Precision mold micropore size feature measurement method

Abstract

The invention discloses a precise mold micropore size feature measurement method, which comprises the following steps: S1, parameter initialization: reading a mold RFID tag to obtain a nominal hole depth H and a curvature radius R, and calculating an initial interlayer spacing dz0 = 0.1 H; s2, dynamic optical scanning: irradiating the surface of the micropore with a radial polarization annular light source, and detecting polarization angle deviation in real time through a polarization sensor; s3, layered image reconstruction: performing wavelet packet denoising on the image output in S2; s4, three-dimensional measurement: mapping the virtual edge coordinates output in S3 to a three-dimensional space; s5, diameter calculation and verification: calculating a horizontal diameter and a vertical diameter line by line in effective lines in the step S4; and S6, outputting a result. The laser incident angle is regulated and controlled in real time through polarization feedback, closed-loop compensation is performed on reflection distortion in combination with a mold curvature radius parameter, the measurement error of the diameter of a curved surface micropore is reduced from + / -8 microns to + / -1.5 microns, and curved surface self-adaptive accurate detection is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mold detection methods, in particular to a precision mold micro-hole size feature measurement method. BACKGROUND

[0002] A mold is a tool used to make shaped articles, which is composed of various parts, and different molds are composed of different parts. It mainly realizes the processing of the shape of an article through the change of the physical state of the formed material. It is known as the "mother of industry".

[0003] In the field of precision mold manufacturing, online detection of micro-hole size (diameter, roundness, depth) directly affects product yield. The current mainstream technology has three bottlenecks: Optical distortion is not compensated: the curvature reflection of curved molds (such as lens mold) causes edge detection distortion, and the traditional method cannot adapt to curvature changes through fixed angle scanning, with a hole diameter measurement error of up to ±8μm; Weak anti-interference ability: noise such as oil stains and scratches in the micro-hole causes edge breakage in the binary image, and the existing algorithm (such as Canny operator) has a false detection rate of more than 15%; Lack of result reliability: the output only depends on a single optical measurement, and there is a lack of a reliable verification mechanism, when the environmental temperature fluctuates ±5℃, the roundness misjudgment rate caused by thermal expansion is 22%. SUMMARY

[0004] The purpose of the present application is to provide a precision mold micro-hole size feature measurement method to solve the problems raised in the background art.

[0005] To achieve the above purpose, the present application provides the following technical solution: a precision mold micro-hole size feature measurement method, comprising the following steps: S1, parameter initialization: reading the mold RFID tag to obtain the nominal hole depth H and the curvature radius R, and calculating the initial layer spacing dz0=0.1H; S2, dynamic optical scanning: irradiating the micro-hole surface with a radial polarized ring light source, and detecting the polarization angle deviation in real time through a polarization sensor ; According to the formula Adjust the axial laser incidence angle dynamically ; Update Execute Z-axis pulse scanning, single layer exposure time , wherein, is the Z-axis speed, and DOF is the optical depth of field; S3, Layered image reconstruction: Wavelet packet denoising is performed on the output image of S2, and entity edge coordinates are segmented by U-Net; based on the entity edge coordinates, a virtual edge is generated by a ray tracing algorithm; the Hausdorff distance between the entity and the virtual edge is calculated . S4, Three-dimensional measurement: mapping the virtual edge coordinates output by S3 to three-dimensional space to reconstruct the actual hole depth Update the effective measurement row range . S5, Diameter calculation and verification: calculate the horizontal diameter and the vertical diameter row by row within the effective row of S4 and . When >10%, eliminate abnormal rows; Based on the edge consistency confidence of S3 =0.95 , . S6, Result output: when 0.95, output the thermal compensation roundness , wherein the roundness is calculated; When 0.95, activate the contact probe review and output the fusion measurement value.

[0006] The S2 interlayer spacing is dynamically adjusted: The subsequent layer spacing dz = min(0.05 , 1 μm).

[0007] The coefficient k of S2 is obtained: when the mold is replaced, the R value is updated from the new mold RFID tag by a wireless reader and k is recalculated.

[0008] The S3 virtual edge generation: Monte Carlo sampling points 2 ; wherein is the entity hole diameter output in S3.

[0009] The loss function of S3 ; wherein ∈[8,12] is set by the material refractive index n.

