Aluminum alloy die casting size on-line detection method and system

Abstract

The application provides an aluminum alloy die casting size online detection method and system, aiming at solving the problems of low detection precision, difficult to distinguish the deviation source and unable to close-loop control under high temperature working condition. The method comprises the following steps: constructing a multi-view image acquisition system and completing global calibration and thermal drift compensation; synchronously acquiring high-temperature die casting structured light images and reconstructing three-dimensional point clouds; registering the point clouds with the room temperature standard model, and constructing the thermal expansion deformation field through the point-by-point normal distance calculation; using non-contact temperature measurement combined with thermocouple calibration to obtain the full-field temperature, and converting the room temperature equivalent size through the point-by-point temperature field driven thermal shrinkage compensation algorithm; combining time sequence analysis and spatial distribution characteristics to determine the deviation source, and outputting the correction amount to realize close-loop control. The system comprises acquisition, processing, control and calibration units. The application realizes high-temperature non-contact high-precision online detection, accurately distinguishes mold wear and process fluctuation, improves detection precision and production efficiency, and is suitable for automatic die casting production line.

Description

Technical Field

[0001] This invention belongs to the field of aluminum alloy die casting quality inspection technology, specifically relating to an online detection method and system for the dimensions of aluminum alloy die castings. Background Technology

[0002] Aluminum alloy die castings offer advantages such as lightweight, high strength, and rapid forming efficiency, making them widely used in the automotive, communications, and electromechanical equipment industries. During the die casting production process, the die-cast parts are typically at a high temperature of 350℃–400℃ after demolding. Direct dimensional inspection at this temperature can lead to significant dimensional deviations due to thermal expansion, resulting in distorted inspection results. Conversely, waiting for the die-cast parts to cool to room temperature before inspection makes real-time monitoring and closed-loop process control impossible, easily resulting in batches of defective products and increasing production costs and material waste.

[0003] Traditional dimensional inspection methods often employ offline contact measurement or manual sampling, which suffer from low inspection efficiency, easy scratching of workpiece surfaces, and inability to reflect the true state of high-temperature forming. Some existing visual inspection technologies do not fully consider the impact of material thermal expansion deformation, equipment thermal drift, and strong self-luminescence interference on measurement results under high-temperature environments, making it difficult to accurately obtain the room temperature equivalent dimensions of die-cast parts. They also cannot effectively distinguish between mold wear and process fluctuations based on dimensional deviations, making it difficult to achieve accurate correction of the die-casting process. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide an online detection method and system for the dimensions of aluminum alloy die castings, so as to solve the problems of high temperature imaging interference, uneven temperature field, system thermal drift, insufficient compensation accuracy, inability to intelligently distinguish the source of deviation and achieve closed-loop control.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A method for online dimensional inspection of aluminum alloy die-cast parts includes the following steps: S1: Construct a multi-view image acquisition system, complete global calibration and online thermal drift compensation, and establish a unified measurement coordinate system; S2: Receives the trigger signal from the die-casting machine, synchronously acquires the structured light image of the high-temperature die-casting part, obtains the temperature distribution data on the surface of the die-casting part, and reconstructs the high-temperature three-dimensional morphology data. S3: Register the high-temperature three-dimensional point cloud with the room temperature standard model. By traversing all the measurement points in the high-temperature point cloud, search for the nearest corresponding point and unit normal vector on the standard model. Calculate the normal distance point by point using the normal projection formula from point to plane, and construct a full-area thermal expansion deformation field covering the surface of the die casting. S4: Based on the temperature distribution data and the thermophysical properties of the aluminum alloy, calculate the difference between the temperature at each measuring point and the room temperature and the corresponding thermal shrinkage compensation amount, and determine the deformation direction based on the normal vector of the measuring point. Correct the three-dimensional coordinates point by point along the surface normal, and convert the high temperature dimension into the room temperature equivalent dimension in reverse. S5: Compare the room temperature equivalent size with the standard size, analyze the source of deviation and output the die-casting process correction amount to achieve closed-loop control; In step S1, the multi-view image acquisition system consists of a structured light projector and an industrial camera, the industrial camera being equipped with a narrowband filter; the system includes a high-temperature resistant reference block, supporting online automatic recalibration to compensate for equipment thermal drift.

