Device model parameter extraction method, device model selection method and electronic equipment

Abstract

The invention provides a device model parameter extraction method, a device model selection method and electronic equipment. The method comprises the following steps: acquiring a parameter curve graph of a device; detecting the position of each scale on a coordinate axis of the parameter curve graph, and associating each scale with scale characters in the parameter curve graph based on the position of each scale to obtain a scale value of each coordinate pixel point on the coordinate axis; scanning a curve point corresponding to each coordinate pixel point on the coordinate axis in the parameter curve graph; and determining model parameters of the device based on the scale value of each coordinate pixel point and the curve point corresponding to each coordinate pixel point. According to the method and the electronic equipment provided by the invention, the extraction of the model parameters is realized through image processing, and the reliability and the accuracy of model parameter extraction are guaranteed while the model parameter extraction cost is reduced.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular, to a method for extracting and selecting device model parameters and an electronic device. Background Art

[0002] Device models can provide detailed parameters of electronic components, providing strong support for circuit design and circuit simulation.

[0003] In practice, to obtain device model parameters, technicians need to measure parameters through experiments, or extract model parameters through simulation software, or use standardized models for modeling. However, measuring parameters through experiments requires specific measuring equipment, a large amount of capital and human resources; extracting model parameters through simulation software requires the prior purchase of expensive electromagnetic field simulation software; using standardized models for modeling will result in large deviations in circuit characteristics and progress.

[0004] Therefore, how to obtain accurate and reliable device model parameters at low cost remains an urgent problem in this field. Summary of the Invention

[0005] The present invention provides a method for extracting and selecting device model parameters and an electronic device, aiming to solve the defects of high cost and poor reliability in extracting device model parameters in the prior art.

[0006] The present invention provides a method for extracting device model parameters, including:

[0007] Obtaining a parameter curve graph of the device;

[0008] Detecting the positions of each scale on the coordinate axis of the parameter curve graph, and based on the positions of each scale, associating each scale with the scale text in the parameter curve graph to obtain the scale values of each coordinate pixel point on the coordinate axis;

[0009] Scanning the curve points corresponding to each coordinate pixel point on the coordinate axis in the parameter curve graph;

[0010] Determining the model parameters of the device based on the scale values of each coordinate pixel point and the curve points corresponding to each coordinate pixel point.

[0011] According to the method for extracting device model parameters provided by the present invention, the scanning of the curve points corresponding to each coordinate pixel point on the coordinate axis in the parameter curve graph includes:

[0012] Scanning and obtaining, one by one, the set of pixel points corresponding to each coordinate pixel point on the first coordinate axis in the parameter curve graph, where the first coordinate axis is any coordinate axis of the parameter curve graph;

[0013] Based on the positional relationship between each coordinate pixel point and the scale on the first coordinate axis, determine the curve points corresponding to each coordinate pixel point from the set of pixel points corresponding to each coordinate pixel point.

[0014] According to a device model parameter extraction method provided by the present invention, the step of determining the curve points corresponding to each coordinate pixel point from the set of pixel points corresponding to each coordinate pixel point based on the positional relationship between each coordinate pixel point and the scale on the first coordinate axis includes:

[0015] In the case where the positional relationship is non - coincident, based on the pixel values of each pixel point in the set of pixel points, and / or the positional relationship between each pixel point in the set of pixel points and the scale on the second coordinate axis, determine the curve points corresponding to the coordinate pixel point from the set of pixel points, where the second coordinate axis is a coordinate axis perpendicular to the first coordinate axis.

[0016] According to a device model parameter extraction method provided by the present invention, the step of determining the curve points corresponding to the coordinate pixel point from the set of pixel points based on the pixel values of each pixel point in the set of pixel points, and / or the positional relationship between each pixel point in the set of pixel points and the scale on the second coordinate axis includes:

[0017] Based on the pixel values of each pixel point in the set of pixel points, determine a plurality of candidate line segments;

[0018] Based on the number of pixel points constituting each candidate line segment, determine the weight of each candidate line segment;

[0019] Based on the weights of the candidate line segments and the distances between the mid - points of each candidate line segment and the previous curve point respectively, determine the target line segment from the candidate line segments; the previous curve point is the curve point corresponding to the previous coordinate pixel point of the coordinate pixel point;

[0020] Based on the target line segment, determine the curve point corresponding to the coordinate pixel point.

[0021] According to a device model parameter extraction method provided by the present invention, the step of determining the curve points corresponding to each coordinate pixel point from the set of pixel points corresponding to each coordinate pixel point based on the positional relationship between each coordinate pixel point and the scale on the first coordinate axis includes:

[0022] In the case where the positional relationship is coincident, based on the previous curve point, determine the curve points corresponding to the coordinate pixel point from the set of pixel points corresponding to the coordinate pixel point; the previous curve point is the curve point corresponding to the previous coordinate pixel point of the coordinate pixel point.

[0023] A method for extracting device model parameters provided by the present invention, the step of associating each scale with the scale text in the parameter curve graph to obtain the scale values of each coordinate pixel point on the coordinate axis includes:

[0024] Based on the change of the distance between adjacent scales, divide the magnitude to which each scale belongs;

[0025] Based on the magnitude to which each scale belongs, associate each scale with the scale text to obtain the scale values of each coordinate pixel point on the coordinate axis.

[0026] A method for extracting device model parameters provided by the present invention, the step of determining the scale text includes:

[0027] Locate the text area in the parameter curve graph;

[0028] Detect the text contour within the text area;

[0029] Based on each text contour detected within the text area, perform text recognition on the parameter curve graph to obtain the scale text in the parameter curve graph.

[0030] The present invention also provides a device selection method, including:

[0031] Determine the target device parameters;

[0032] Match the target device parameters with the model parameters of each candidate device, and determine the candidate device with the highest matching degree as the target device;

[0033] The model parameters of the candidate device are determined based on the device model parameter extraction method described above.

[0034] The present invention also provides a device model parameter extraction device, including:

[0035] An acquisition unit that acquires the parameter curve graph of the device;

[0036] A scale detection unit for detecting the positions of each scale on the coordinate axis of the parameter curve graph, and based on the positions of each scale, associating each scale with the scale text in the parameter curve graph to obtain the scale values of each coordinate pixel point on the coordinate axis;

[0037] A curve scanning unit for scanning the curve points corresponding to each coordinate pixel point on the coordinate axis in the parameter curve graph;

[0038] A modeling unit for determining the model parameters of the device based on the scale values of each coordinate pixel point and the curve points corresponding to each coordinate pixel point.

