A method and system for measuring the cure temperature of an aluminum pigment coating

Abstract

This invention relates to the field of temperature measurement technology and discloses a method and system for measuring the curing temperature of aluminum pigment coatings. The method includes: S1, acquiring a radiation signal; S2, signal processing and temperature acquisition: processing the acquired radiation signal using at least one of the following two processing modes to obtain the true temperature of the coating: Mode A, calculating the true temperature of the coating by solving the polarization radiation transmission equations based on the radiation signal; Mode B, extracting high-dimensional transient radiation features from the radiation signal and calculating the true temperature of the coating by matching them with a pre-constructed transient micro-polarization fingerprint database; S3, temperature mapping distribution; S4, identifying and alerting to anomalies. This invention can solve the problems of infrared thermometry being easily interfered with by environmental reflections, difficulty in obtaining fine temperature distribution, and difficulty in adapting to frequent changes in coating or workpiece design during the curing of aluminum pigment coatings.

Description

Technical Field

[0001] This invention relates to the field of temperature measurement technology, and specifically to a method and system for measuring the curing temperature of aluminum pigment coatings. Background Technology

[0002] In the precision manufacturing process of high-end consumer electronics products, such as the metal casing production line for laptops or tablets, workpieces coated with aluminum pigment often require high-temperature curing to ensure the coating's adhesion, hardness, scratch resistance, and the final metallic luster and color uniformity. This curing process is usually completed in a curing oven, and the precision of its temperature control directly determines the quality and yield of the final product.

[0003] Existing non-contact infrared temperature measurement systems are prone to distortion due to the extremely high reflectivity of the aluminum pigment coating itself. While receiving the heat radiation from the coating itself, they are also susceptible to strong interference from the heating elements and reflected radiation from the furnace wall inside the curing oven.

[0004] The metal casings of modern consumer electronics are becoming increasingly thinner and lighter. Their low heat capacity results in extremely fast temperature responses. When ultra-thin metal casings pass through a curing oven at high speed, single-point infrared thermography at a fixed location can only capture instantaneous point temperatures, making it difficult to depict a precise temperature profile along the entire curing path. Furthermore, due to the thinness of the metal casing, its internal heterogeneous structures, such as battery compartments, antenna areas, or reinforcing ribs, significantly interfere with uniform heat transfer. Even when covered by a coating, this can create significant temperature distribution differences on the coating surface. Simultaneously, as product designs become increasingly complex, with diverse workpiece geometries and material distributions, these factors all affect the heat conduction and temperature distribution on the coating surface. Current technologies, such as multi-angle infrared sensor arrays and more complex signal processing methods, attempt to differentiate and compensate for environmental reflections by collecting radiation information from different angles. However, these methods still cannot penetrate the coating surface to sense and quantify the local thermal characteristics differences of the workpiece, thus failing to accurately reflect the true temperature distribution on the coating surface.

[0005] Even more challenging is the need for production lines to frequently switch between different colors, metallic effects, and coating formulations or workpiece designs to meet consumer demands for personalized appearances. Even the same product may have multiple surface treatments, such as matte metallic and glossy metallic. These different coating formulations, especially the particle size, shape, and arrangement of aluminum pigments, as well as the thickness and composition of the clear coat, directly affect the optical properties and effective emissivity of the coating. Each coating switch means that previously established emissivity correction methods or calibration parameters for that specific coating may no longer be applicable. If each switch requires time-consuming and complex offline recalibration, it will severely impact production efficiency and flexibility, which is impractical for the consumer electronics industry, which pursues rapid iteration and customized production. Summary of the Invention

[0006] This invention discloses a method and system for measuring the curing temperature of aluminum pigment coatings, aiming to solve problems in the prior art such as infrared temperature measurement distortion caused by the high reflectivity of aluminum pigment coatings, the influence of workpiece geometry and material diversity on temperature distribution, and the poor adaptability of traditional temperature measurement methods when production lines frequently switch coating formulations or workpiece designs.

[0007] To achieve the above objectives, the technical solution of the present invention is as follows: In a first aspect, the present invention discloses a method for measuring the curing temperature of an aluminum pigment coating, the method comprising the following steps: S1. Acquiring radiation signals: For a target workpiece that is in a high-speed motion state and has an aluminum pigment coating on its surface, a multispectral infrared polarization sensor array is used to synchronously acquire radiation signals from the surface of the target workpiece in different infrared spectral bands and different polarization directions. S2. Signal Processing and Temperature Acquisition: The acquired radiation signal is processed using at least one of the following two processing modes to obtain the true temperature of the coating: Mode A: Based on the radiation signal, the true temperature of the coating is calculated by solving the polarization radiation transmission equations. Mode B: Extract high-dimensional transient radiation features from the radiation signal and calculate the true temperature of the coating by matching them with a pre-built transient micro-polarization fingerprint database; S3. Temperature Mapping Distribution: All the acquired single-point temperature data, corresponding to the real-time motion information of the target workpiece, are reconstructed and mapped into a temperature distribution map covering the entire surface of the target workpiece through coordinate transformation and spatial interpolation. S4. Identify and alert to anomalies: Based on the preset temperature threshold range and uniformity standard threshold range, identify abnormal areas in the temperature distribution map and issue an early warning.

[0008] Furthermore, in the above method, step S2, which involves calculating the true temperature of the coating by solving the polarization radiation transport equations based on the radiation signal, includes: Before the target workpiece enters the measurement area, a radiation measurement is performed on a preset reference point to obtain the spectral radiance of the reference point. Based on the known emissivity and known true temperature of the reference point, the ambient spectral radiance is calculated by back-calculating using Planck's radiation law. Based on the collected radiation signals, a set of polarization radiation transmission equations based on Fresnel's equations is established, and the thermal emission radiation of the target workpiece itself and the reflected radiation from the environment are obtained by solving the equations. By pre-setting the emissivity ratio of the coating at two wavelengths, and based on the direct proportional relationship between the thermal emission radiation of the target workpiece itself and the Planck blackbody radiation function when environmental reflection radiation is ignored, the initial temperature estimate of the coating is obtained. Based on the functional relationship between the coating emissivity ratio and the actual temperature of the coating and the environmental spectral radiance at two wavelengths, an inversion equation set is established. Starting from the initial temperature estimate, the inversion equation set is solved by numerical iteration to obtain the actual temperature of the coating and the effective emissivity of the coating at two wavelengths.

[0009] More specifically, in some implementations, the step of establishing an inversion equation set based on the functional relationship between the coating emissivity ratio and the coating's true temperature and the ambient spectral radiance at two wavelengths includes: For the first wavelength Second wavelength ,satisfy: ; ; / =f( The functional relationship was obtained by conducting offline spectral radiometric calibration experiments on standard samples of the coating. In the formula, For coating at wavelength The measured spectral radiance, For coating at wavelength Effective emissivity at the following levels Let be the Planck blackbody radiation function. As a reference point at wavelength The measured environmental spectral radiance, This represents the actual temperature of the coating.

