrical and optical parameters of the multiple chips in the first parameter set, the second hidden layer 904 can employ a random forest model or a recursive feature elimination (RFE) algorithm to select at least one feature from the derived features that contributes the most to the electrical parameter to be predicted. For example, assuming the first hidden layer 902 inputs four derived features into the second hidden layer 904, the random forest model in the second hidden layer 904 predicts the turn-on voltage of the chip using these four derived features, further outputs a score for each of the four derived features, and then uses, for example, the top three derived features with the highest scores as the associated features; while the RFE algorithm trains the initial prediction module 90 using all derived features, removes the worst derived features after multiple rounds of training, and retains the optimal combination of derived features.

[0161] Optionally, an activation function can be provided at the output of each hidden layer. The activation function can be a rectified linear unit (ReLU) function.

[0162] For example, such as Figure 9As shown, assuming that the number of optical parameters (X1, X2, X3) of multiple chips in the first parameter set is three, and the number of electrical parameters (Y1, Y2) of multiple chips in the first parameter set is two, the initial prediction module 90 includes a first hidden layer 902, a second hidden layer 904, and a third hidden layer 906. The first hidden layer 902 includes 64 neurons (L1, L2, ..., L64), the second hidden layer 904 includes 128 neurons (M1, M2, ..., M128), and the third hidden layer 906 includes 64 neurons (N1, N2, ..., N64). After the optical parameters of multiple chips in the first parameter set are input into the first hidden layer 902, under the joint learning of the first hidden layer 902, the second hidden layer 904, and the third hidden layer 906, the optical parameters of multiple chips in the first parameter set are mapped to the electrical parameters of multiple chips in the first parameter set. Optionally, in this exemplary embodiment, optical parameters X1, X2, and X3 can correspond to brightness, wavelength, and brightness uniformity, electrical parameters Y1 and Y2 can correspond to operating voltage and operating current, the derived features output by the first hidden layer 902 can be the product of wavelength and brightness uniformity, and the correlation relationships output by the second hidden layer 904 can be the correlation relationships between brightness and operating current, and the correlation relationships between wavelength and the slope of the IV curve.

[0163] In this embodiment, a neural network machine learning algorithm is used to train the initial prediction model. The neural network adopts a three-layer feedforward structure, i.e., three hidden layers. Based on this, by iteratively learning the training data, the mapping relationship between the optical parameters and electrical parameters of the LED chip can be established, and finally a target prediction model with high prediction accuracy can be obtained.

[0164] In an exemplary embodiment, when the number of optical parameters of multiple chips in the first parameter set is 3 to 5, the number of hidden layers is 3, and the number of neurons included in the 3 hidden layers along the data input direction to the data output direction is 64, 128 and 64 respectively; the inactivation probability of the random inactivation layer is 0.2 or 0.2 to 0.3.

[0165] In this embodiment, since the number of optical parameters of multiple chips in the first parameter set is small and the information that optical parameters can provide is relatively limited, the limited number of hidden layers can avoid the bad phenomenon of overfitting during training and improve the prediction accuracy of the target prediction model obtained in the final training.

[0166] In an exemplary embodiment, when the number of optical parameters of multiple chips in the first parameter set is 6 to 8, the number of hidden layers is 4, and the number of neurons included in the 4 hidden layers along the data input direction to the data output direction are 128, 256, 128 and 64 respectively; the inactivation probability of the random inactivation layer is 0.25 or 0.25 to 0.35.

[0167] In the case of four hidden layers, the fourth hidden layer can be positioned between the first and second hidden layers. The fourth hidden layer, based on the first hidden layer, transforms the derived features output by the first hidden layer into high-dimensional complex features that precisely correspond to electrical parameters such as operating voltage and operating current. This high-dimensional complex feature is then input to the second hidden layer. For example, if the chip has a wavelength of 520nm and a brightness of 160, the fourth hidden layer can map this to obtain an operating voltage of 3.3V for the chip.

[0168] In this embodiment, since the number of optical parameters of multiple chips in the first parameter set is large, by increasing the number of neurons included in each hidden layer, the ability of the target prediction model to extract information from optical parameters can be improved, thereby improving the prediction accuracy of the target prediction model.