[0039] The present invention also provides a device selection device, including:

[0040] A parameter determination unit for determining target device parameters;

[0041] A device matching unit for matching the target device parameters with the model parameters of each candidate device, and determining the candidate device with the highest matching degree as the target device;

[0042] The model parameters of the candidate devices are determined based on the device model parameter extraction method as described above.

[0043] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the device model parameter extraction method or the device selection method as described above is implemented.

[0044] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the device model parameter extraction method or the device selection method as described above is implemented.

[0045] The present invention also provides a computer program product, including a computer program. When the computer program is executed by a processor, the device model parameter extraction method or the device selection method as described above is implemented.

[0046] The device model parameter extraction, selection method, and electronic device provided by the present invention associate each scale on the coordinate axis of the parameter curve graph with the scale text to obtain the scale values of each coordinate pixel point on the coordinate axis. In addition, each curve point corresponding to the coordinate pixel point is obtained by scanning, so that the curve point and the scale value are associated, and the model parameters of the device corresponding to the parameter curve graph are obtained, thereby realizing the extraction of model parameters through image processing, while reducing the cost of model parameter extraction and ensuring the reliability and accuracy of model parameter extraction. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0048] Figure 1 is one of the flow diagrams of the device model parameter extraction method provided by the present invention;

[0049] Figure 2 is the flow diagram of the curve point scanning method provided by the present invention;

[0050] Figure 3 is one of the partial screenshots of the parameter curve graph provided by the present invention;

[0051] Figure 4 is the second partial screenshot of the parameter curve graph provided by the present invention;

[0052] Figure 5 is the third partial screenshot of the parameter curve graph provided by the present invention;

[0053] Figure 6 is the frequency impedance graph provided by the present invention;

[0054] Figure 7 is the fourth partial screenshot of the parameter curve graph provided by the present invention;

[0055] Figure 8 is the second flow schematic diagram of the device model parameter extraction method provided by the present invention;

[0056] Figure 9 is the flow schematic diagram of the device selection method provided by the present invention;

[0057] Figure 10 is the structural schematic diagram of the device model parameter extraction device provided by the present invention;

[0058] Figure 11 is the structural schematic diagram of the device selection device provided by the present invention;

[0059] Figure 12 is the structural schematic diagram of the electronic device provided by the present invention. Detailed implementation manners

[0060] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Apparently, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art without making creative efforts based on the embodiments in the present invention belong to the scope of protection of the present invention.

[0061] Currently, the acquisition of device model parameters is mainly achieved through experimental measurement methods, simulation methods, standard model methods, and data manual methods.

[0062] Among them, the experimental measurement method is a method of obtaining model parameters by actually measuring the characteristics of devices. For example, for an inductor, devices such as an LCR (Resistor - Inductor - Capacitor) meter or a network analyzer can be used to measure its inductance value and series resistance value, so as to establish a model. However, the test results of some devices cannot directly give the RLC values of the model. In this case, circuit simulation software such as Keysight ADS and SPICE is also needed to perform circuit simulation on the device, and compare the simulation results with the measured results to fit the best model parameters.

[0063] First, the experimental measurement method requires devices for measuring the parameters of specific passive devices, and the purchase price of these devices is usually high. Second, the experimental measurement method also requires researchers to design and manufacture corresponding test fixtures for fixing passive devices for testing. For devices whose model parameters cannot be directly extracted, circuit simulation software is additionally needed to perform circuit simulation and compare the simulation results with the measured results to fit the best parameters.

[0064] Therefore, the experimental measurement method requires a large investment of funds and manpower in extracting device model parameters.

[0065] The simulation method is a method of performing three - dimensional modeling on passive devices using electromagnetic field simulation software such as Ansys HFSS and CST Studio Suite. The parameter model of the device can be extracted through the simulation method.

[0066] When performing three - dimensional modeling, it is usually necessary to know the structural shape and size information of the device, as well as the material property information used in the device, such as parameters like conductivity, permittivity, and dielectric loss factor. However, these parameters usually involve trade secrets, and device manufacturers generally do not provide them, or the provided parameters deviate from the actual situation. In addition, the electromagnetic field simulation software required for modeling also needs to be purchased separately and is expensive.

[0067] The standard model method is a modeling method applicable to some passive devices with standardized models, such as the ideal linear model of an inductor and the ideal equivalent circuit model of a capacitor. Modeling can be directly performed using the standard model according to the specifications and characteristic data of the device.

[0068] However, there are some limitations in using the ideal device model. Compared with the RLC equivalent model circuit, there are large deviations in circuit characteristics and accuracy in the ideal device model. Therefore, it is difficult to accurately match the actual situation when using the ideal device model for simulation.

[0069] The data - sheet method is a method of obtaining model parameters by referring to the data sheets of devices provided by manufacturers.

[0070] It should be noted that data manuals generally only provide electrical characteristic parameters, such as breakdown voltage values, DC impedance values, impedance-frequency diagrams, etc. These parameters do not directly give the RLC equivalent circuit of the device and its parameter values. Therefore, it is still necessary to manually read the frequency-impedance diagram and then calculate the RLC model of the device. Moreover, since the frequency-impedance curve corresponding to the model cannot be fitted and compared with the impedance diagram in the data manual, the accuracy of the parameters cannot be determined, which leads to certain uncertainties and errors in both model parameter extraction and device selection.

[0071] To solve the above problems, the present invention provides a method for extracting device model parameters. Figure 1 It is one of the schematic flowcharts of the method for extracting device model parameters provided by the present invention, as Figure 1 shown, the method includes:

[0072] Step 110, obtaining a parameter curve diagram of the device.

[0073] The device here, that is, the device for which model parameter extraction is required, can be various resistors, capacitors, inductors, magnetic beads, etc., and the embodiments of the present invention do not make specific limitations in this regard.

[0074] The parameter curve diagram of the device can be an image recorded in the data manual provided by the manufacturer and obtained through forms such as photographing or screenshotting. For example, it can be a frequency-impedance diagram. Further, after obtaining the parameter curve diagram through photographing or screenshotting, a series of image processing operations can be performed on the parameter curve diagram. For example, through image segmentation, the area irrelevant to the parameter curve in the parameter curve diagram can be cropped off.

[0075] Step 120, detecting the positions of each scale on the coordinate axes of the parameter curve diagram, and based on the positions of each scale, associating each scale with the scale text in the parameter curve diagram to obtain the scale values of each coordinate pixel point on the coordinate axes.