[0010] Preferably, by using the pre-set emissivity ratio of the coating at two wavelengths, and based on the direct proportional relationship between the thermal emission radiation of the target workpiece itself and the Planck blackbody radiation function when ignoring ambient reflected radiation, the initial temperature estimate of the coating is obtained. This includes: based on the thermal emission radiation of the coating itself at two wavelengths, and combined with the pre-set emissivity ratio of the coating at two wavelengths, the initial temperature estimate of the coating is obtained by using Wien's approximation law or a lookup table method.

[0011] Based on the above, this invention further proposes that the numerical iterative solution method adopts the Newton-Raphson method or the least squares method.

[0012] Furthermore, in the above method, step S2, which involves extracting high-dimensional transient radiation features from the radiation signal and calculating the true temperature of the coating by matching them with a pre-constructed transient micro-polarization fingerprint database, includes: The acquired radiation signals are preprocessed to obtain transient radiation intensity values. All acquired transient radiation intensity values ​​are combined in a preset order to form an initial feature vector. The dynamic change rate of the initial feature vector in the continuous time series is calculated to obtain dynamic change information. The dynamic change information is used as a new feature dimension to expand the initial feature vector and form a higher-dimensional transient micro-polarization feature vector. In a controlled environment, a reference plate with a standard coating having an equivalent structure to the coating on the target workpiece is heated at an ultra-high heating rate to simulate the actual production process. During the heating process, the following data are collected simultaneously: full polarization radiation data of the standard coating surface, the actual temperature data of the standard coating obtained by a contact sensor, and microstructure evolution image data of the standard coating obtained by a microscopic imaging system. Standard transient micro-polarization feature vectors are generated based on fully polarized radiation data, microstructure states are determined based on microstructure evolution image data, and standard transient micro-polarization feature vectors, real temperatures, and microstructure states at the same time stamp are correlated to form entries in the transient micro-polarization fingerprint database. The transient micro-polarization feature vectors acquired in real time are compared with the transient micro-polarization fingerprint database, and the true temperature of the current measurement point is calculated based on the comparison results.

[0013] Based on the above, this invention further proposes a step of comparing the transient micro-polarization feature vector acquired in real time with a transient micro-polarization fingerprint database, and calculating the true temperature of the current measurement point based on the comparison result, including: Calculate the cosine similarity or Euclidean distance between the transient micro-polarization feature vectors acquired in real time and the feature vectors stored in the transient micro-polarization fingerprint database; Based on the cosine similarity being higher than a preset threshold or the Euclidean distance being lower than a preset threshold, one or more entries matching the transient micro-polarization feature vectors acquired in real time are selected from the transient micro-polarization fingerprint database. The actual temperature is calculated by using a weighted average or interpolation method to calculate the temperature values ​​corresponding to all matching entries.

[0014] Furthermore, in the above method, the step of identifying abnormal areas in the temperature distribution map and issuing an early warning based on a preset temperature threshold range and a uniformity standard threshold range includes: Based on the temperature distribution map, the real-time temperature value of each coordinate point on the temperature distribution map is compared with the preset temperature threshold range for the corresponding coordinate point to identify abnormal coordinate points; Spatial clustering of abnormal coordinate points is performed to form one or more abnormal connected regions; For each abnormally connected region, the region's temperature statistical characteristics are calculated, and these characteristics are compared with preset temperature threshold ranges and uniformity standard threshold ranges. If the comparison result exceeds the threshold range, an early warning is triggered.

[0015] Based on the above, the present invention further proposes that the method also includes a pattern coordination and verification step, comprising: When obtaining the true temperature using mode A, a transient micro-polarization feature vector for verification is simultaneously generated based on the radiation signal. The transient micro-polarization feature vector used for verification is input into the transient micro-polarization fingerprint database corresponding to mode B for matching, and the temperature reference value and the corresponding confidence level are obtained. If the deviation between the actual temperature calculated by Mode A and the temperature reference value exceeds the preset tolerance and / or the confidence level is lower than the preset threshold, then the calculation result of Mode A is determined to be unreliable, and at least one of the following operations is triggered: enable the output result of Mode B as the final true temperature, trigger the remeasurement and calibration of the ambient spectral radiance, and output a warning message with a low confidence level flag.

[0016] Secondly, the present invention also discloses an aluminum pigment coating curing temperature measurement system, which is used to implement the above method, and the system includes: One or more processors; Memory, used to store one or more programs; When one or more programs are executed by one or more processors, the one or more processors implement the following modules: The radiation signal acquisition module is used to synchronously acquire radiation signals from the surface of a target workpiece with an aluminum pigment coating that is in high-speed motion, using a multispectral infrared polarization sensor array in different infrared spectral bands and different polarization directions. The signal processing and temperature acquisition module is used to process the acquired radiation signal using at least one of the following two processing modes to obtain the true temperature of the coating: Mode A: Based on the radiation signal, the true temperature of the coating is calculated by solving the polarization radiation transmission equations. Mode B: Extract high-dimensional transient radiation features from the radiation signal and calculate the true temperature of the coating by matching them with a pre-built transient micro-polarization fingerprint database; The temperature mapping distribution module is used to reconstruct and map all the acquired single-point temperature data, corresponding to the real-time motion information of the target workpiece, into a temperature distribution map covering the entire surface of the target workpiece through coordinate transformation and spatial interpolation. The anomaly identification and alert module is used to identify abnormal areas in the temperature distribution map based on the preset temperature threshold range and uniformity standard threshold range, and to issue early warning alerts.

[0017] This technical solution provides an integrated hardware and software platform to achieve automated, high-precision measurement and real-time monitoring of the curing temperature of aluminum pigment coatings. It effectively solves the problems of insufficient accuracy and poor adaptability of traditional temperature measurement systems in complex industrial environments, thereby significantly improving production efficiency and product quality.