[0169] In an exemplary embodiment, when the number of optical parameters of multiple chips in the first parameter set is greater than or equal to 10, the number of hidden layers is 5, and the number of neurons included in the 5 hidden layers along the data input direction to the data output direction are 256, 512, 256, 128 and 64, respectively; the inactivation probability of the randomly deactivated layer is 0.3 or 0.3~0.4.

[0170] In this embodiment, since there are many optical parameters of multiple chips in the first parameter set, it is necessary to use a neural network with a large number of hidden layers to fit the mapping relationship between the numerous optical parameters and electrical parameters, so as to improve the information extraction capability of the target prediction model obtained by the final training, thereby improving the prediction accuracy of the target prediction model obtained by the final training.

[0171] In one exemplary embodiment, such as Figure 10 As shown, the initial prediction module 90 also includes a preprocessing layer 1002; the parameter values ​​of the optical parameters of multiple chips in the first parameter set are input into the first hidden layer 902 to obtain the derived features of the multiple chips, including:

[0172] Through the preprocessing layer 1002, the parameter values ​​of the optical parameters of multiple chips in the first parameter set are normalized and data augmented in sequence to obtain the processed parameter values ​​of the optical parameters of multiple chips in the first parameter set.

[0173] The processing parameter values ​​of the optical parameters of multiple chips in the first parameter set are input into the first hidden layer 902 to obtain the derived features of the optical parameters of multiple chips in the first parameter set.

[0174] The normalization process for the optical parameters of multiple chips in the first parameter set refers to eliminating the unit range of optical parameters such as brightness and wavelength of multiple chips in the first parameter set. On the one hand, this makes the optical parameters with different unit ranges comparable. On the other hand, it avoids the problem of excessively large or small optical parameter values ​​leading to excessive data processing or data fluctuations in the initial prediction module 90, thereby improving the reliability of the target prediction model obtained in the final training.

[0175] For example, normalization can be achieved by dividing the optical parameter values ​​of multiple chips in the first parameter set by the average value of the optical parameters; or normalization can be achieved by limiting the fluctuation range of the optical parameter values ​​of multiple chips in the first parameter set. The normalized optical parameter value = optical parameter value / (maximum fluctuation range of optical parameter - minimum fluctuation range of optical parameter), so that the normalized optical parameter value range is [0, 1]. For example, if the brightness fluctuation range is set to 50~200, then for a training chip with a brightness of 100, its normalized brightness = brightness / (maximum fluctuation range of brightness - minimum fluctuation range of brightness) = 100 / (200 - 50) ≈ 0.67.

[0176] Data augmentation of the optical parameters of multiple chips in the first parameter set refers to adding Gaussian noise with an amplitude of 5% to the optical parameters such as brightness and wavelength of multiple chips in the first parameter set to simulate the measurement error that the initial prediction module 90 may have in the actual test process. By expanding the sample size of the training data, the generalization ability of the target prediction model obtained in the final training is improved.

[0177] For example, such as Figure 10 As shown, the parameter values ​​of the optical parameters (X1, X2, X3) of multiple chips in the first parameter set are input to the preprocessing layer 1002. The preprocessing layer 1002 performs normalization and data augmentation on the parameter values ​​of the optical parameters of multiple chips in the first parameter set in sequence. The optical parameters after normalization and data augmentation are X'1, X'2, and X'3. Based on this, the processed parameter values ​​of the optical parameters of multiple chips in the first parameter set are obtained. The processed parameter values ​​of the optical parameters of multiple chips in the first parameter set are input to the first hidden layer 902 to obtain the derived features of the optical parameters of multiple chips in the first parameter set.

[0178] In this embodiment, the initial prediction model also includes a preprocessing layer. Based on this, the parameter values ​​of the optical parameters of multiple chips in the first parameter set are normalized and data augmented sequentially through the preprocessing layer. On the one hand, this can improve the convergence speed and training stability of the initial prediction model during the training process. On the other hand, the introduction of Gaussian noise can improve the robustness of the initial prediction model to data fluctuations, avoid the overfitting defect of the target prediction model, and significantly improve the prediction accuracy and generalization ability of the target prediction model obtained by the final training.