[0006] Furthermore, the structured light projector is a laser with a wavelength of 532nm.

[0007] Furthermore, in step S2, the system is hardwired to the die-casting machine PLC and hardware synchronous triggering is used; high-temperature self-emission is suppressed by short exposure and adaptive grating brightness adjustment, and high-temperature three-dimensional point cloud is obtained by phase calculation.

[0008] Furthermore, in step S4, an infrared thermal imager is used to acquire the temperature distribution across the entire field and calibrate it using thermocouples; based on the point-by-point temperature and thermal expansion coefficient, thermal expansion compensation is performed point by point to obtain the equivalent size at room temperature.

[0009] Furthermore, in step S5, the sources of deviation are distinguished based on the spatial distribution and temporal trend of the deviation: uniform and slow deviation is determined to be mold wear, and random fluctuation deviation is determined to be process fluctuation; mold compensation parameters or die-casting process adjustment parameters are output respectively.

[0010] An online dimension detection system for aluminum alloy die castings, used to implement the method, includes: a multi-view image acquisition unit for synchronous acquisition of the three-dimensional morphology and surface temperature of the high-temperature die casting; a data processing unit for three-dimensional reconstruction, registration, thermal expansion compensation, and deviation analysis; a feedback control unit for communicating with the die casting machine and issuing correction commands; and an online calibration unit for online compensation of system thermal drift.

[0011] Compared with the prior art, the present invention has the following beneficial effects: This invention achieves non-contact, high-precision, online dimensional inspection of aluminum alloy die-casting parts under high-temperature conditions through global calibration and online thermal drift compensation of a multi-view image acquisition system, synchronous hardware 3D topography acquisition under high-temperature conditions, registration of high-temperature 3D point clouds with room-temperature standard models and calculation of thermal expansion deformation fields, inverse mapping conversion of high-temperature dimensions to room-temperature equivalent dimensions based on the overall temperature distribution, analysis of spatial distribution and temporal variation characteristics of dimensional deviations, intelligent decomposition of deviation sources, and real-time output of die-casting process corrections. It effectively eliminates measurement errors caused by thermal expansion and uneven temperature fields, and can automatically distinguish between mold wear and process fluctuations based on the inspection results, accurately outputting mold compensation or die-casting process adjustment parameters. It solves the problems of lagging, insufficient accuracy, susceptibility to high-temperature interference, and inability to achieve closed-loop control of the production process in traditional inspection methods, significantly improving the dimensional qualification rate, inspection stability, and overall production efficiency of die-casting parts. Attached Figure Description

[0012] Figure 1 This is a schematic diagram of the process flow of this processing method. Detailed Implementation

[0013] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and should not be construed as limiting the scope of the invention.

[0014] Example 1 like Figure 1 As shown, an online dimension detection method for aluminum alloy die castings is applicable to high-temperature online dimension detection in aluminum alloy die casting production lines. The demolding temperature of the workpiece is 350℃~400℃. The system is deployed at the mold opening and part removal station and can achieve non-contact, high-precision, fully automatic dimension detection and process closed-loop control in high-temperature, dusty, and vibration environments.

[0015] Specifically, the steps include the following: Step S1: Construct a multi-view image acquisition system, complete global calibration and online thermal drift compensation, and establish a unified measurement coordinate system. Specifically, firstly, a multi-view image acquisition system is deployed around the die-casting machine's part-removal station. The system includes multiple sets of structured light projectors with specific wavelengths and industrial cameras. Narrow-band filters matching the projection wavelength are installed at the front end of each camera to filter out self-radiation from high-temperature workpieces and interference from ambient stray light. After the hardware installation is complete, global calibration is performed. This involves sequentially using the Zhang Zhengyou calibration method to calibrate the camera's intrinsic parameters, using a distortion correction model to compensate for lens distortion, and using a hand-eye calibration algorithm to calibrate the extrinsic parameters, thus unifying all cameras and projectors to the same world coordinate system. The preferred wavelength of the structured light projector is 532nm.