[0018] The beneficial effects of this invention are: This invention discloses a method for measuring the curing temperature of aluminum pigment coatings. By employing a multispectral infrared polarization sensor array to simultaneously acquire radiation signals from the target workpiece surface across different infrared spectral bands and polarization directions, and combining two innovative signal processing modes (inversion calculation based on the polarization radiative transfer equation set and matching calculation based on a transient micro-polarization fingerprint database), it effectively separates the thermal radiation of the aluminum pigment coating itself from the reflected radiation from the environment, overcoming the strong interference problem of high-reflectivity coatings on infrared temperature measurement in existing technologies. Furthermore, by combining the acquired single-point temperature data with the real-time motion information of the target workpiece, a temperature distribution map covering the entire workpiece surface is reconstructed and mapped, achieving comprehensive and refined monitoring of the surface temperature of high-speed moving workpieces, overcoming the limitation of traditional point-based temperature measurement in depicting fine temperature curves. Finally, by identifying and alerting abnormal areas based on preset thresholds, temperature deviations during the curing process can be detected and corrected in a timely manner, effectively preventing quality problems such as color difference, uneven gloss, or coating peeling. This method not only improves the accuracy and reliability of temperature measurement but also enhances its adaptability to different coating formulations and workpiece designs, significantly improving the level of curing process control and product quality in industrial production. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of a method for measuring the curing temperature of an aluminum pigment coating provided by the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. It should be noted that the embodiments described herein (including but not limited to methods, systems, devices, etc.) are only a part of the implementation of the present invention, and not all of them. Generally, the contents described and shown in the accompanying drawings (such as method steps, system modules, component connections or structural configurations, etc.) can be adjusted, combined or designed in many different ways.

[0021] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention, without inventive effort, through equivalent substitution, sequence adjustment, structural modification, or logical reconstruction, are within the scope of protection of the present invention.

[0022] See Figure 1 This invention proposes a method for measuring the curing temperature of aluminum pigment coatings, the method comprising the following steps: S1. Acquiring radiation signals: For a target workpiece that is in a high-speed motion state and has an aluminum pigment coating on its surface, a multispectral infrared polarization sensor array is used to synchronously acquire radiation signals from the surface of the target workpiece in different infrared spectral bands and different polarization directions.

[0023] S2. Signal Processing and Temperature Acquisition: The acquired radiation signal is processed using at least one of the following two processing modes to obtain the true temperature of the coating: Mode A: Based on the radiation signal, the true temperature of the coating is calculated by solving the polarization radiation transmission equations. Mode B: Extract high-dimensional transient radiation features from the radiation signal, and calculate the true temperature of the coating by matching them with a pre-built transient micro-polarization fingerprint database.

[0024] S3. Temperature Mapping Distribution: All the acquired single-point temperature data, corresponding to the real-time motion information of the target workpiece, are reconstructed and mapped into a temperature distribution map covering the entire surface of the target workpiece through coordinate transformation and spatial interpolation.

[0025] S4. Identify and alert to anomalies: Based on the preset temperature threshold range and uniformity standard threshold range, identify abnormal areas in the temperature distribution map and issue an early warning.

[0026] This invention introduces a multispectral infrared polarization sensor array to collect radiation signals, and combines the inversion of polarization radiation transfer equations or the matching and calculation of transient micro-polarization fingerprint database to achieve accurate measurement of the true temperature of aluminum pigment coating on the surface of high-speed moving target workpieces. Furthermore, through temperature mapping distribution and anomaly identification prompts, it effectively solves the problems of temperature measurement distortion, difficulty in obtaining fine temperature distribution, and poor adaptability of measurement methods in the prior art.

[0027] This invention provides a method for measuring the curing temperature of aluminum pigment coatings. Its core lies in acquiring rich radiation signals through a multispectral infrared polarization sensor array and combining it with advanced signal processing modes to achieve accurate measurement of the true temperature of the coating on the surface of a high-speed moving workpiece.

[0028] Specifically, in step S1, it is necessary to acquire radiation signals. For a target workpiece in high-speed motion with an aluminum pigment coating on its surface, a multispectral infrared polarization sensor array can be used to simultaneously acquire radiation signals from the surface of the target workpiece in different infrared spectral bands and different polarization directions. For example, an array composed of multiple infrared sensors can be deployed, each equipped with a filter and a polarizer, enabling it to acquire data in multiple bands such as short-wave infrared and mid-wave infrared, and in multiple polarization directions such as 0 degrees, 45 degrees, and 90 degrees. When the target workpiece passes through the measurement area at high speed, the sensor array will synchronously acquire radiation data of its surface in different spectral and polarization dimensions at an extremely high sampling frequency.

[0029] In step S2, the collected radiation signal needs to be processed to obtain the true temperature of the coating. This invention provides at least one of two processing modes.

[0030] Mode A: Based on the radiation signal, the true temperature of the coating is calculated by solving the polarization radiative transfer equations. For example, a set of polarization radiative transfer equations, including Fresnel equations, can be established using the collected multispectral polarization radiation data, combined with the optical property model of the coating and the environmental radiation model.

[0031] Mode B: High-dimensional transient radiation features are extracted from the radiation signal and matched with a pre-constructed transient micro-polarization fingerprint database to deduce the true temperature of the coating. For example, the acquired radiation signal can be pre-processed to extract transient features such as radiation intensity, polarization degree, and spectral shape over a continuous time series, and the dynamic change rate of these features can be calculated to form a high-dimensional transient micro-polarization feature vector. This feature vector contains a unique "fingerprint" of the coating's microstructure evolution and temperature changes during curing. Then, this real-time acquired feature vector is compared with a transient micro-polarization fingerprint database pre-constructed through offline experiments. The fingerprint database stores standard transient micro-polarization feature vectors of coatings under different true temperatures and microstructure states. By calculating similarity (such as cosine similarity) or distance (such as Euclidean distance), the best-matching fingerprint entry is found, thereby deducing the true temperature of the current measurement point.

[0032] In step S3, temperature mapping distribution needs to be performed. All acquired single-point temperature data, corresponding to the real-time motion information of the target workpiece, are reconstructed and mapped into a temperature distribution map covering the entire surface of the target workpiece through coordinate transformation and spatial interpolation. For example, single-point temperature data collected by the sensor array at different times will be associated with the precise location of the target workpiece at the corresponding time. By projecting these discrete temperature point data onto a two-dimensional plane or three-dimensional model of the target workpiece, and using spatial interpolation algorithms (such as Kriging interpolation, inverse distance weighted interpolation, etc.), a continuous, high-resolution temperature distribution map can be generated, intuitively displaying the temperature uniformity and local hot spots across the entire workpiece surface.

[0033] In step S4, anomalies need to be identified and alerted. Based on preset temperature threshold ranges and uniformity standard threshold ranges, abnormal areas in the temperature distribution map are identified, and warnings are issued. For example, an ideal curing temperature range and an allowable temperature gradient or uniformity deviation range can be set. The system analyzes the generated temperature distribution map in real time, comparing the real-time temperature of each point with the preset temperature thresholds to identify abnormal points that exceed the range. Furthermore, these abnormal points are spatially clustered to form abnormal connected regions, and statistical characteristics such as the average temperature and maximum temperature difference of these regions are calculated. If these statistical characteristics exceed the preset uniformity standard threshold, the system will immediately trigger a warning, such as through audible and visual alarms, displaying the location of the abnormal area, or sending control commands to the system, so that operators can intervene in a timely manner to adjust the curing process parameters.