[0179] In one exemplary embodiment, the method further includes:

[0180] The measured parameter values ​​of the electrical parameters of multiple chips in the first parameter set are normalized through the preprocessing layer to obtain the processed measured parameter values ​​of the electrical parameters of multiple chips in the first parameter set.

[0181] The above-mentioned determination of the loss value corresponding to the first parameter set based on the measured and predicted parameter values ​​of the electrical parameters of multiple chips in the first parameter set includes:

[0182] Based on the processed measured parameter values ​​and predicted parameter values ​​of the electrical parameters of multiple chips in the first parameter set, the loss value corresponding to the first parameter set is determined.

[0183] Similarly, electrical parameters such as operating voltage and operating current can also be normalized by dividing by the average value of the electrical parameters, or by limiting the fluctuation range of the measured parameter values ​​of the electrical parameters of multiple chips in the first parameter set.

[0184] In this embodiment, in addition to normalizing the optical parameter values ​​of multiple chips in the first parameter set, the electrical parameter measurement values ​​of multiple chips in the first parameter set are also normalized. This eliminates the unit range between the processed measurement parameter values ​​and the predicted parameter values ​​of multiple chips, making them comparable and improving the training efficiency and reliability of the initial prediction model.

[0185] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0186] Based on the same inventive concept, this application also provides an electrical parameter prediction device for implementing the electrical parameter prediction method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more embodiments of the electrical parameter prediction device provided below can be found in the limitations of the electrical parameter prediction method described above, and will not be repeated here.

[0187] In one exemplary embodiment, such as Figure 11 As shown, an electrical parameter prediction device is provided, comprising: a first determining module 1102 and a second determining module 1104, wherein:

[0188] The first determining module 1102 is used to determine the parameter values ​​of the optical parameters of the target chip;

[0189] The second determining module 1104 is used to determine the electrical parameters of the target chip based on the parameter values ​​of the optical parameters through the target prediction model; the electrical parameters include the electrical parameters obtained by testing the target chip through electroluminescence testing.

[0190] In an exemplary embodiment, the electrical parameters of the target chip include the operating current and series resistance of the target chip. Regarding determining the values ​​of the electrical parameters of the target chip based on the values ​​of the optical parameters using a target prediction model, the second determining module 1104 is specifically configured to input the values ​​of the optical parameters into the target prediction model to obtain the ideality factor, the saturation current of the target chip, and the series resistance of the target chip; in response to a selection operation for the operating voltage, determine a first operating voltage; and based on the first operating voltage, the saturation current, the series resistance, and the ideality factor, determine the operating current of the target chip at the first operating voltage using a diode current equation.

[0191] In an exemplary embodiment, the optical parameters of the target chip include the wavelength, brightness, and brightness uniformity of the target chip; the electrical parameters of the target chip include the operating voltage and operating current of the target chip; in determining the parameter values ​​of the electrical parameters of the target chip based on the parameter values ​​of the optical parameters through the target prediction model, the second determining module 1104 is specifically used to input the wavelength, brightness, and brightness uniformity of the target chip into the target prediction model to obtain the operating voltage and operating current of the target chip.

[0192] In an exemplary embodiment, the second determining module 1104 is further configured to acquire training data including at least two parameter sets, each parameter set including parameter values ​​of optical parameters and measured parameter values ​​of electrical parameters of multiple chips, wherein the parameter values ​​of optical parameters of multiple chips are obtained by photoluminescence testing, and the measured parameter values ​​of electrical parameters of multiple chips are obtained by electroluminescence testing; input the parameter values ​​of optical parameters of multiple chips in the first parameter set into the initial prediction model to obtain the predicted parameter values ​​of electrical parameters of multiple chips in the first parameter set, wherein the at least two parameter sets include the first parameter set; determine the loss value corresponding to the first parameter set based on the measured parameter values ​​and predicted parameter values ​​of electrical parameters of multiple chips in the first parameter set; optimize the parameters of the initial prediction model based on the loss value corresponding to the first parameter set until the training termination condition is met to obtain the target prediction model.