[0016] To address the issue of decreased accuracy due to thermal drift in high-temperature environments, this solution incorporates a high-temperature resistant reference block within the testing area and supports an online automatic recalibration process. This allows for automatic compensation of thermal drift errors during production, ensuring long-term measurement stability. By using a fixed reference block, this solution enables dynamic calibration of the measurement reference under high-temperature conditions. Unlike traditional single-calibration methods, this significantly improves the long-term reliability of the system under harsh conditions.

[0017] Step S2: Receive the trigger signal from the die-casting machine, synchronously acquire structured light images of the high-temperature die-cast parts, and reconstruct the high-temperature three-dimensional morphology data. Specifically, the system is connected to the die-casting machine's PLC via a hardwired connection, receives the mold opening or part removal trigger signal in real time, and immediately outputs multiple synchronous pulses upon receiving the signal to control all structured light projectors and industrial cameras to work synchronously, avoiding asynchronous acquisition errors caused by workpiece movement or rapid temperature drop.

[0018] The projector projects an coded grating, and the camera acquires stripe images in short-exposure mode. The grating brightness is adaptively adjusted based on the workpiece temperature to suppress interference from high-temperature strong reflections. After image acquisition, the system uses a three-step phase-shifting method for phase calculation. A quality-guided phase unwrapping algorithm is used to obtain a continuous phase distribution. Then, combined with calibration parameters, the three-dimensional coordinates are calculated using triangulation principles. Finally, high-temperature three-dimensional point cloud data is generated through point cloud stitching, filtering, and redundancy removal.

[0019] Step S3: Register the high-temperature 3D point cloud with the room-temperature standard model and calculate the thermal expansion deformation field of the entire region. Specifically, the high-temperature 3D point cloud and the preset room-temperature standard CAD model are preprocessed, and noise and invalid points are removed by passing through filtering. Then, the Iterative Closest Point (ICP) algorithm is used for rigid body registration to achieve spatial alignment between the measured point cloud and the standard model.

[0020] After registration, a thermal expansion deformation field is constructed by calculating the point-by-point normal distance. This process is a key innovative step in achieving accurate dimensional compensation at high temperatures. The specific implementation method is as follows: First, the registered high-temperature 3D point cloud data and the room-temperature standard model data are subjected to spatial position normalization processing to ensure that the two sets of data are in the same coordinate system and their orientations are perfectly aligned. For each data point in the high-temperature 3D point cloud... Search for the closest Euclidean distance on the standard model of room temperature. and obtain the standard model in Unit normal vector at point .

[0021] With unit normal vector Based on this, calculate the high temperature point. Relative to standard model points The normal projection distance d is the amount of thermal expansion deformation of the current point under high temperature. It is calculated using the formula for the normal projection from a point to a plane, specifically:

[0022] In the formula, "·" represents vector dot product, d is positive indicating expansion at that point, and d is negative indicating contraction. By traversing all measuring points in the high-temperature point cloud and performing the above normal distance calculation point by point, the set of normal deformation amounts covering the entire surface of the die-casting can be obtained. By associating and mapping the deformation amount of each point with its spatial coordinates, a global thermal expansion deformation field can be formed. This thermal expansion deformation field can accurately reflect the non-uniform expansion characteristics of the die-casting under high-temperature conditions caused by uneven temperature distribution, wall thickness differences, and different heat dissipation rates. Unlike the traditional method of using a single average expansion coefficient for overall compensation, this scheme can accurately characterize the actual expansion magnitude and direction at each location through point-by-point normal distance calculation. This provides a reliable basis for the accurate conversion of high-temperature dimensions to room-temperature equivalent dimensions, significantly improving the accuracy and reliability of the final test results.

[0023] Step S4: Obtain the overall temperature distribution of the die casting. Based on the thermophysical properties of aluminum alloy, the high-temperature dimensions are converted to equivalent dimensions at room temperature. In this step, while acquiring the three-dimensional morphology, the system uses a non-contact temperature measurement method to obtain the overall temperature distribution of the die casting. Temperature calibration is performed using data collected by the thermocouples built into the mold to improve the accuracy and reliability of the temperature field.