[0034] The aluminum pigment coating curing temperature measurement method proposed in this invention aims to solve the problems in the prior art when aluminum pigment coatings are cured on high-speed moving workpieces, such as the susceptibility of infrared temperature measurement to environmental reflection interference, the difficulty in obtaining fine temperature distribution, and the difficulty in adapting to frequent coating or workpiece design changes.

[0035] This invention introduces a multispectral infrared polarization sensor array, enabling simultaneous acquisition of radiation signals from the surface of a target workpiece across different infrared spectral bands and polarization directions. This multidimensional data acquisition method provides richer information compared to traditional single-spectral or unpolarized infrared thermometry, laying the foundation for accurately distinguishing between the coating's own thermal radiation and environmental reflected radiation.

[0036] In terms of signal processing and temperature acquisition, this invention provides two innovative modes: Mode A and Mode B.

[0037] Mode A, by solving the polarization radiation transport equations, can invert and calculate the true temperature of the coating from complex radiation signals. Traditional methods often struggle to effectively separate ambient reflected radiation, leading to distorted measurement results. This invention, however, utilizes polarization information combined with the transport equations established by Fresnel's equations to more accurately eliminate environmental interference, thereby obtaining the true temperature of the coating. This is particularly crucial for high-reflectivity aluminum pigment coatings.

[0038] Mode B introduces transient micro-polarization fingerprint database matching and estimation. In actual production, coating formulations or workpiece designs frequently change, making traditional methods relying on pre-calibration difficult to adapt. This invention extracts high-dimensional transient radiation features and matches them with a pre-constructed fingerprint database to quickly and adaptively estimate the coating temperature. This method eliminates the need for complex offline recalibration for each switch, significantly improving production efficiency and measurement flexibility.

[0039] Furthermore, this invention also includes temperature mapping distribution and anomaly identification and alerting steps. By combining all single-point temperature data with the real-time motion information of the target workpiece, and utilizing coordinate transformation and spatial interpolation, a temperature distribution map covering the entire surface of the target workpiece is reconstructed and mapped. This solves the problem that traditional point-based temperature measurement cannot depict a fine temperature curve, enabling operators to fully understand the temperature uniformity of the workpiece surface. Based on this, according to preset temperature thresholds and uniformity standards, abnormal areas in the temperature distribution map are identified and early warning alerts are issued, enabling timely detection and correction of potential quality problems during the curing process and avoiding product defects.

[0040] In summary, this invention, through a series of technical means including multispectral infrared polarization sensing, polarization radiative transfer equation inversion or transient micro-polarization fingerprint database matching, temperature mapping distribution, and anomaly identification and alerts, forms a complete, efficient, and adaptive method for measuring the curing temperature of aluminum pigment coatings. Compared to existing technologies, this invention can more accurately measure the true temperature of the aluminum pigment coating on the surface of high-speed moving workpieces, effectively overcome environmental reflection interference, adapt to the rapidly changing needs of production lines, and provide comprehensive temperature distribution information and early warning information, thereby significantly improving the control accuracy of the curing process and product quality.

[0041] In actual industrial production environments, target workpieces are typically in high-speed motion, and the optical properties of their surface aluminum pigment coatings are complex. Furthermore, interference from environmental radiation and the nonlinear coupling between coating emissivity and temperature pose challenges to directly and accurately solving the polarization radiative transfer equations to invert the true coating temperature. Failure to fully consider these complex factors may affect the accuracy and reliability of the temperature inversion results. To address this, this invention further proposes a detailed implementation of Mode A to improve the accuracy and robustness of temperature measurements under complex operating conditions.

[0042] In the aluminum pigment coating curing temperature measurement method proposed in this invention, step S2, which involves calculating the true temperature of the coating based on the radiation signal by solving the polarization radiation transport equations, specifically includes: Before the target workpiece enters the measurement area, a radiation measurement is performed on a preset reference point to obtain the spectral radiance of the reference point. Based on the known emissivity of the reference point... and known true temperature The ambient spectral radiance is obtained by inverse calculation using Planck's radiation law.

[0043] The total radiation received by the reference point includes both its own emitted and reflected ambient radiation. = = For opaque surfaces, according to Kirchhoff's laws, we know that... ,in, For the reference point spectral radiance, For reference point emissivity, Reflectance at reference point Let Planck's blackbody radiation function be the reference point. From the ambient spectral radiance, we can deduce: .

[0044] This step aims to accurately characterize the background radiation field in the measurement environment, accurately obtain the ambient spectral radiance of the reference point, and provide the necessary environmental parameters for solving the subsequent radiative transfer equation, thereby effectively separating the thermal emission radiation of the target workpiece itself from the reflected radiation of the environment.

[0045] Based on the collected radiation signals, a set of polarization radiation transport equations based on Fresnel's equations is established, and the thermal emission radiation of the target workpiece and the reflected radiation from the environment are obtained by solving these equations. By solving this set of equations, the total radiation signal can be decomposed into two parts: the thermal emission radiation of the target workpiece and the reflected radiation from the environment. This is the key to accurately inverting the true temperature of the target workpiece.

[0046] In some embodiments of this invention, a set of polarization radiative transfer equations based on Fresnel equations is established. This set of equations can describe the coating's own thermal emission radiation, environmental reflected radiation, and their respective contributions in different polarization directions. When polarization information is complex, Stokes parameters or Mueller matrix methods can be used for more refined polarization analysis. These equations can be solved using numerical iterative algorithms, such as the Newton-Raphson method or the least squares method, thereby separating the coating's own thermal emission radiation and establishing an inversion equation set to calculate the coating's true temperature.

[0047] When measuring the target surface, the received spectral radiance is the superposition of its own thermal emission and ambient reflected radiation. = ,in The total radiation of the target workpiece. For coating at wavelength The effective emissivity is given by T, where T is the true temperature of the coating. Based on the above formula, different polarization components or different observation angles are established. The following system of equations, for example: parallel components and vertical components At two observation angles and The equation is established as follows: .

[0048] Based on Fresnel equations and optical properties, a solvable set of equations was further established, and the thermal emission radiation components of the coating itself were obtained by numerical iterative algorithm. and environmental reflected radiation components .

[0049] In some embodiments of the present invention, when the degree of polarization of the emitted radiation is close to 0, the polarization radiation transport equations are specifically established as follows: ; ; = ; For emitted radiation, the degree of polarization is typically close to 0. ≈ ≈ ; - = - ; polarization degree of reflected radiation ; = ; - = ; = ; = - = + - = + - .

[0050] In the above formula, The total radiance measured in the parallel polarization direction. This represents the total radiance measured in the vertical polarization direction. This represents the total radiance. This represents the total reflected radiance. This represents the parallel polarization component of the reflected radiation. This represents the vertical polarization component of the reflected radiation. The total emitted radiance, For the parallel polarization component of the emitted radiation, This represents the vertical polarization component of the emitted radiation. The degree of polarization of the reflected radiation is calculated in advance using Fresnel's equations or obtained through experimental calibration.