[0076] Specifically, after obtaining the parameter curve diagram, a coordinate system corresponding to the parameter curve diagram can be established by detecting the coordinate axes and the scales on the coordinate axes in the parameter curve diagram.

[0077] In this process, through the object detection method for images, the coordinate axes in the parameter curve diagram can be detected, thereby determining the positions of the coordinate axes, and each scale on the coordinate axes can be detected, thereby determining the positions of each scale on the coordinate axes.

[0078] After determining the positions of each scale on the coordinate axis, the scales can be associated with the scale texts in the parameter curve graph, and thus the scale values corresponding to each scale on the coordinate axis can be obtained. It can be understood that here, the scale texts in the parameter curve graph can be obtained by performing text recognition on the parameter curve graph through OCR (Optical Character Recognition) technology. Through OCR technology, not only can the scale texts in the parameter curve graph be recognized, but also the positions of the scale texts in the parameter curve graph can be obtained.

[0079] Thus, when associating the scales with the scale texts in the parameter curve graph, the positions of the scales can be matched with the positions of the scale texts, and the scale texts with successful position matching are determined as the scale values corresponding to the corresponding scales. Alternatively, considering the case where the scales are relatively dense, the distances between adjacent scales can also be determined based on the positions of the scales, and by comparing the changes in the distances, the scales reflecting the magnitude changes can be determined from the scales, and these scales are preferentially associated with the scale texts. The embodiments of the present invention do not make specific limitations on this.

[0080] After realizing the association between the scales on the coordinate axis and the scale texts, the scale values corresponding to each coordinate pixel point on the coordinate axis can be determined based on the scale values reflected by the scale texts associated with the scales, thereby realizing the establishment of the coordinate system.

[0081] Step 130, scan the curve points corresponding to each coordinate pixel point on the coordinate axis in the parameter curve graph.

[0082] Specifically, after completing the establishment of the coordinate axis, the pixel points corresponding to each coordinate pixel point on the coordinate axis can be traversed in the parameter curve graph, and the curve points corresponding to each coordinate pixel point can be selected therefrom. For example, in a frequency-impedance graph, there are a horizontal coordinate axis representing frequency and a vertical coordinate axis representing impedance. Here, taking the horizontal coordinate axis as an example, for each coordinate pixel point on the horizontal coordinate axis representing frequency, scan in the frequency-impedance graph to determine a column of pixel points corresponding to each coordinate pixel point in the coordinate system, and select the curve point corresponding to the coordinate pixel point from this column of pixel points, that is, the pixel point reflecting the impedance value at the frequency corresponding to the coordinate pixel point.

[0083] Furthermore, the way to select the curve points can be to select the pixel point with the highest pixel value in the set of pixel points corresponding to the coordinate pixel point in a black-and-white curve graph, or to select the pixel point with the pixel value consistent with the curve color in the set of pixel points corresponding to the coordinate pixel point in a color curve graph.

[0084] Step 140, determine the model parameters of the device based on the scale values of the coordinate pixel points and the curve points corresponding to the coordinate pixel points.

[0085] Specifically, after obtaining the scale values of each coordinate pixel point on the coordinate axis and the curve points corresponding to each coordinate pixel point on the coordinate axis, the relationship between the scale value represented by each coordinate pixel point and the curve point can be constructed. For example, for a frequency-impedance diagram, the relationship between frequency and impedance can be obtained. It can be understood that the relationship between the scale value represented by each coordinate pixel point obtained thereby and the curve point can be used as the model parameters of the device for device selection and circuit simulation applications.

[0086] The method provided by the embodiment of the present invention associates each scale on the coordinate axis of the parameter curve diagram with the scale text to obtain the scale value of each coordinate pixel point on the coordinate axis. In addition, the curve points corresponding to each coordinate pixel point are obtained by scanning, thereby associating the curve points with the scale values to obtain the model parameters of the device corresponding to the parameter curve diagram, so as to realize the extraction of model parameters through image processing, while reducing the cost of model parameter extraction and ensuring the reliability and accuracy of model parameter extraction.

[0087] Based on the above embodiments, Figure 2 is a schematic flowchart of the curve point scanning method provided by the present invention, as Figure 2 shown, step 130 includes:

[0088] Step 131, scan one by one and obtain the set of pixel points corresponding to each coordinate pixel point on the first coordinate axis in the parameter curve diagram, where the first coordinate axis is any coordinate axis of the parameter curve diagram;

[0089] Step 132, based on the positional relationship between each coordinate pixel point and the scale on the first coordinate axis, determine the curve point corresponding to each coordinate pixel point from the set of pixel points corresponding to each coordinate pixel point.

[0090] Specifically, after the coordinate system is established, any coordinate axis in the coordinate system can be used as the first coordinate axis. By traversing the pixel points corresponding to each coordinate pixel point on the first coordinate axis in the parameter curve diagram, the set of pixel points corresponding to each coordinate pixel point can be obtained. Here, the set of pixel points may include each pixel point corresponding to the coordinate pixel point in the coordinate system of the parameter curve diagram. For example, taking the horizontal coordinate axis as the first coordinate axis and scanning each coordinate pixel point based on the horizontal coordinate axis, the set of pixel points for each coordinate pixel point is the pixel points with the same abscissa as the coordinate pixel point in the coordinate system, that is, the pixel points in the same column as the coordinate pixel point; or taking the vertical coordinate axis as the first coordinate axis and scanning each coordinate pixel point based on the vertical coordinate axis, the set of pixel points for each coordinate pixel point is the pixel points with the same ordinate as the coordinate pixel point in the coordinate system, that is, the pixel points in the same row as the coordinate pixel point.

[0091] When performing a scan of each coordinate pixel point, for the horizontal axis, it is possible to scan column by column from left to right along the direction of the axis; for the vertical axis, it is possible to scan row by row from bottom to top along the direction of the axis.

[0092] For any coordinate pixel point, after scanning to obtain the corresponding pixel point set of this coordinate pixel point, it is possible to select the curve point corresponding to this coordinate pixel point from the pixel point set according to the positional relationship between this coordinate pixel point and the scale, that is, whether this coordinate pixel point coincides with the scale on the first axis.

[0093] It can be understood that for the convenience of viewing, grids are often divided in the parameter curve graph, and the division of the grids is based on the scales on the axes, that is, grid lines are extended at the scales. In this way, if the coordinate pixel point coincides with the scale on the first axis, the pixel values of each pixel point in the pixel point set corresponding to this coordinate pixel point may all be the pixel values of the grid color. In this case, it is not easy to determine which specific pixel point is the curve point through the pixel values of each pixel point in the pixel point set.