[0024] After obtaining the overall temperature distribution data, the system performs real-time comparison and calibration between the infrared temperature measurement results and the temperature data collected by the thermocouples built into the mold to eliminate temperature measurement errors caused by factors such as the oxide layer on the casting surface, residual release agent, and environmental radiation. The specific calibration method is as follows: using the steady-state temperature collected by the thermocouple at the corresponding position in the mold cavity as the reference temperature, the difference between the infrared temperature measurement point and the reference temperature is calculated, and the overall temperature field is corrected point by point to make the temperature value at each measurement point closer to the actual temperature of the casting. The calibration formula is as follows:

[0025] In the formula: This is the calibrated temperature value of the i-th measuring point; The raw temperature value collected by the infrared device; The reference temperature is measured by the thermocouple built into the mold. This is the reference temperature obtained by infrared sampling at the corresponding location of the thermocouple.

[0026] Through the above calibration, the accuracy of the overall temperature data can be controlled within ±1℃, providing a reliable input for the subsequent point-by-point temperature field driven thermal shrinkage compensation algorithm.

[0027] Meanwhile, this step uses a point-by-point temperature field driven thermal shrinkage compensation algorithm to inversely convert the high-temperature dimensions into equivalent dimensions at room temperature. Specifically, the point-by-point temperature field driven thermal shrinkage compensation algorithm uses the actual temperature of each point in the three-dimensional point cloud as the driving quantity, and combines the material's thermal expansion coefficient to achieve accurate restoration of the high-temperature point cloud to the equivalent dimensions at room temperature.

[0028] First, calculate the temperature difference between each measuring point and the standard room temperature. The formula for calculating the temperature difference is:

[0029] in, Let i be the measured temperature at the i-th measuring point. This is the standard room temperature.

[0030] The thermal contraction compensation is calculated point by point according to the theory of thermal expansion. The formula for calculating the compensation is as follows:

[0031] Where α is the linear thermal expansion coefficient of aluminum alloy. This represents the thermal shrinkage compensation amount for the i-th measuring point.

[0032] The deformation direction is determined based on the normal vector of the measuring point, and the three-dimensional coordinates are corrected point by point along the surface normal. The coordinate correction formula is:

[0033] in, The three-dimensional coordinates of the i-th measuring point before correction. Let be the unit normal vector corresponding to the i-th measurement point. The corrected room temperature equivalent three-dimensional coordinates are shown. The dimensional deviations caused by high-temperature expansion are eliminated sequentially using the above formulas. After traversing all measurement points, a global thermal shrinkage compensation point cloud is formed, making the high-temperature measurement data equivalent to the true dimensions under standard room temperature conditions, providing a unified benchmark for subsequent dimensional comparison, deviation analysis, and quality judgment. This algorithm adopts a point-by-point independent calculation method, which can adapt to the actual working conditions of complex curved surfaces, uneven wall thickness, and uneven temperature distribution in castings, and has higher accuracy and applicability compared to traditional overall compensation methods.

[0034] Step S5: After obtaining the room temperature equivalent dimensions and completing the deviation calculation, this solution uses a combination of time series analysis and spatial distribution characteristics to intelligently determine the source of deviation. It can automatically distinguish whether the dimensional deviation is caused by mold wear or process fluctuation, providing a basis for subsequent accurate output of compensation parameters.

[0035] Specifically, in terms of time series analysis, the system continuously collects and stores dimensional deviation data of multiple production runs, forming a time-series variation curve with the production run as the horizontal axis and the critical dimensional deviation value as the vertical axis. The system performs trend fitting and stationarity analysis on the time-series curve, calculates the magnitude, direction, and rate of change of deviation between consecutive runs, and determines whether the deviation exhibits a continuous increasing, continuous decreasing, or random fluctuation characteristic through time-series differencing. The calculation of the deviation increment between adjacent runs uses the following formula:

[0036] In the formula: This represents the deviation increment between the k-th module and the (k-1)-th module (k is the current production module, k≥2); This represents the critical dimension deviation value for the k-th module; This represents the critical dimensional deviation value for the (k-1)th mold. This applies to 5-10 consecutive mold runs. Satisfy 0 < If the deviation increment is ≤ 0.002mm / mold, that is, when the deviation increment is stable, without drastic jumps and shows a unidirectional increasing trend, it indicates that the deviation is caused by continuous wear of the mold cavity.