[0051] By pre-setting the emissivity ratio of the coating at two wavelengths, and based on the direct proportionality between the thermal emission radiation of the target workpiece itself and the Planck blackbody radiation function (ignoring ambient reflected radiation), an initial temperature estimate of the coating is obtained. This initial estimate serves as a good starting point for numerical iterative solutions, helping to accelerate convergence and avoid getting trapped in local optima. Based on the functional relationship between the coating emissivity ratio and the actual coating temperature, and the ambient spectral radiance at the two wavelengths, an inversion equation set is established. Starting with the initial temperature estimate, the inversion equation set is solved using a numerical iterative method to obtain the actual temperature of the coating and its effective emissivity at the two wavelengths.

[0052] This invention effectively quantifies and separates the interference of environmental radiation on the radiation signal of the target workpiece by introducing precise measurement of the ambient spectral radiance, thus overcoming the challenges posed by the complexity of the environmental background in traditional methods. By establishing a set of polarization radiative transfer equations based on Fresnel's equations, the thermal emission radiation of the target workpiece itself and the reflected radiation from the environment can be analyzed and distinguished more accurately, thereby providing a more reliable physical model for the inversion of the true temperature. In addition, by obtaining an initial temperature estimate through a preset coating emissivity ratio, an efficient and stable starting point is provided for subsequent numerical iterative solutions, significantly improving the solution efficiency and convergence of the nonlinear inversion equations. Finally, by combining the functional relationship between the coating emissivity ratio and the true temperature, the true temperature and effective emissivity of the coating can be obtained simultaneously through iterative solutions, thus overcoming the errors caused by unknown or assumed constant emissivity in single-wavelength temperature measurements.

[0053] Through the above technical solutions, this invention can significantly improve the accuracy and robustness of aluminum pigment coating curing temperature measurement. Specifically, by accurately measuring the ambient spectral radiance, the interference of ambient background radiation on the measurement results is effectively eliminated; by establishing a set of polarization radiative transfer equations based on Fresnel's equations, accurate separation of the target workpiece's own thermal emission radiation and ambient reflected radiation is achieved; by obtaining the initial temperature estimate and using a numerical iterative solution method, the complexity of solving nonlinear equations is overcome, improving inversion efficiency and accuracy. These improvements enable the acquisition of high-precision true coating temperature and effective emissivity even under high-speed motion and complex environmental conditions, providing more reliable data support for quality control of the aluminum pigment coating curing process, thereby effectively avoiding coating defects or performance degradation caused by inaccurate temperature measurement.

[0054] In some preferred embodiments, it is assumed that the target workpiece passes through the curing oven at a speed of 120 m / min on an aluminum pigment coating curing production line. To accurately measure its surface temperature, a reference blackbody with a known emissivity (e.g., 0.98) and a known true temperature (e.g., 350 Kelvin) is placed near the measurement area before the target workpiece enters the measurement area. A multispectral infrared polarization sensor array measures the radiation of this reference blackbody and inversely derives the spectral radiance of the current environment according to Planck's radiation law. When the target workpiece enters the measurement area, the sensor array simultaneously acquires radiation signals from its surface in multiple infrared spectral bands and different polarization directions. Based on these acquired radiation signals, a set of polarization radiative transfer equations is established using Fresnel equations to separate the thermal emission radiation of the target workpiece itself from the reflected radiation from the environment. To obtain an initial temperature estimate, a preset emissivity ratio (e.g., 0.85) of the aluminum pigment coating at two specific wavelengths (e.g., 3.9 μm and 4.5 μm) can be used, combined with the direct proportionality between its own thermal emission radiation and Planck's blackbody radiation function, for calculation. Subsequently, using this initial temperature estimate as a starting point, and combining the functional relationship between the coating emissivity ratio and the true temperature obtained through offline calibration experiments, as well as the previously measured environmental spectral radiance, an inversion equation set was established. This inversion equation set can be numerically solved iteratively using the Newton-Raphson method until convergence, ultimately yielding the accurate true temperature of the coating on the target workpiece surface and its effective emissivity at wavelengths of 3.9 μm and 4.5 μm. This method ensures the accuracy of coating curing temperature measurement even under high-speed production and complex environmental interference.

[0055] Specifically, the steps for establishing the inversion equation set based on the functional relationship between the coating emissivity ratio and the coating's true temperature and the environmental spectral radiance at the two wavelengths can be implemented in the following way.

[0056] Specifically, the steps for establishing the inversion equation system described above include: For the first wavelength Second wavelength ,satisfy: ; ; / =f( The functional relationship was obtained by conducting offline spectral radiometric calibration experiments on standard samples of the coating. In the formula, For coating at wavelength The measured spectral radiance, For coating at wavelength Effective emissivity at the following levels Let be the Planck blackbody radiation function. As a reference point at wavelength The measured environmental spectral radiance, This represents the actual temperature of the coating.

[0057] in, Indicates at wavelength The spectral radiance of the coating surface radiation signal acquired by the multispectral infrared polarization sensor array is then processed. This refers to the coating at wavelength The effective emissivity reflects the ability of the coating surface to emit radiation. It is Planck's blackbody radiation function, used to describe the radiation of a blackbody at temperature T at wavelength T. The radiation intensity at that location. It is the ambient spectral radiance, the value of which is obtained by performing radiation measurements on a preset reference point before the target workpiece enters the measurement area and then inverting the result using Planck's radiation law.

[0058] Furthermore, the functional relationship / =f( This is crucial auxiliary information. The functional relationship is obtained through offline spectral radiometric calibration experiments on standard samples with the same or equivalent characteristics as the coating of the target workpiece. Under controlled experimental conditions, by accurately measuring the spectral radiance and effective emissivity of the standard samples at different real temperatures, an empirical functional relationship between the emissivity ratio and the real temperature can be established, thus providing prior knowledge for subsequent inversion calculations.

[0059] This invention constructs a closed system of equations by explicitly defining the above three equations, such that, given... , and , In this case, the true temperature T of the coating and the effective emissivity at both wavelengths can be solved. and Specifically, the first two equations describe the physical relationship between the measured spectral radiance, coating emissivity, and true temperature. The third equation introduces an inherent property of the coating material: the functional relationship between the ratio of its emissivity at two specific wavelengths and the true temperature. It is precisely this functional relationship, obtained through offline calibration, that makes the previously underdetermined set of equations solvable, allowing for the accurate inversion calculation of the coating's true temperature from the acquired radiation signal through numerical iterative solutions.