[0094] Therefore, in the process of determining the curve point, it is possible to select the curve point from the pixel point set by judging the positional relationship between the coordinate pixel point and the scale and applying the selection method of the curve point corresponding to the positional relationship. For example, in the case where the coordinate pixel point coincides with the scale on the first axis, it is possible to determine the curve point corresponding to this coordinate pixel point based on the curve point corresponding to the coordinate pixel point adjacent to this coordinate pixel point. In the case where the coordinate pixel point does not coincide with the scale on the first axis, it is possible to determine the curve point corresponding to this coordinate pixel point based on the pixel values of each pixel point in the pixel point set.

[0095] Based on any of the above embodiments, step 132 includes:

[0096] In the case where the positional relationship is non - coincidence, based on the pixel values of each pixel point in the pixel point set, and / or, the positional relationship between each pixel point in the pixel point set and the scale on the second axis, determine the curve point corresponding to the coordinate pixel point from the pixel point set, where the second axis is an axis perpendicular to the first axis.

[0097] Specifically, the second axis is the axis before the first axis in the coordinate system. In the parameter curve graph, the first axis is perpendicular to the second axis.

[0098] When the positional relationship between the coordinate pixel point and the scale is non - coincident, it indicates that not all the pixel points in the pixel point set corresponding to the coordinate pixel point are grid pixel points. In this case, based on the positional relationship between each pixel point in the pixel point set and the scale on the second coordinate axis, that is, whether each pixel point is located on the grid line extended from the scale on the second coordinate axis, the pixel points located on the grid line can be screened out from the pixel point set to avoid interference of this part of pixel points on the selection of curve points.

[0099] In addition, curve points can also be screened out based on the pixel values of the remaining pixel points in the pixel point set. For example, pixel points with pixel values consistent with the curve color can be screened out from the pixel point set as curve points; or, non - background - color pixel points, that is, pixel points that may be curve points, can be screened out from the pixel point set, and then the curve points can be determined from them.

[0100] For example, Figure 3 is one of the partial screenshots of the parameter curve graph provided by the present invention. As Figure 3 shown, in a certain column in the figure, white pixel points can be screened out through pixel values, and pixel points located on the grid line can be screened out through the relationship between the pixel points and the scale on the vertical coordinate axis. The pixel points thus retained, that is, Figure 3 the two points marked in black in are respectively the curve points on the impedance Z curve and the curve points on the equivalent series resistance ESR (Equalized Series Resistance). According to circuit principles, the impedance value Z is greater than the equivalent series resistance ESR. When extracting the impedance Z curve, the point with a larger ordinate can be selected from these two points as the curve point of the impedance Z curve.

[0101] It should be noted that during the process of determining curve points, based on the positional relationship between each pixel point in the pixel point set and the scale on the second coordinate axis, the pixel points located on the grid line can be screened out from the pixel point set. This operation may cause the accidental deletion of curve points located on the grid line, resulting in the inability to extract the correct curve points from the remaining pixel points in the pixel point set. To address this issue, when curve points cannot be extracted from the remaining pixel points in the pixel point set, the curve point corresponding to the previous coordinate pixel point of the coordinate pixel point can be recorded as the previous curve point, and the curve point corresponding to this coordinate pixel point can be determined based on the previous curve point. For example, when the extracted curve is in an increasing trend, add 1 to the ordinate of the previous curve point as the ordinate of the curve point corresponding to this coordinate pixel point; when the extracted curve is in a decreasing trend, subtract 1 from the ordinate of the previous curve point as the ordinate of the curve point corresponding to this coordinate pixel point; when the extracted curve is neither in an increasing nor a decreasing trend, use the ordinate of the previous curve point as the ordinate of the curve point corresponding to this coordinate pixel point.

[0102] Based on any of the above embodiments, in step 132, determining the curve point corresponding to the coordinate pixel point from the pixel point set based on the pixel values of the pixel points in the pixel point set, and / or, the positional relationship between the pixel points in the pixel point set and the scales on the second coordinate axis, includes:

[0103] Determine a plurality of candidate line segments based on the pixel values of the pixel points in the pixel point set;

[0104] Determine the weight of each candidate line segment based on the number of pixel points constituting each candidate line segment;

[0105] Determine the target line segment from the candidate line segments based on the weights of the candidate line segments and the distances between the midpoints of the candidate line segments and the previous curve point respectively; the previous curve point is the curve point corresponding to the previous coordinate pixel point of the coordinate pixel point;

[0106] Determine the curve point corresponding to the coordinate pixel point based on the target line segment.

[0107] Specifically, Figure 4 is the second partial screenshot of the parameter curve graph provided by the present invention, Figure 4 The solid line in is the content of the screenshot of the parameter curve graph, and the dotted line is an auxiliary line drawn for easy understanding and explanation. Among them, each point on the vertical dotted line, that is, each pixel point in the pixel point set corresponding to the coordinate pixel point. Figure 4 It can be seen from that there are multiple situations where curves overlap and then separate in the column where the coordinate pixel point is located. At this time, the candidate line segments existing in the pixel point set can be marked first, for example Figure 4 L1 and L2 marked in.

[0108] At this time, specifically, based on the pixel values of the pixel points in the pixel point set, each pixel point that may be a curve point in the pixel point set can be screened out, and whether each pixel point is connected into a line segment can be judged by the positions between the pixel points that may be curve points, thereby obtaining candidate line segments. It can be understood that in the case of multiple curves overlapping and separating, multiple candidate line segments can be obtained.

[0109] It can be understood that the candidate line segments are composed of pixel points, and the more pixel points that make up the candidate line segments, the higher the probability that the candidate line segments are parts of the curve to be extracted. Therefore, the weights of the candidate line segments can be determined based on the number of pixel points constituting the candidate line segments. For example, the weights of the candidate line segments can be determined based on the following formula:

[0110] weight = 2 deltaL

[0111] In the formula, weight represents the weight of the candidate line segment, and deltaL is the number of pixel points constituting the candidate line segment.

[0112] After obtaining the weights of the candidate line segments, the target line segment can be determined from the candidate line segments by combining the weights of the candidate line segments and the distances between the midpoints of the candidate line segments and the previous curve point respectively. Here, the previous curve point is the curve point corresponding to the previous coordinate pixel point of the coordinate pixel point. The smaller the distance between the previous curve point and the midpoint of any candidate line segment, the greater the probability that the candidate line segment is connected to the previous curve point, and the higher the probability that the candidate line segment is a segment of the curve to be extracted, that is, the target line segment.