[0193] In an exemplary embodiment, the optical parameters of the multiple chips include the wavelengths of the multiple chips; the first determining module 1102 is further configured to group the multiple chips based on the wavelengths of the multiple chips in the first parameter set to obtain at least two chip groups in the first parameter set; determine the weight of each chip group based on the number of chips included in the at least two chip groups; the weight of the chip group is negatively correlated with the number of chips included; in terms of determining the loss value corresponding to the first parameter set based on the measured parameter value and predicted parameter value of the electrical parameters of the multiple chips in the first parameter set, the second determining module 1104 is specifically configured to determine the loss value corresponding to each chip in the first parameter set based on the measured parameter value and predicted parameter value of the electrical parameters of each chip in the first parameter set; and perform a weighted summation of the loss values ​​corresponding to each chip in the first parameter set based on the weight of the chip group corresponding to each chip in the first parameter set to obtain the loss value corresponding to the first parameter set.

[0194] In an exemplary embodiment, the initial prediction model includes a first hidden layer, a second hidden layer, and a third hidden layer. Regarding inputting the parameter values ​​of the optical parameters of multiple chips in the first parameter set into the initial prediction model to obtain the predicted parameter values ​​of the electrical parameters of the multiple chips in the first parameter set, the second determining module 1104 is specifically used to input the parameter values ​​of the optical parameters of the multiple chips in the first parameter set into the first hidden layer to obtain the derived features of the multiple chips in the first parameter set; input the derived features of the multiple chips in the first parameter set into the second hidden layer to obtain the correlation between the electrical parameters and optical parameters of the multiple chips in the first parameter set; and input the parameter values ​​of the optical parameters of the multiple chips in the first parameter set and the correlation between the electrical parameters and optical parameters into the third hidden layer to obtain the predicted parameter values ​​of the electrical parameters of the multiple chips in the first parameter set.

[0195] In an exemplary embodiment, the initial prediction model further includes a preprocessing layer; in terms of inputting the parameter values ​​of the optical parameters of multiple chips in the first parameter set into the first hidden layer to obtain the derived features of the multiple chips, the second determining module 1104 performs normalization processing and data augmentation on the parameter values ​​of the optical parameters of multiple chips in the first parameter set sequentially through the preprocessing layer to obtain the processed parameter values ​​of the optical parameters of multiple chips in the first parameter set; the processed parameter values ​​of the optical parameters of multiple chips in the first parameter set are input into the first hidden layer to obtain the derived features of the optical parameters of multiple chips in the first parameter set.

[0196] Each module in the aforementioned electrical parameter prediction device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0197] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 12As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores data related to the electrical parameter prediction method. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When the computer program is executed by the processor, it implements an electrical parameter prediction method.

[0198] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 13 As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When executed by the processor, the computer program implements an electrical parameter prediction method. The display unit is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0199] Those skilled in the art will understand that Figure 12 or Figure 13The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0200] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0201] In one exemplary embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above-described method embodiments.

[0202] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0203] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0204] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0205] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0206] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for predicting electrical parameters, characterized in that, include: Determine the parameter values ​​of the optical parameters of the target chip; Based on the optical parameters, the electrical parameters of the target chip are determined using a target prediction model; the electrical parameters include those obtained by testing the target chip using electroluminescence.

2. The method according to claim 1, characterized in that, The electrical parameters of the target chip include its operating current and series resistance. The step of determining the electrical parameters of the target chip based on the optical parameters using a target prediction model includes: The parameter values ​​of the optical parameters are input into the target prediction model to obtain the ideality factor, the saturation current of the target chip, and the series resistance of the target chip; In response to the selection operation for the operating voltage, a first operating voltage is determined; Based on the first operating voltage, the saturation current, the series resistance, and the ideality factor, the operating current of the target chip at the first operating voltage is determined using the diode current equation.