[0060] The above technical solution clarifies the specific mathematical form of the inversion equations, providing a solid theoretical foundation for subsequent numerical iterative solutions. In particular, by introducing a functional relationship between the coating emissivity ratio obtained from offline calibration and the true temperature, the problem of unknown or difficult-to-determine emissivity in multispectral thermometry is effectively solved, improving the accuracy and reliability of temperature inversion. This explicit equation construction method enables precise measurement of the curing temperature of aluminum pigment coatings even under complex high-speed motion and environmental radiation effects, providing high-quality input data for subsequent temperature distribution map reconstruction and anomaly identification.

[0061] In some embodiments of the present invention described above, when calculating the true temperature of the coating by solving the polarization radiative transfer equations, it is necessary to obtain an initial temperature estimate of the coating as the starting point for the numerical iterative solution. Specifically, the initial temperature estimate of the coating can be obtained in the following ways.

[0062] The method of obtaining an initial temperature estimate of the coating by using the pre-set emissivity ratio of the coating at two wavelengths, based on the direct proportional relationship between the thermal emission radiation of the target workpiece itself and the Planck blackbody radiation function when ignoring ambient reflected radiation, includes: calculating the initial temperature estimate of the coating based on its own thermal emission radiation at the two wavelengths and in combination with the pre-set emissivity ratio of the coating at the two wavelengths using Wien's approximation law or a lookup table method.

[0063] The "Wien approximation law calculation" refers to simplifying the blackbody radiation formula using Wien's approximation law. When the product of wavelength and temperature is small, Planck's law can be approximated as Wien's approximation law. By measuring the self-thermal emission radiation at two specific wavelengths and combining it with a preset ratio of the coating's emissivity at these two wavelengths, the initial temperature estimate of the coating can be directly calculated using the mathematical form of Wien's approximation law. This method has high computational efficiency and accuracy within a specific temperature and wavelength range.

[0064] The "lookup table method" refers to pre-establishing a lookup table containing the relationships between different intrinsic thermal emission values, emissivity ratios, and corresponding initial temperature estimates. During actual measurement, based on the measured intrinsic thermal emission at two wavelengths and the preset emissivity ratio, a direct match is performed in the lookup table to quickly obtain the initial temperature estimate of the coating. This lookup table can be pre-generated using a large amount of experimental data or theoretical calculations to cover the possible operating temperature range.

[0065] This invention obtains the initial temperature estimate of the coating by employing Wien's approximation law or a lookup table method. The working principle is that a reasonable initial value is needed to accelerate convergence and improve computational stability during the iterative process of calculating the true temperature of the coating. Wien's approximation law can simplify the radiative transfer equation under certain conditions, making the calculation of the initial temperature more direct and efficient. Meanwhile, the lookup table method avoids complex real-time calculations by using pre-stored empirical data or calculation results, thus providing a rapid initial temperature estimate close to the true value. Both methods utilize the direct proportionality between the coating's own thermal emission radiation at different wavelengths and the Planck blackbody radiation function, combined with a preset emissivity ratio, thereby providing a reliable starting point for subsequent numerical iterative solutions under the simplified condition of neglecting environmental reflected radiation.

[0066] The above technical solutions enable efficient and accurate acquisition of the initial temperature estimate of the coating. Using Wien's approximation law, a relatively accurate initial temperature can be obtained with lower computational complexity under certain conditions, providing a good starting point for subsequent numerical iterative solutions and accelerating the convergence of the iteration process. The lookup table method further improves the real-time performance of the initial temperature estimate, particularly suitable for industrial scenarios with high computational speed requirements, significantly shortening the initial temperature acquisition time and thus improving the response speed and efficiency of the entire temperature measurement method. Both methods effectively support the subsequent numerical iterative solution of the inversion equations, ensuring the accuracy and stability of the final calculated true temperature.

[0067] Specifically, in some embodiments of the above-mentioned method for measuring the curing temperature of aluminum pigment coatings, the numerical iterative solution method used to solve the inversion equation set can be further clarified.

[0068] According to the above method for measuring the curing temperature of aluminum pigment coating, the numerical iterative solution method adopts the Newton-Raphson method or the least squares method.

[0069] The numerical iterative solution method refers to a method that approximates the exact solution through a series of approximate calculations when solving nonlinear equation systems or optimization problems. Specifically, the Newton-Raphson method is an efficient iterative method that determines the next iteration point by linearly approximating the function at the current point (i.e., using the first-order term of the Taylor expansion) and solving the linear equation, thus quickly converging to the root of the equation. This method is suitable for solving nonlinear equation systems with good differentiability. The least squares method is an optimization method that aims to minimize the sum of squared errors and is commonly used for curve fitting and parameter estimation. In inversion problems, when the established inversion equation system can be transformed into an optimization problem, the least squares method can iteratively adjust the parameters to minimize the sum of squared residuals between the model's predicted values ​​and the actual measured values, thereby obtaining the true temperature and effective emissivity of the coating. In practical applications, an appropriate numerical iterative solution method can be selected based on factors such as the specific form of the inversion equation system, convergence requirements, and computational resources.

[0070] This invention provides a concrete and operable algorithmic basis for solving the inversion equations by explicitly specifying the numerical iterative solution method as either the Newton-Raphson method or the least squares method. The Newton-Raphson method utilizes the local linearity of functions, enabling rapid convergence to the solution of the equations, and is particularly suitable for inversion equations with good derivative properties. The least squares method finds the optimal solution by minimizing the sum of squared errors, providing a robust solution even when the inversion equations contain noise or uncertainty. Both methods effectively handle nonlinear inversion problems, ensuring a gradual approximation of the coating's true temperature and effective emissivity during the iteration process.

[0071] The above technical solution clarifies the specific algorithm for the numerical iterative solution of the inversion equations, improving the accuracy and efficiency of temperature inversion calculations. The Newton-Raphson method accelerates the convergence process and reduces computation time; while the least squares method provides a more robust solution even with measurement errors or model uncertainties, thus improving the reliability of obtaining the true temperature of the coating. This helps ensure accurate measurement of the curing temperature of the aluminum pigment coating under high-speed motion, providing a solid data foundation for subsequent temperature distribution map reconstruction and anomaly identification.

[0072] Specifically, in the above-mentioned method for measuring the curing temperature of aluminum pigment coatings, step S2, which involves extracting high-dimensional transient radiation features from the radiation signal and calculating the true temperature of the coating by matching them with a pre-constructed transient micro-polarization fingerprint database, can be implemented as follows: The acquired radiation signal is preprocessed to obtain transient radiation intensity values. All the acquired transient radiation intensity values ​​are combined in a preset order to form an initial feature vector. The dynamic change rate of the initial feature vector in a continuous time series is calculated to obtain dynamic change information. The dynamic change information is used as a new feature dimension to expand the initial feature vector to form a higher-dimensional transient micro-polarization feature vector. In a controlled environment, a reference plate with a standard coating having an equivalent structure to the coating of the target workpiece is heated at an ultra-high heating rate simulating the actual production process. During the heating process, the following data are collected simultaneously: full polarization radiation data of the surface of the standard coating, the actual temperature data of the standard coating obtained by a contact sensor, and the microstructure evolution image data of the standard coating obtained by a microscopic imaging system. Based on the full polarization radiation data, a standard transient micro-polarization feature vector is generated. Based on the microstructure evolution image data, the microstructure state is determined. The standard transient micro-polarization feature vector, the actual temperature, and the microstructure state at the same time stamp are associated to form entries in the transient micro-polarization fingerprint database. The transient micro-polarization feature vectors acquired in real time are compared with the transient micro-polarization fingerprint database, and the true temperature of the current measurement point is calculated based on the comparison results.