[0113] Therefore, the probability that the candidate line segment is the target line segment can be measured by combining the weight of the candidate line segment and the distance between the midpoint of the candidate line segment and the previous curve point, so as to select the target line segment from the candidate line segments. For example, the average value of the ordinates of each pixel point in the candidate line segment can be calculated, that is, the ordinate of the midpoint of the candidate line segment is obtained, and the absolute value of the difference between this value and the ordinate of the previous curve point is calculated as the distance between the two. The smaller the value obtained by dividing the distance by the weight of the candidate line segment, the more likely the candidate line segment is the target line segment.

[0114] After determining the target line segment from the candidate line segments, the midpoint of the target line segment can be used as the curve point corresponding to the coordinate pixel point.

[0115] Based on any of the above embodiments, step 132 includes:

[0116] In the case where the positional relationship is coincidence, based on the previous curve point, the curve point corresponding to the coordinate pixel point is determined from the pixel point set corresponding to the coordinate pixel point; the previous curve point is the curve point corresponding to the previous coordinate pixel point of the coordinate pixel point.

[0117] Specifically, Figure 5 It is the third partial screenshot of the parameter curve graph provided by the present invention. As Figure 3 shown, in the parameter curve graph, there is a situation where the curve intersects the grid. For example, at the position indicated by the arrow. Here, assuming that the first coordinate axis is the horizontal coordinate axis, the coordinate pixel point on the horizontal coordinate axis coincides with the scale on the horizontal coordinate axis, and the grid is divided longitudinally from the scale. Therefore, all the pixel points in the pixel point set corresponding to the coordinate pixel point are grid pixel points, and it is difficult to determine the curve point from the pixel values.

[0118] In response to this situation, the curve point corresponding to the previous coordinate pixel of the coordinate pixel can be denoted as the previous curve point, and the curve point corresponding to the coordinate pixel is determined based on the previous curve point. For example, when the extracted curve is in an increasing trend, add 1 to the ordinate of the previous curve point as the ordinate of the curve point corresponding to the coordinate pixel; when the extracted curve is in a decreasing trend, subtract 1 from the ordinate of the previous curve point as the ordinate of the curve point corresponding to the coordinate pixel; when the extracted curve is neither in an increasing nor decreasing trend, use the ordinate of the previous curve point as the ordinate of the curve point corresponding to the coordinate pixel.

[0119] In particular, for the first coordinate axis, the set of pixel points corresponding to the first coordinate pixel on this axis is usually the set of pixel points used to form the second coordinate axis. Similarly, the curve point corresponding to this coordinate pixel cannot be obtained, and the curve point corresponding to this coordinate pixel can be selectively ignored.

[0120] Based on any of the above embodiments, in step 120, the associating the scales with the scale texts in the parameter curve graph to obtain the scale values of the coordinate pixels on the coordinate axis includes:

[0121] Dividing the scales into different magnitudes based on the change in the distance between adjacent scales;

[0122] Based on the magnitudes to which the scales belong, associating the scales with the scale texts to obtain the scale values of the coordinate pixels on the coordinate axis.

[0123] Specifically, considering the actual processing situation, the scale distributions of some coordinate axes are not evenly distributed. For example Figure 6 is the frequency impedance graph provided by the present invention, Figure 6 in which, for the horizontal coordinate axis of frequency and the vertical coordinate axis of impedance (Impedance-ESR), both are exponentially increasing or decreasing, the distances between adjacent scales are not the same, and the scale intervals within the same magnitude are arranged in a decreasing order.

[0124] In response to this situation, after identifying the positions of the scales on the coordinate axis of the parameter curve graph, the distance between every two adjacent scales can be calculated, and based on the change in the distance, the scales can be divided into different magnitudes. For example, in Figure 6Among them, the scale values 1 - 10 corresponding to the 1st scale to the 10th scale are all within the same order of magnitude; when crossing an order of magnitude, the distance between adjacent scales arranges to repeat that of the previous order of magnitude, that is, the distance between the 10th scale and the 11th scale is the same as the distance between the 1st scale and the 2nd scale, and the distances of other scales after the 11th scale repeat the distances of other scales after the 2nd scale, and so on. The distance of the first scale of the new order of magnitude is larger than the distance of the last scale of the previous order of magnitude. For example, Figure 6 the distance between the 10th scale and the 11th scale in Figure 6 is much larger than the distance between the 9th scale and the 10th scale.

[0125] After determining the order of magnitude to which each scale belongs, the association between the scale and the scale text can be realized based on the order of magnitude to which each scale belongs. For example, on a logarithmic coordinate axis, generally only the scale positions that are multiples of 10 will be marked with scale values. For example, Figure 6 the scale positions of 1, 10, and 100 in Figure 6 . Therefore, when the scale distance changes from decreasing to increasing, it can indicate that the order of magnitude has changed, that is, a scale that is a multiple of 10 appears. At this time, the scale text obtained through text recognition can be associated with the scale as the scale value of the scale. Generally, the smallest scale value is associated with the grid line closest to the origin side of the coordinate.

[0126] After completing the association between the scale and the scale text, that is, after obtaining the scale values of each scale, the scale values of each coordinate pixel point on the coordinate axis can be calculated based on a preset formula. It can be understood that the coordinate pixel points here can include non-scale pixel points. For example, Figure 7 is the fourth partial screenshot of the parameter curve graph provided by the present invention. On the Figure 7 shown logarithmic coordinate axis, the calculation formula for the scale value can be:

[0127]

[0128] In the formula, the coordinate pixel point for which the scale value is to be calculated is denoted as x, and X is the calculated scale value. Figure 7 On the shown logarithmic coordinate axis, starting from 0, it increases in steps of 1 in sequence. Assume x0 = 10, and the corresponding pixel point serial number is P0Idx; x1 = 100, and the corresponding pixel point serial number is P1Idx; x = X, and the corresponding pixel serial number is PixIdx.

[0129] Based on any of the above embodiments, the steps for determining the scale text include:

[0130] Locate the text area in the parameter curve graph;

[0131] Detect the text contour within the text area;

[0132] Based on each text contour detected within the text region, perform text recognition on the parameter curve graph to obtain the scale text in the parameter curve graph.

[0133] Specifically, before associating each scale on the coordinate axis with the scale text, it is necessary to first identify the scale text from the parameter curve graph.