[0037] Furthermore, the time-series trend recognition model is extended: in addition to recognizing continuous, slowly increasing / decreasing trends, the system is also equipped with a sudden change detection algorithm: if the deviation increment between adjacent modules... If the damage exceeds a preset threshold (e.g., >0.05mm) and exhibits localized spatial characteristics, it is determined to be sudden mold damage (e.g., cracking or chipping). If the time-series curve shows periodic fluctuations (e.g., a sinusoidal waveform with each production cycle), it is determined to be periodic interference from the part removal or painting process. The system matches the corresponding correction strategy library based on the identified trend type.

[0038] In terms of spatial distribution characteristic analysis, the system maps deviation values ​​onto the surface of the die-cast part's 3D model, forming a deviation spatial distribution cloud map. This identifies the location, range, concentration, geometric shape of the deviation areas, and their correspondence with the mold forming surface. The concentration of the deviation areas is calculated using the following formula:

[0039] In the formula: C is the concentration of the deviation area (range 0~1); Nc is the number of deviation measurement points concentrated in the corresponding area of ​​the mold forming surface; Ntotal is the total number of all deviation measurement points. If the deviation area is highly concentrated, fixed in position, and completely corresponds to the contact surface of the mold cavity, and the calculated concentration C ≥ 0.8 (i.e., more than 80% of the deviation measurement points are concentrated on the mold forming surface), exhibiting a uniform deviation characteristic across the entire region, then the deviation originates from mold wear. If the deviation area is scattered, has no fixed position, appears randomly across regions, has no fixed correspondence with the mold cavity, and the concentration C < 0.5, accompanied by local abrupt changes and irregular distribution characteristics, then the deviation originates from fluctuations in die casting process parameters.

[0040] Furthermore, deviation morphology analysis is introduced to assist in the judgment: the system pre-stores CAD structural feature data of the mold cavity, including the parting surface position, slider position, insert position, and ejector pin distribution. The calculated deviation spatial distribution cloud map is overlaid and compared with the above mold structural features: if the geometric shape of the deviation area (such as strip, ring, or specific contour) coincides with the parting surface, slider mating surface, or insert edge height of the mold, it is strongly judged as mold wear or change in mating clearance; If the deviation area exhibits thermal features (such as local depressions or protrusions) consistent with the cooling water channel layout or casting wall thickness distribution, the deformation is determined to be caused by process parameters (such as insufficient cooling efficiency or insufficient holding pressure). By combining morphological matching degree, the determination result that relies solely on the numerical threshold C is corrected.

[0041] A combined judgment is made based on comprehensive time series analysis and spatial distribution characteristics. When the deviation simultaneously satisfies the condition of continuous and slow increase in time series ( When the deviation is stable at 0~0.002mm / mold and spatially concentrated in the mold forming area (C≥0.8), the source of the deviation is determined to be mold wear; when the deviation exhibits irregular and random changes in time sequence ( When the fluctuations are severe and have no fixed direction, and are spatially dispersed without a fixed location (C<0.5), the source of the deviation is determined to be process fluctuation.

[0042] Furthermore, for intermediate states that cannot be clearly categorized: when the calculation results do not meet any of the aforementioned clear judgment conditions for mold wear or process fluctuations (e.g., spatial concentration 0.5 ≤ C < 0.8, or the time series trend is not obvious): If the system determines that the deviation is 'compound deviation' or 'unknown anomaly', it will not directly output mold compensation or process correction instructions. Instead, it will trigger an alarm signal, push the deviation data and 3D cloud map to the human-machine interface, and request manual intervention for analysis. At the same time, the system will automatically record the data of this mold for subsequent big data analysis and model iteration.

[0043] Based on the above judgment results, the system can automatically output differentiated correction strategies: for mold wear, output the mold compensation amount for the corresponding area; for process fluctuations, output the adjustment amount of process parameters such as injection speed, holding pressure, and cooling time, thereby realizing closed-loop control of detection-analysis-correction.

[0044] Example 2 A high-temperature online dimensional inspection system for aluminum alloy die castings, used to implement the method, includes: a multi-view structured light acquisition unit, a temperature synchronous acquisition unit, a data processing unit, a thermal drift online calibration unit, and a process closed-loop control unit; the units are electrically interconnected through industrial Ethernet and high-speed signal trigger lines, and are deployed as a whole at the die casting machine's mold opening and part removal station, and can operate continuously and stably in industrial environments with high temperatures of 350℃-400℃, dust, and vibration.