[0073] Specifically, the transient radiation intensity value refers to the instantaneous quantitative data reflecting the radiation characteristics of the target workpiece surface, acquired by a multispectral infrared polarization sensor array within an extremely short time. These transient radiation intensity values ​​can include radiation intensity information from different spectral bands and different polarization directions. The initial feature vector is formed by arranging and combining these transient radiation intensity values ​​according to a preset logical order, with the aim of initially capturing the radiation characteristics of the coating. Further, the dynamic change rate refers to the rate or trend of numerical change of the initial feature vector at continuous time points, which can be obtained, for example, by calculating the difference or gradient of the feature vector at adjacent time points. The dynamic change information, as an important indicator reflecting the dynamic evolution of the coating state, is expanded into the initial feature vector, thereby forming a higher-dimensional transient micro-polarization feature vector. This vector can more comprehensively and finely characterize the instantaneous physical state and microstructure changes of the coating during the curing process.

[0074] The standard coating refers to a coating sample that is highly similar to the target workpiece coating in terms of material composition, structural characteristics, and curing behavior, with the aim of simulating real-world working conditions under controlled conditions. The ultra-high heating rate aims to reproduce the temperature changes of a workpiece rapidly passing through the heating zone on an actual production line, ensuring the authenticity and applicability of the fingerprint database data. The fully polarized radiation data refers to data collected by a polarization sensor during the heating process, containing complete polarization information such as radiation intensity, degree of polarization, and polarization angle. The real temperature data is typically obtained directly through high-precision contact or non-contact temperature sensors (such as thermocouples, fiber optic thermometers, etc.) and serves as the benchmark for establishing the fingerprint database. The microstructure evolution image data is acquired in real-time or near real-time using a microscopic imaging system (such as an infrared microscope, optical microscope, etc.) to record the changes in the internal microstructure (such as crystal growth, phase transition, pore formation, etc.) of the coating during the heating and curing process.

[0075] Therefore, the standard transient micro-polarization feature vector is generated based on fully polarized radiation data acquired under controlled conditions, and it has the same dimension and structure as the transient micro-polarization feature vector acquired in real time. The microstructure state is obtained from the analysis of microstructure evolution image data and is used to describe the microstructural characteristics of the coating at specific temperatures and time points. Correlating these standard transient micro-polarization feature vectors, the actual temperature, and the microstructure state acquired at the same time stamp forms entries in the transient micro-polarization fingerprint database. These entries collectively constitute the transient micro-polarization fingerprint database, providing a reference for subsequent real-time temperature estimation. In practical applications, the transient micro-polarization feature vector acquired in real time is compared with the transient micro-polarization fingerprint database. The purpose is to find the entry in the fingerprint database that best matches the current real-time measurement state, and then calculate the actual temperature of the current measurement point based on the actual temperature recorded in the matching entry.

[0076] This invention achieves precise measurement of the curing temperature of aluminum pigment coatings by constructing a high-dimensional transient micro-polarization feature vector and a transient micro-polarization fingerprint library. Specifically, in the feature extraction stage, not only is the instantaneous radiation intensity considered, but its dynamic rate of change over time is also introduced. This allows the resulting transient micro-polarization feature vector to more comprehensively capture the complex physicochemical changes of the coating during curing, including microstructure evolution and compositional changes, which are often accompanied by transient fluctuations in radiation characteristics. By simulating the ultra-high heating rate of actual production processes in a controlled environment and simultaneously acquiring fully polarized radiation data, real temperature data, and microstructure evolution image data, a transient micro-polarization fingerprint library containing multi-dimensional information can be established. This fingerprint library closely correlates the radiation characteristics, real temperature, and microstructure state of the coating, forming a comprehensive "fingerprint" system. When the transient micro-polarization feature vector of the target workpiece is acquired in real time, it can be compared with the pre-constructed fingerprint library to quickly and accurately find the best-matching fingerprint entry, thereby calculating the real temperature of the current measurement point. This feature-matching-based mechanism effectively avoids measurement errors caused by unknown or varying emissivity in traditional radiation thermometry methods, and is particularly suitable for target workpieces with high-speed movement and complex surface characteristics.

[0077] Through the above technical solution, this invention overcomes the challenges faced by traditional radiation thermometry methods in measuring the curing temperature of target workpieces with high-speed movement and complex surface characteristics (such as aluminum pigment coatings). Specifically, by extracting high-dimensional transient radiation features and combining them with their dynamic change information, it can more sensitively and accurately reflect subtle changes in the coating curing process, thereby improving the accuracy and robustness of temperature measurement. Furthermore, by constructing a transient micro-polarization fingerprint database containing fully polarized radiation data, true temperature data, and microstructure evolution image data, this invention can establish a more direct and reliable mapping relationship between radiation features and true temperature, effectively solving the problem of difficulty in accurately determining the coating emissivity as it changes with temperature, wavelength, polarization state, and degree of curing. This temperature estimation method based on fingerprint database matching enables accurate measurement of the coating's true temperature even without precisely knowing the coating's emissivity, making it particularly suitable for industrial production scenarios with extremely high requirements for curing temperature accuracy. This helps optimize production processes, improve product quality, and reduce scrap rates.

[0078] In some embodiments of the present invention described above, a technical solution is proposed to extract high-dimensional transient radiation features from the radiation signal and calculate the true temperature of the coating by matching them with a pre-constructed transient micro-polarization fingerprint database. However, in its implementation, the basic solution does not provide specific, quantitative methods for accurately quantifying the similarity between the real-time acquired feature vectors and the feature vectors stored in the fingerprint database, or for reliably calculating the true temperature when multiple matching entries exist. This lack of specific matching and calculation mechanisms may lead to uncertainty in the matching results or affect the accuracy and robustness of temperature calculation in complex and variable environments.

[0079] To address this, the present invention further proposes a step of comparing the transient micro-polarization feature vector acquired in real time with the transient micro-polarization fingerprint database, and calculating the true temperature of the current measurement point based on the comparison result, including: Calculate the cosine similarity or Euclidean distance between the transient micro-polarization feature vectors acquired in real time and the feature vectors stored in the transient micro-polarization fingerprint database; Based on the cosine similarity being higher than a preset threshold or the Euclidean distance being lower than a preset threshold, one or more entries matching the transient micro-polarization feature vectors acquired in real time are selected from the transient micro-polarization fingerprint database. The actual temperature is calculated by using a weighted average or interpolation method to calculate the temperature values ​​corresponding to all matching entries.