[0134] To identify the scale text in the parameter curve graph, it is first necessary to locate the text region in the parameter curve graph. Here, the text region is the text region beside the coordinate axis in the parameter curve graph that can reflect the coordinate axis scale, and the text region may contain the scale text of the coordinate axis. The positioning of the text region can be achieved through a pre-trained object recognition model. It can be understood that after locating the text region, further processing can be performed on the text region to ensure its accuracy and reliability. Specifically, morphological operations can be applied to connect or fill the text region, and then dilation and erosion processing can be performed on the text region.

[0135] After completing the positioning of the text region, text contour detection can be performed within the text region. Here, the text contour detection can be achieved through Opencv contour detection.

[0136] After completing the text contour detection, each text contour in the text region can be traversed, and the text corresponding to each text contour in the parameter curve graph can be recognized, thereby obtaining the text corresponding to each text contour in the text region, that is, the scale text in the parameter curve graph. Here, the text recognition can be achieved through OCR technology.

[0137] Based on any of the above embodiments, Figure 8 is the second flow schematic diagram of the device model parameter extraction method provided by the present invention. As Figure 8 shown, the method includes:

[0138] First, import the frequency impedance curve graph:

[0139] It can be understood that the frequency impedance curve graph here is the parameter curve graph for which model parameter extraction is required.

[0140] Second, start splitting the picture to obtain the picture unit region, the coordinate axis grid region, and the coordinate scale value region:

[0141] Here, the picture splitting refers to splitting the parameter curve graph, that is, the frequency impedance curve graph, to obtain each functional region in the parameter curve graph. Among them, the image unit region is the region in the parameter curve graph that identifies the coordinate unit. For example Figure 6The area marked with "Frequency (kHz)" and "Impedance - ESR (ohm)"; the coordinate axis grid area is the coordinate system area representing the curve, for example Figure 6 The area with grids distributed in it; the coordinate scale value area is the area marked with scale text, for example Figure 6 The area marked with "0.1" to "100000" for frequency and the area marked with "0.001" to "1000" for ESR.

[0142] Subsequently, extract the unit information from the picture unit area:

[0143] Here, the extraction of unit information can be achieved through OCR.

[0144] In addition, extract the coordinate axis scale values from the coordinate axis grid area to establish a coordinate system:

[0145] The coordinate axes in the coordinate axis grid area diagram can be detected to determine the positions of the coordinate axes, and each scale on the coordinate axes can be detected to determine the positions of each scale on the coordinate axes.

[0146] After determining the positions of each scale on the coordinate axes, the scales can be associated with the scale text in the scale value area, so as to obtain the scale values corresponding to each scale on the coordinate axes, and then determine the scale values corresponding to each coordinate pixel point on the coordinate axes, thereby realizing the establishment of the coordinate system.

[0147] Next, extract the curve information from the coordinate axis grid area:

[0148] After the coordinate axes are established, by traversing the pixel points corresponding to each coordinate pixel point on the coordinate axes in the coordinate axis grid area, and selecting the curve points corresponding to each coordinate pixel point from them, and associating the scale values of each coordinate pixel point on the coordinate axes with the curve points corresponding to each coordinate pixel point on the coordinate axes, the curve information can be obtained, that is, the model parameters corresponding to the frequency - impedance curve diagram.

[0149] The method provided by the embodiment of the present invention associates each scale on the coordinate axes of the parameter curve diagram with the scale text to obtain the scale values of each coordinate pixel point on the coordinate axes. In addition, the curve points corresponding to each coordinate pixel point are obtained by scanning, so as to associate the curve points with the scale values, and obtain the model parameters of the device corresponding to the parameter curve diagram, thereby realizing the extraction of model parameters through image processing, while reducing the cost of model parameter extraction and ensuring the reliability and accuracy of model parameter extraction.

[0150] Based on any of the above embodiments, Figure 9 is the flow schematic diagram of the device selection method provided by the present invention, as Figure 9As shown, the method includes:

[0151] Step 910, determine the target device parameters.

[0152] Here, the target device parameters refer to the device parameters that the target device is expected to have when selecting a device. The target device parameters can be pre-determined or calculated and inferred according to circuit design requirements or simulation requirements.

[0153] For example, for a capacitor in an RLC series structure, the impedance formula can be expressed as:

[0154] Z = R + jwL + 1 / (jwC)

[0155] In the formula, Z is the impedance, R is the resistance, L is the inductance, C is the capacitance, w = 2πf, and f is the frequency.

[0156] The RLC parameters of passive devices can be deduced through the frequency point position coordinate information. Taking the capacitor as an example, the deduction process is as follows: First, find the position of the resonance point and obtain the resonance frequency Fres. Then, take the frequency point Fw that is 10 times lower than the resonance point. This point belongs to the low-frequency band, and the device is capacitive. It can be calculated that Z = 1 / (2πFw*C), so C = 1 / (2πFw*Z). At the resonance point, the device is purely resistive, R = Z; the resonance frequency It is calculated that L = (2πFres) 2 . Above, the calculation of the RLC parameters of a group of capacitors is completed and can be used as the target device parameters.

[0157] Step 920, match the target device parameters with the model parameters of each candidate device, and determine the candidate device with the highest matching degree as the target device;

[0158] The model parameters of the candidate devices are determined based on the device model parameter extraction method provided in any of the above embodiments.

[0159] Specifically, the model parameters of each candidate device obtained based on the device model parameter extraction method provided in the above embodiments can be pre-stored. After obtaining the target device parameters, the target device parameters are respectively matched with the model parameters of each candidate device, and the candidate device with the highest matching degree is used as the target device that meets the expectations.

[0160] It can be understood that considering the model parameters of each candidate device obtained by the device model parameter extraction method provided based on the above embodiments, which can be reflected as the curve points corresponding to each coordinate pixel point in the coordinate system. When performing model parameter matching, the target model parameters can be converted into a target curve in the coordinate system and fitted with the model curves corresponding to the model parameters in the same coordinate system. It can be understood that the smaller the area enclosed between the target curve and the model curve, the higher the fitting degree between the target region and the model curve, and the higher the matching degree between the target model parameters and the model parameters. After curve fitting the model curve corresponding to the target model parameters with the model curves of each candidate device respectively, the candidate device with the highest fitting degree can be selected as the target device, thereby completing device selection.