[0010] The thermal expansion coefficient α of S6: ; n is the refractive index of the mold material, which is read by an RFID tag.

[0011] The roundness calculation condition of S6: must be met simultaneously ≥0.95 and ≤2 pm, otherwise jump to S6.

[0012] The control system of the method, characterized in comprising: A parameter initialization module: obtaining the pre-stored nominal hole depth H and the curvature radius R in the mold tag through an RFID reader; Calculating an initial scanning layer spacing dz0=0.1H to provide a reference parameter for optical scanning; Storing and outputting R to a dynamic optical control module to realize the closed-loop starting point of curvature compensation; A dynamic optical control module: comprising a polarization monitoring unit, a light source control unit, and a pulse scanning unit, for eliminating curved surface reflection distortion; A layered reconstruction module: comprising an image preprocessing unit: performing wavelet packet denoising on the input scanning image to remove environmental interference noise points; An entity segmentation unit: extracting micro-hole entity edge coordinates through a U-Net neural network and outputting a binary entity edge image; A virtual generation unit: generating a theoretical virtual edge based on the entity edge coordinates using a ray tracing algorithm; A deviation analysis unit: calculating the Hausdorff distance between the entity and the virtual edge Outputting the calibrated edge coordinates and to the downstream module to reconstruct the real edge with anti-interference; A three-dimensional measurement module: comprising a space mapping unit: mapping the edge coordinates output by the layered reconstruction module to a three-dimensional space to reconstruct the actual hole depth ; A row screening unit: calculating the effective measurement row range according to the actual hole depth Excluding the aperture / hole bottom distortion area to realize space mapping and diameter calculation; A diameter calculation unit: reading coordinate values row by row and column by column within the effective row, and calculating the vertical diameter and the horizontal diameter in combination with the optical scale; A decision output module: comprising: A confidence calculation unit: calculating edge consistency; A thermal compensation unit: inverting the thermal expansion coefficient according to the temperature sensor data and outputting the compensated diameter; A roundness calculation unit: calculating the micro-hole roundness in combination with the horizontal / vertical compensated diameter; A decision execution unit: judging whether to activate the contact probe for fusion re-inspection according to the calculation results to realize result reliability guarantee.

[0013] Compared with the prior art, the beneficial effects of the present application are: One, the present application adjusts the laser incidence angle in real time through polarization feedback (dynamic optical control module), and the curvature radius parameter of the mold is combined to close loop compensate the reflection distortion, so that the measurement error of the curved micro-hole diameter is reduced from ±8um to ±1.5um, and the self-adaptive precise detection of the curved surface is realized.

[0014] Two, the present application adopts a three-level processing chain (layered reconstruction module) of wavelet packet-U-Net-light ray tracing, and can still accurately segment the entity edge under the scene that the oil coverage reaches 30%, and the false detection rate is reduced from 15% to 3.2%, and the anti-interference reconstruction robustness is improved.

[0015] Three, the present application is based on virtual-real edge deviation quantitative confidence (decision output module), when the confidence is insufficient, the contact type reinspection is automatically switched, so that the roundness misjudgment rate under high temperature working condition is reduced by 18 percentage points, and a reliable decision mechanism is constructed.

[0016] Four, the initial layer spacing is calculated according to the self-adaptive calculation of hole depth (parameter initialization module), and the effective row intelligent truncation (three-dimensional measurement module) is combined, so that the scanning time is shortened by 47% compared with the traditional full hole detection, and the detection efficiency is optimized. DETAILED DESCRIPTION

[0017] Fig. 1 The flowchart of the present application is shown in the figure; Fig. 2 The system framework diagram of the present application is shown in the figure. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor belong to the scope of protection of the present application.