3. The method according to claim 1, characterized in that, The optical parameters of the target chip include the wavelength, brightness, and brightness uniformity of the target chip; the electrical parameters of the target chip include the operating voltage and operating current of the target chip. The step of determining the electrical parameters of the target chip based on the optical parameters using a target prediction model includes: The wavelength, brightness, and brightness uniformity of the target chip are input into the target prediction model to obtain the operating voltage and operating current of the target chip.

4. The method according to claim 1, characterized in that, The method further includes: Acquire training data comprising at least two parameter sets, each parameter set comprising parameter values ​​of optical parameters and measured parameter values ​​of electrical parameters of multiple chips, wherein the parameter values ​​of optical parameters of multiple chips are obtained by photoluminescence testing, and the measured parameter values ​​of electrical parameters of multiple chips are obtained by electroluminescence testing; The parameter values ​​of the optical parameters of multiple chips in the first parameter set are input into the initial prediction model to obtain the predicted parameter values ​​of the electrical parameters of multiple chips in the first parameter set, wherein the at least two parameter sets include the first parameter set; Based on the measured and predicted parameter values ​​of the electrical parameters of multiple chips in the first parameter set, the loss value corresponding to the first parameter set is determined; The parameters of the initial prediction model are optimized based on the loss values ​​corresponding to the first parameter set until the training termination condition is met, thereby obtaining the target prediction model.

5. The method according to claim 4, characterized in that, The optical parameters of the plurality of chips include the wavelengths of the plurality of chips; the method further includes: Based on the wavelengths of multiple chips in the first parameter set, the multiple chips are grouped to obtain at least two chip groups in the first parameter set; The weight of each chipset is determined based on the number of chips included in at least two of the chipsets; the weight of the chipset is negatively correlated with the number of chips it includes. The step of determining the loss value corresponding to the first parameter set based on the measured parameter values ​​and predicted parameter values ​​of the electrical parameters of multiple chips in the first parameter set includes: Based on the measured and predicted electrical parameter values ​​of each chip in the first parameter set, the loss value corresponding to each chip in the first parameter set is determined. Based on the weights of the chipsets corresponding to each chip in the first parameter set, the loss values ​​corresponding to each chip in the first parameter set are weighted and summed to obtain the loss value corresponding to the first parameter set.

6. The method according to claim 4, characterized in that, The initial prediction model includes a first hidden layer, a second hidden layer, and a third hidden layer; the step of inputting the parameter values ​​of the optical parameters of the multiple chips in the first parameter set into the initial prediction model to obtain the predicted parameter values ​​of the electrical parameters of the multiple chips in the first parameter set includes: The parameter values ​​of the optical parameters of the multiple chips in the first parameter set are input into the first hidden layer to obtain the derived features of the multiple chips in the first parameter set. The derived features of the multiple chips in the first parameter set are input into the second hidden layer to obtain the correlation between the electrical parameters and optical parameters of the multiple chips in the first parameter set; The optical parameter values ​​of multiple chips in the first parameter set and the correlation between the electrical parameters and the optical parameters are input into the third hidden layer to obtain the predicted parameter values ​​of the electrical parameters of multiple chips in the first parameter set.

7. The method according to claim 6, characterized in that, The initial prediction model further includes a preprocessing layer; the step of inputting the parameter values ​​of the optical parameters of the multiple chips in the first parameter set into the first hidden layer to obtain the derived features of the multiple chips includes: Through the preprocessing layer, the parameter values ​​of the optical parameters of the multiple chips in the first parameter set are sequentially normalized and data-enhanced to obtain the processed parameter values ​​of the optical parameters of the multiple chips in the first parameter set. The processing parameter values ​​of the optical parameters of the multiple chips in the first parameter set are input into the first hidden layer to obtain the derived features of the optical parameters of the multiple chips in the first parameter set.

8. An electrical parameter prediction device, characterized in that, The device includes: The first determining module is used to determine the parameter values ​​of the optical parameters of the target chip; The second determining module is used to determine the electrical parameters of the target chip based on the parameter values ​​of the optical parameters using a target prediction model; the electrical parameters include electrical parameters obtained by testing the target chip through electroluminescence testing.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images