[0045] The multi-view structured light acquisition unit consists of at least two sets of structured light projectors and two industrial area array cameras. The front end of the camera lens is equipped with a narrow-band filter that is strictly matched with the wavelength of the projected light to filter out the thermal radiation of the high-temperature casting itself and the interference of stray light on site. The camera and projector are both installed on both sides of the part picking robot through high-precision fixed brackets to ensure that the acquisition view covers the entire formed surface of the casting and realizes blind-spot-free three-dimensional shape reconstruction.

[0046] The temperature synchronization acquisition unit includes a non-contact infrared thermometer and at least two sets of built-in thermocouples. The infrared thermometer is used to acquire real-time temperature distribution data of the entire casting. The thermocouples are fixed on the surface of the mold cavity and used to perform online calibration of the temperature field to ensure that the temperature data and the three-dimensional morphology data are strictly synchronized in time and space, providing a unified benchmark for subsequent thermal expansion compensation.

[0047] The data processing unit is an embedded industrial control computer equipped with high-performance image processing and point cloud calculation software. It is used to execute all algorithm processes, including 3D point cloud reconstruction, ICP registration, thermal expansion deformation field calculation, point-by-point temperature field compensation, room temperature equivalent size conversion, size deviation determination and deviation source differentiation.

[0048] The online thermal drift calibration unit includes a high-temperature resistant ceramic reference block fixedly installed in the acquisition field of view. Its deformation is less than 0.005 mm under high temperature environment, which can serve as a long-term stable measurement reference. The system automatically performs three-dimensional acquisition and calibration calculation on the reference block at set intervals, and corrects the measurement drift caused by the thermal deformation of the camera and projector under high temperature conditions in real time, so as to ensure the long-term operating accuracy of the system.

[0049] The process closed-loop control unit is a PLC control module with industrial bus communication function. It can achieve bidirectional data interaction with the die-casting machine host, robot host, and spraying machine host through PROFINET or EtherNet / IP protocol. When the system determines that the deviation is caused by mold wear, it outputs a mold compensation signal. When the system determines that the deviation is caused by process fluctuation, it outputs process parameter adjustment signals such as injection speed, boost pressure, cooling time, and spraying volume, so as to realize fully automatic closed-loop control of detection-analysis-adjustment.

[0050] The embodiments of the present invention are given for illustrative and descriptive purposes only, and are not intended to be exhaustive or to limit the invention to the forms disclosed. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described in order to better illustrate the principles and practical application of the invention, and to enable those skilled in the art to understand the invention and to design various embodiments with various modifications suitable for a particular purpose.