[0161] The method provided by the embodiments of the present invention associates each scale on the coordinate axis of the parameter curve graph with the scale text to obtain the scale values of each coordinate pixel point on the coordinate axis. In addition, each curve point corresponding to each coordinate pixel point is obtained by scanning, thereby associating the curve point with the scale value to obtain the model parameters of the device corresponding to the parameter curve graph, so as to realize the extraction of model parameters through image processing, while reducing the cost of model parameter extraction and ensuring the reliability and accuracy of model parameter extraction. Based on this, device selection is carried out, greatly reducing the selection difficulty and effectively improving the selection success rate.

[0162] Figure 10 is a schematic structural diagram of a device model parameter extraction device provided by the present invention, as Figure 10 shown, the device includes:

[0163] An acquisition unit 1010, which acquires a parameter curve graph of a device;

[0164] A scale detection unit 1020, configured to detect the positions of each scale on the coordinate axis of the parameter curve graph, and based on the positions of each scale, associate each scale with the scale text in the parameter curve graph to obtain the scale values of each coordinate pixel point on the coordinate axis;

[0165] A curve scanning unit 1030, configured to scan the curve points corresponding to each coordinate pixel point on the coordinate axis in the parameter curve graph;

[0166] A modeling unit 1040, configured to determine the model parameters of the device based on the scale values of each coordinate pixel point and the curve points corresponding to each coordinate pixel point.

[0167] The device provided by the embodiment of the present invention associates each scale on the coordinate axis of the parameter curve graph with scale text to obtain the scale values of each coordinate pixel point on the coordinate axis. In addition, by scanning, the curve points corresponding to each coordinate pixel point are obtained. Thus, the curve points and the scale values are associated to obtain the model parameters of the device corresponding to the parameter curve graph, so as to realize the extraction of model parameters through image processing, while reducing the cost of model parameter extraction and ensuring the reliability and accuracy of model parameter extraction.

[0168] Based on any of the above embodiments, the curve scanning unit includes:

[0169] A scanning subunit, configured to scan one by one and obtain a set of pixel points corresponding to each coordinate pixel point on the first coordinate axis in the parameter curve graph, where the first coordinate axis is any coordinate axis of the parameter curve graph;

[0170] A selection subunit, configured to determine the curve points corresponding to the respective coordinate pixel points from the set of pixel points corresponding to the respective coordinate pixel points based on the positional relationship between the respective coordinate pixel points and the scale on the first coordinate axis.

[0171] Based on any of the above embodiments, the selection subunit is specifically configured to:

[0172] In the case where the positional relationship is non - coincident, determine the curve points corresponding to the coordinate pixel points from the set of pixel points based on the pixel values of the respective pixel points in the set of pixel points, and / or the positional relationship between the respective pixel points in the set of pixel points and the scale on the second coordinate axis, where the second coordinate axis is a coordinate axis perpendicular to the first coordinate axis.

[0173] Based on any of the above embodiments, the selection subunit is specifically configured to:

[0174] Determine a plurality of candidate line segments based on the pixel values of the respective pixel points in the set of pixel points;

[0175] Determine the weights of the respective candidate line segments based on the number of pixel points constituting each candidate line segment;

[0176] Determine a target line segment from the respective candidate line segments based on the weights of the respective candidate line segments and the distances between the mid - points of the respective candidate line segments and the previous curve point; the previous curve point is the curve point corresponding to the previous coordinate pixel point of the coordinate pixel point;

[0177] Determine the curve point corresponding to the coordinate pixel point based on the target line segment.

[0178] Based on any of the above embodiments, the selection subunit is specifically configured to:

[0179] When the positional relationship is coincidence, based on the previous curve point, determine the curve point corresponding to the coordinate pixel point from the set of pixel points corresponding to the coordinate pixel point; the previous curve point is the curve point corresponding to the previous coordinate pixel point of the coordinate pixel point.

[0180] Based on any of the above embodiments, the scale detection unit is specifically configured to:

[0181] Divide the magnitudes to which the respective scales belong based on the change in the distance between the adjacent scales;

[0182] Associate the respective scales with the scale text based on the magnitudes to which the respective scales belong, to obtain the scale values of the respective coordinate pixel points on the coordinate axis.

[0183] Based on any of the above embodiments, the scale detection unit is further configured to:

[0184] Locate the text area in the parameter curve graph;

[0185] Detect the text contour within the text area;

[0186] Perform text recognition on the parameter curve graph based on each text contour detected within the text area, to obtain the scale text in the parameter curve graph.

[0187] Based on any of the above embodiments, Figure 11 is a structural schematic diagram of the device selection device provided by the present invention, as Figure 11 shown, the device includes:

[0188] A parameter determination unit 1110, configured to determine target device parameters;

[0189] A device matching unit 1120, configured to match the target device parameters with the model parameters of each candidate device, and determine the candidate device with the highest matching degree as the target device;

[0190] The model parameters of the candidate device are determined based on the device model parameter extraction method described in the above embodiments.

[0191] The device provided by the embodiment of the present invention associates the respective scales on the coordinate axis of the parameter curve graph with the scale text, to obtain the scale values of the respective coordinate pixel points on the coordinate axis. Additionally, by scanning, the curve points corresponding to the respective coordinate pixel points are obtained, and thus the curve points and the scale values are associated to obtain the model parameters of the device corresponding to the parameter curve graph, thereby realizing the extraction of the model parameters through image processing. While reducing the cost of extracting the model parameters, the reliability and accuracy of the extraction of the model parameters are ensured. Based on this, device selection is performed, greatly reducing the selection difficulty and effectively improving the selection success rate.

[0192] Figure 12 An entity structure diagram of an electronic device is illustrated, as Figure 12 shown. The electronic device may include: a processor 1210, a communications interface 1220, a memory 1230, and a communication bus 1240. Among them, the processor 1210, the communications interface 1220, and the memory 1230 communicate with each other through the communication bus 1240. The processor 1210 may call logic instructions in the memory 1230 to execute a device model parameter extraction method, which includes:

[0193] Obtain a parameter curve graph of the device;

[0194] Detect the positions of the scales on the coordinate axes of the parameter curve graph, and based on the positions of the scales, associate the scales with the scale texts in the parameter curve graph to obtain the scale values of the coordinate pixel points on the coordinate axes;

[0195] Scan the curve points corresponding to the coordinate pixel points on the coordinate axes in the parameter curve graph;

[0196] Based on the scale values of the coordinate pixel points and the curve points corresponding to the coordinate pixel points, determine the model parameters of the device.

[0197] The processor 1210 may also call logic instructions in the memory 1230 to execute a device selection method, which includes:

[0198] Determine the target device parameters;

[0199] Match the target device parameters with the model parameters of each candidate device, and determine the candidate device with the highest matching degree as the target device;

[0200] The model parameters of the candidate devices are determined based on the device model parameter extraction method.