[0019] Please refer to Figs. 1-2 , the present application provides a technical scheme: a precision mold micro-hole size feature measurement method, comprising the following steps: S1, parameter initialization: reading the mold RFID tag to obtain the nominal hole depth H and the curvature radius R, and calculating the initial layer spacing dz0=0.1H; S2, dynamic optical scanning: irradiating the micro-hole surface with a radial polarized ring light source, and detecting the polarization angle deviation in real time through a polarization sensor ; According to the formula , the axial laser incidence angle is dynamically adjusted ; According to the updated , the Z-axis pulse scanning is performed, and the single-layer exposure time is wherein, is the Z-axis velocity, and DOF is the optical depth of field; S3, layered image reconstruction: wavelet packet denoising is performed on the S2 output image, and entity edge coordinates are segmented by U-Net; based on the entity edge coordinates, a virtual edge is generated by a ray tracing algorithm; Hausdorff distance of the entity and the virtual edge is calculated ; S4, three-dimensional measurement: mapping the virtual edge coordinates output by S3 to three-dimensional space to reconstruct the actual hole depth updating the effective measurement row range ; S5, diameter calculation and verification: calculating the horizontal diameter and the vertical diameter row by row within the effective row of S4 ; ; when >10%, the abnormal row is rejected; based on the edge consistency confidence calculated in S3 = ; S6, result output: when 0.95, the thermal compensation roundness is output wherein, the roundness is calculated; when 0.95, the contact probe reinspection is activated and the fusion measurement value is output.

[0020] The S2 interlayer spacing is dynamically adjusted: the subsequent interlayer spacing dz = min(0.05 , 1 μm).

[0021] The coefficient k of the S2 is obtained: when the mold is replaced, the R value is updated from the new mold RFID tag by a wireless reader and k is recalculated.

[0022] The S3 virtual edge generation: Monte Carlo sampling points 2 ; wherein, is the entity aperture output in S3.

[0023] The loss function of the S3 ; wherein, is set by the material refractive index n.

[0024] The thermal expansion coefficient α of the S6: ; n is the refractive index of the mold material, which is read by an RFID tag.

[0025] The roundness calculation condition of S6: Need to meet at the same time ≥0.95 and ≤2μm, otherwise jump to S6.

[0026] The control system of the method, characterized in that it comprises: Parameter initialization module: obtain the pre-stored nominal hole depth H and curvature radius R in the mold tag through the RFID reader; Calculate the initial scanning layer spacing dz0=0.1H, which provides a reference parameter for optical scanning; Store and output R to the dynamic optical control module to realize the closed-loop starting point of curvature compensation; Dynamic optical control module: including polarization monitoring unit, light source control unit and pulse scanning unit, for eliminating the distortion of curved surface reflection; Layered reconstruction module: including image preprocessing unit: wavelet packet denoising is performed on the input scanning image to remove environmental interference noise points; Entity segmentation unit: extract the micro-hole entity edge coordinates through the U-Net neural network, and output the binary entity edge image; Virtual generation unit: based on the entity edge coordinates, the theoretical virtual edge is generated by using the ray tracing algorithm; Deviation analysis unit: calculate the Hausdorff distance between the entity and the virtual edge , output the calibrated edge coordinates and to the downstream module to reconstruct the real edge with anti-interference; Three-dimensional measurement module: including space mapping unit: map the edge coordinates output by the layered reconstruction module to three-dimensional space, and reconstruct the actual hole depth ; Line screening unit: calculate the effective measurement line range according to the actual hole depth , exclude the aperture / hole bottom distortion area, realize space mapping and diameter calculation; Diameter calculation unit: read the coordinate values row by row and column by column within the effective line, and calculate the vertical diameter and horizontal diameter combined with the optical scale; Decision output module: including: Confidence calculation unit: calculate the edge consistency; Thermal compensation unit: according to the temperature sensor data, the thermal expansion coefficient is inverted, and the compensated diameter is output; Roundness calculation unit: combined with the horizontal / vertical compensated diameter, the micro-hole roundness is calculated; Decision execution unit, according to the calculation result, judge whether to activate the contact probe for fusion re-inspection, realize the result reliability guarantee.

[0027] It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting; it is not intended to be used to limit the scope of the present application. Unless defined otherwise, all technical and scientific terms used herein have the same meanings as commonly understood by one of ordinary skill in the art to which this application belongs. The singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise.