Claims

1. A method for online dimensional inspection of aluminum alloy die-casting parts, characterized in that, Includes the following steps: S1: Construct a multi-view image acquisition system, complete global calibration and online thermal drift compensation, and establish a unified measurement coordinate system; S2: Receives the trigger signal from the die-casting machine, synchronously acquires the structured light image of the high-temperature die-casting part, obtains the temperature distribution data on the surface of the die-casting part, and reconstructs the high-temperature three-dimensional morphology data. S3: Register the high-temperature three-dimensional point cloud with the room-temperature standard model and calculate the thermal expansion deformation field of the entire region; S4: Based on the temperature distribution data and the thermophysical properties of the aluminum alloy, the high-temperature dimensions are converted into equivalent dimensions at room temperature. S5: Compare the room temperature equivalent size with the standard size, analyze the source of deviation and output the die-casting process correction amount to achieve closed-loop control; In step S3, after registration is completed, the registered high-temperature three-dimensional point cloud data and the room temperature standard model data are subjected to spatial position normalization processing so that the two sets of data are in the same coordinate system and their attitudes are completely aligned. For each data point in the high-temperature 3D point cloud Search for the closest Euclidean distance on the standard model of room temperature. and obtain the standard model in Unit normal vector at point ; With unit normal vector Based on this, calculate the high temperature point. Relative to standard model points The normal projection distance d is the amount of thermal expansion deformation of the current point under high temperature. It is calculated using the normal projection formula from a point to a plane, specifically: In the formula, "·" represents the vector dot product operation, a positive d indicates that the point expands, and a negative d indicates that the point contracts. In step S4, the reverse conversion of high-temperature dimensions to room-temperature equivalent dimensions is as follows: using the steady-state temperature value collected by the thermocouple at the corresponding position in the mold cavity as the reference temperature, the difference between the infrared temperature measurement point and the reference temperature is calculated, and the temperature field of the entire field is corrected point by point to make the temperature value of each measurement point closer to the actual temperature of the casting. The calibration formula is as follows: In the formula: This is the calibrated temperature value of the i-th measuring point; The raw temperature value collected by the infrared device; The reference temperature is measured by the thermocouple built into the mold. This is the reference temperature for infrared data collected at the corresponding location of the thermocouple; Through the above calibration, the accuracy of the temperature data across the entire field can be controlled within ±1℃, providing a reliable input for the subsequent point-by-point temperature field-driven thermal shrinkage compensation algorithm. Meanwhile, this step uses a point-by-point temperature field driven thermal shrinkage compensation algorithm to inversely convert the high-temperature dimensions into room-temperature equivalent dimensions. Specifically, the point-by-point temperature field driven thermal shrinkage compensation algorithm uses the actual temperature of each point in the three-dimensional point cloud as the driving quantity, and combines the material's thermal expansion coefficient to achieve accurate restoration of the high-temperature point cloud to the room-temperature equivalent dimensions. First, calculate the temperature difference between each measuring point and the standard room temperature. The formula for calculating the temperature difference is: in, Let i be the measured temperature at the i-th measuring point. Standard room temperature; The thermal contraction compensation is calculated point by point according to the theory of thermal expansion. The formula for calculating the compensation is as follows: Where α is the linear thermal expansion coefficient of aluminum alloy. This represents the thermal shrinkage compensation amount for the i-th measuring point; The deformation direction is determined based on the normal vector of the measuring point, and the three-dimensional coordinates are corrected point by point along the surface normal. The coordinate correction formula is: in, The three-dimensional coordinates of the i-th measuring point before correction. Let be the unit normal vector corresponding to the i-th measurement point. These are the corrected room temperature equivalent three-dimensional coordinates.

2. The method for online dimensional inspection of aluminum alloy die-casting parts according to claim 1, characterized in that, In step S1, the multi-view image acquisition system consists of a structured light projector and an industrial camera, the industrial camera being equipped with a narrowband filter; the system includes a high-temperature resistant reference block, supporting online automatic recalibration to compensate for equipment thermal drift.

3. The method for online dimensional inspection of aluminum alloy die-casting parts according to claim 2, characterized in that, The structured light projector is a laser with a wavelength of 532nm.

4. The method for online dimensional inspection of aluminum alloy die-casting parts according to claim 1, characterized in that, In step S2, the system is hardwired to the die-casting machine PLC and hardware synchronous triggering is used; high-temperature self-emission is suppressed by short exposure and adaptive grating brightness adjustment, and high-temperature three-dimensional point cloud is obtained by phase calculation.

5. The method for online dimensional inspection of aluminum alloy die-casting parts according to claim 1, characterized in that, In step S4, an infrared thermal imager is used to acquire the temperature distribution across the entire field and calibrate it with a thermocouple; based on the point-by-point temperature and thermal expansion coefficient, thermal expansion compensation is performed point by point to obtain the equivalent size at room temperature.

6. The method for online dimensional inspection of aluminum alloy die-casting parts according to claim 1, characterized in that, In step S5, the sources of deviation are distinguished according to the spatial distribution and temporal trend of the deviation: uniform and slow deviation is determined to be mold wear, and random fluctuation deviation is determined to be process fluctuation; mold compensation parameters or die casting process adjustment parameters are output respectively.

7. An online dimensional inspection system for aluminum alloy die-cast parts, used to implement the method described in any one of claims 1 to 6, characterized in that, include: A multi-view image acquisition unit is used for the simultaneous acquisition of the three-dimensional morphology and surface temperature of high-temperature die-cast parts; The data processing unit is used for 3D reconstruction, registration, thermal expansion compensation, and deviation analysis. The feedback control unit is used to communicate with the die-casting machine and issue correction commands. Online calibration unit for online compensation of system thermal drift.

Images