[0201] In addition, when the logical instructions in the above-mentioned memory 1230 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0202] On the other hand, the present invention also provides a computer program product. The computer program product includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the device model parameter extraction method provided by each of the above methods. This method includes:

[0203] Obtain the parameter curve graph of the device;

[0204] Detect the positions of each scale on the coordinate axes of the parameter curve graph, and based on the positions of each scale, associate each scale with the scale text in the parameter curve graph to obtain the scale values of each coordinate pixel point on the coordinate axes;

[0205] Scan the curve points corresponding to each coordinate pixel point on the coordinate axes in the parameter curve graph;

[0206] Based on the scale values of each coordinate pixel point and the curve points corresponding to each coordinate pixel point, determine the model parameters of the device.

[0207] When the computer program is executed by a processor, the computer can also execute the device selection method provided by each of the above methods. This method includes:

[0208] Determine the target device parameters;

[0209] Match the target device parameters with the model parameters of each candidate device, and determine the candidate device with the highest matching degree as the target device;

[0210] The model parameters of the candidate device are determined based on the device model parameter extraction method.

[0211] In another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements a device model parameter extraction method provided by the above-mentioned various methods. The method includes:

[0212] Obtain the parameter curve graph of the device;

[0213] Detect the positions of the scales on the coordinate axes of the parameter curve graph, and based on the positions of the scales, associate the scales with the scale texts in the parameter curve graph to obtain the scale values of the coordinate pixel points on the coordinate axes;

[0214] Scan the curve points corresponding to the coordinate pixel points on the coordinate axes in the parameter curve graph;

[0215] Based on the scale values of the coordinate pixel points and the curve points corresponding to the coordinate pixel points, determine the model parameters of the device.

[0216] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements a device selection method provided by the above-mentioned various methods. The method includes:

[0217] Determine the target device parameters;

[0218] Match the target device parameters with the model parameters of each candidate device, and determine the candidate device with the highest matching degree as the target device;

[0219] The model parameters of the candidate devices are determined based on the device model parameter extraction method.

[0220] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative labor.

[0221] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0222] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for extracting device model parameters, characterized in that, it includes: Obtaining the parameter curve graph of the device; Detecting the positions of the scales on the coordinate axes of the parameter curve graph, and associating each scale with the scale text in the parameter curve graph based on the positions of the scales to obtain the scale values of the coordinate pixel points on the coordinate axes; Scanning the curve points corresponding to the coordinate pixel points on the coordinate axes in the parameter curve graph; Determining the model parameters of the device based on the scale values of the coordinate pixel points and the curve points corresponding to the coordinate pixel points.

2. The method for extracting device model parameters according to claim 1, characterized in that, the scanning of the curve points corresponding to the coordinate pixel points on the coordinate axes in the parameter curve graph includes: Scanning one by one and obtaining the set of pixel points corresponding to the coordinate pixel points on the first coordinate axis in the parameter curve graph, where the first coordinate axis is any coordinate axis of the parameter curve graph; Based on the positional relationship between each coordinate pixel point and the scale on the first coordinate axis, determining the curve points corresponding to each coordinate pixel point from the set of pixel points corresponding to each coordinate pixel point.

3. The method for extracting device model parameters according to claim 2, characterized in that, the determining of the curve points corresponding to each coordinate pixel point from the set of pixel points corresponding to each coordinate pixel point based on the positional relationship between each coordinate pixel point and the scale on the first coordinate axis includes: In the case where the positional relationship is non - coincident, determining the curve point corresponding to the coordinate pixel point from the set of pixel points based on the pixel values of the pixel points in the set of pixel points, and / or the positional relationship between the pixel points in the set of pixel points and the scale on the second coordinate axis, where the second coordinate axis is the coordinate axis perpendicular to the first coordinate axis.

4. The method for extracting device model parameters according to claim 3, characterized in that, the determining of the curve point corresponding to the coordinate pixel point from the set of pixel points based on the pixel values of the pixel points in the set of pixel points, and / or the positional relationship between the pixel points in the set of pixel points and the scale on the second coordinate axis includes: Determining a plurality of candidate line segments based on the pixel values of the pixel points in the set of pixel points; Determining the weights of the candidate line segments based on the number of pixel points constituting each candidate line segment; Determining the target line segment from the candidate line segments based on the weights of the candidate line segments and the distances between the mid - points of the candidate line segments and the previous curve point respectively, where the previous curve point is the curve point corresponding to the previous coordinate pixel point of the coordinate pixel point; Determining the curve point corresponding to the coordinate pixel point based on the target line segment.

5. The method for extracting device model parameters according to claim 2, characterized in that, the determining of the curve points corresponding to each coordinate pixel point from the set of pixel points corresponding to each coordinate pixel point based on the positional relationship between each coordinate pixel point and the scale on the first coordinate axis includes: In the case where the positional relationship is coincidence, based on the previous curve point, determine the curve point corresponding to the coordinate pixel point from the set of pixel points corresponding to the coordinate pixel point; the previous curve point is the curve point corresponding to the previous coordinate pixel point of the coordinate pixel point.

6. The method for extracting device model parameters according to any one of claims 1 to 5, wherein, the associating the scales with the scale texts in the parameter curve graph to obtain the scale values of the coordinate pixel points on the coordinate axis includes: divide the magnitudes to which the scales belong based on the change in the distance between adjacent scales; associate the scales with the scale texts based on the magnitudes to which the scales belong to obtain the scale values of the coordinate pixel points on the coordinate axis.

7. The method for extracting device model parameters according to any one of claims 1 to 5, wherein, the step of determining the scale text includes: locate the text area in the parameter curve graph; detect the text contours in the text area; perform text recognition on the parameter curve graph based on each text contour detected in the text area to obtain the scale text in the parameter curve graph.

8. A device selection method, wherein, includes: determine the target device parameters; match the target device parameters with the model parameters of each candidate device, and determine the candidate device with the highest matching degree as the target device; the model parameters of the candidate device are determined based on the method for extracting device model parameters according to any one of claims 1 to 7.

9. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein, when the processor executes the program, the method for extracting device model parameters according to any one of claims 1 to 7 or the device selection method according to claim 8 is implemented.

10. A non-transitory computer-readable storage medium, on which a computer program is stored, wherein, when the computer program is executed by a processor, the method for extracting device model parameters according to any one of claims 1 to 7 or the device selection method according to claim 8 is implemented.

Images