Claims

1. A method for measuring the dimensional characteristics of microholes in a precision mold, characterized in that, Includes the following steps: S1. Parameter initialization: Read the RFID tag of the mold to obtain the nominal hole depth H and radius of curvature R, and calculate the initial interlayer spacing dz0=0.1H; S2. Dynamic optical scanning: The surface of the micropores is illuminated by a radially polarized ring light source, and the polarization angle deviation is detected in real time by a polarization sensor. ; According to the formula Dynamically adjust the axial laser incident angle ; After the update Perform Z-axis pulse scan, single-layer exposure time ,in, Z-axis velocity, DOF optical depth of field; S3. Layered Image Reconstruction: Perform wavelet packet denoising on the output image of S2, segment entity edge coordinates using U-Net; generate virtual edges based on entity edge coordinates using a ray tracing algorithm; calculate the Hausdorff distance between entities and virtual edges. ; S4. 3D Measurement: Map the virtual edge coordinates output by S3 to 3D space to reconstruct the actual hole depth. Update valid measurement row range ; S5. Diameter Calculation and Verification: Calculate the horizontal diameter row by row within the valid rows of S4. and vertical diameter ; when Remove abnormal rows when the percentage is >10%; Based on S3 Calculate edge consistency confidence = ; S6. Output Results: When At 0.95, the output thermal compensation roundness is... ,in, Calculate roundness; when At 0.95, the contact probe is activated for re-inspection and the fusion measurement value is output.

2. The method for measuring the micro-hole size characteristics of a precision mold according to claim 1, characterized in that: The interlayer spacing in S2 is dynamically adjusted: Subsequent interlayer spacing dz=min(0.05) (1μm).

3. The method for measuring the micro-hole size characteristics of a precision mold according to claim 1, characterized in that: The coefficient k of S2 is obtained by updating the R value from the RFID tag of the new mold and recalculating k through a wireless reader when the mold is changed.

4. The method for measuring the micro-hole size characteristics of a precision mold according to claim 1, characterized in that: The S3 virtual edge generation: Monte Carlo sampling points 2 ;in, This refers to the solid aperture output in S3.

5. The method for measuring the micro-hole size characteristics of a precision mold according to claim 1, characterized in that: The loss function of S3 ; in, ∈[8,12] is set by the material's refractive index n.

6. The method for measuring the micro-hole size characteristics of a precision mold according to claim 1, characterized in that: The coefficient of thermal expansion α of S6: n is the refractive index of the mold material, which is read by an RFID tag.

7. The method for measuring the micro-hole size characteristics of a precision mold according to claim 1, characterized in that: The roundness calculation conditions for S6 are as follows: Must meet simultaneously ≥0.95 and ≤2μm, otherwise jump to S6.

8. A control system for implementing the method according to any one of claims 1 to 7, characterized in that: include: Parameter initialization module: Obtains the nominal hole depth H and radius of curvature R pre-stored in the mold tag through an RFID reader; The initial interlayer spacing dz0 = 0.1H is calculated to provide a reference parameter for optical scanning; Store and output R to the dynamic optical control module to realize the closed-loop starting point of curvature compensation; Dynamic optical control module: includes polarization monitoring unit, light source control unit and pulse scanning unit, used to eliminate surface reflection distortion; Layered reconstruction module: includes image preprocessing unit: performs wavelet packet denoising on input scanned image to remove environmental interference noise; Entity segmentation unit: Extracts the coordinates of the micropore entity edges using a U-Net neural network and outputs a binarized entity edge image; Virtual generation unit: Based on the coordinates of the entity edge, a theoretical virtual edge is generated using a ray tracing algorithm; Deviation analysis unit: Calculates the Hausdorff distance between the entity and the virtual edge. Output calibrated edge coordinates and Downstream modules reconstruct realistic edges to resist interference; 3D Measurement Module: Includes a spatial mapping unit: maps the edge coordinates output by the layered reconstruction module to 3D space to reconstruct the actual borehole depth. , ; Row filtering unit: based on actual hole depth Calculate the effective measurement range, exclude the distortion area at the orifice / bottom, and realize spatial mapping and diameter calculation; Diameter calculation unit: Reads coordinate values ​​row by row and column by column within the effective row, and calculates the vertical and horizontal diameters in conjunction with the optical scale; Decision output module: includes: Confidence calculation unit: calculates edge consistency; Thermal compensation unit: Calculates the coefficient of thermal expansion based on temperature sensor data and outputs the compensated diameter; Roundness calculation unit: Calculates the roundness of micro-holes by combining horizontal / vertical compensation diameter; The decision-making and execution unit determines whether to activate the contact probe for fusion re-inspection based on the calculation results, thereby ensuring the reliability of the results.

Images