[0080] Specifically, when comparing feature vectors, cosine similarity or Euclidean distance can be used as indicators to measure the similarity between two transient micro-polarization feature vectors. Cosine similarity measures the directional similarity between two vectors, and for high-dimensional data, it effectively reflects the closeness of feature patterns. Euclidean distance directly reflects the geometric distance between two vectors in multidimensional space; the smaller the distance, the higher the similarity. In practical applications, an appropriate similarity or distance metric can be selected based on the specific data characteristics and matching accuracy requirements.

[0081] Furthermore, to ensure the accuracy and reliability of the matching, after calculating the similarity or distance, filtering is performed based on preset thresholds. For example, when using cosine similarity, two feature vectors are considered a match only if the value is higher than the preset similarity threshold; when using Euclidean distance, a match is considered successful only if the value is lower than the preset distance threshold. By setting reasonable thresholds, irrelevant fingerprint database entries can be effectively excluded, improving the accuracy of the matching. One or more selected matching entries represent historical data records in the fingerprint database that are closest to the current real-time measurement state.

[0082] Furthermore, when multiple matching fingerprint entries are selected, a weighted average or interpolation method can be used to calculate the more accurate true temperature. The weighted average method assigns different weights to each matching entry based on its similarity or distance; entries with higher similarity or closer distances are given greater weight, thus combining multiple matching results to obtain the final temperature value. Interpolation methods, such as linear or polynomial interpolation, can construct a functional relationship using the matching entries and their corresponding temperature values ​​to calculate the true temperature corresponding to the current feature vector. Both methods aim to improve the accuracy and robustness of temperature estimation by fusing multiple relevant information.

[0083] This invention addresses the lack of precise definitions and quantification standards in the matching and temperature estimation processes of the aforementioned basic schemes by introducing specific similarity calculation methods (cosine similarity or Euclidean distance), threshold filtering mechanisms, and multi-entry temperature estimation methods (weighted average or interpolation). Specifically, the introduction of cosine similarity or Euclidean distance allows for precise quantification of the similarity between the real-time acquired transient micro-polarization feature vector and the standard feature vector in the transient micro-polarization fingerprint database, thus providing an objective basis for subsequent matching. By judging similarity or distance using preset thresholds, irrelevant fingerprint database entries can be effectively filtered out, ensuring that the selected matching entries are highly correlated with the current measurement state and avoiding temperature estimation errors caused by fuzzy matching. When multiple highly matched fingerprint database entries exist, the weighted average or interpolation method is used to comprehensively process the temperature values ​​corresponding to these entries. This fully utilizes the rich information contained in the fingerprint database, further smoothing noise and reducing random errors through data fusion, and improving the accuracy and stability of temperature estimation. It is precisely because of these specific and quantitative processing steps that the process of estimating the true temperature of the coating from high-dimensional transient radiation characteristics becomes more scientific and reliable.

[0084] Through the above technical solutions, this invention significantly improves the accuracy and reliability of the method for measuring the curing temperature of aluminum pigment coatings. Specifically, using cosine similarity or Euclidean distance for feature vector comparison provides a clear quantitative standard for the matching process, avoiding errors from subjective judgment. Threshold-based screening of matching entries effectively eliminates the influence of low-correlation data on temperature estimation, ensuring the accuracy of the matching results. More importantly, when multiple matching entries exist, the weighted average method or interpolation method is used to comprehensively estimate the true temperature, fully utilizing multi-source information in the fingerprint database. This effectively reduces the uncertainty that a single match may bring, improving the robustness and anti-interference ability of temperature estimation. Compared to simply performing general "comparison" and "estimation," this invention provides a refined data processing flow, making the final obtained coating true temperature closer to the actual value, providing a more solid data foundation for subsequent curing process control and anomaly identification.

[0085] In some preferred embodiments, the present invention is implemented as follows: Assume that the transient micro-polarization feature vector acquired in real time is V_real = [v1, v2, ..., vn], and the transient micro-polarization fingerprint database stores multiple standard feature vectors and their corresponding real temperatures, such as V_lib1 (T1), V_lib2 (T2), ..., V_libk (Tk).

[0086] First, calculate the cosine similarity between V_real and each V_lib_i in the fingerprint database. For example, if the calculated cosine similarity with V_lib1 is 0.95, with V_lib2 is 0.92, with V_lib3 is 0.88, and with V_lib4 is 0.70.

[0087] Secondly, a preset cosine similarity threshold of 0.90 is set. Based on this threshold, V_lib1 and V_lib2 are selected as matching entries because their similarity (0.95 and 0.92) is higher than 0.90, while V_lib3 and V_lib4 are excluded.

[0088] Finally, for the selected matching entries V_lib1 (T1) and V_lib2 (T2), a weighted average method can be used to estimate the true temperature. For example, the true temperature T_real can be calculated as (0.95 * T1 + 0.92 * T2) / (0.95 + 0.92) based on similarity as the weight. Alternatively, if the temperature data in the fingerprint database is continuous, interpolation, such as linear interpolation, can be used to estimate T_real based on the position of V_real in the feature space, combined with the feature vectors of V_lib1 and V_lib2 and their corresponding temperatures. In this way, even if the real-time measurement data is not completely consistent with a single entry in the fingerprint database, a more accurate and reliable temperature estimation result can be obtained by integrating information from multiple highly similar entries.

[0089] In some embodiments of the present invention described above, a scheme is proposed to identify abnormal areas in a temperature distribution map and issue early warnings based on preset temperature threshold ranges and uniformity standard threshold ranges. However, in practical applications, simply performing point-to-point temperature comparisons may lead to misjudging isolated, transient temperature fluctuations as abnormalities, or failing to effectively identify regional anomalies with specific spatial distribution characteristics that significantly impact curing quality. If these problems are not addressed, false alarms or missed alarms may occur, affecting production efficiency and product quality. Therefore, the present invention further proposes an optimized scheme for the above-mentioned anomaly identification and early warning steps by introducing spatial clustering and regional statistical analysis to more accurately identify and locate abnormal areas.

[0090] The steps described above, which involve identifying abnormal regions in the temperature distribution map and issuing warnings based on preset temperature threshold ranges and uniformity standard threshold ranges, include: Based on the temperature distribution map, the real-time temperature value of each coordinate point in the temperature distribution map is compared with the preset temperature threshold range for the corresponding coordinate point to identify abnormal coordinate points; Spatial clustering is performed on the abnormal coordinate points to form one or more abnormal connected regions; For each abnormally connected region, the region's temperature statistical characteristics are calculated, and these characteristics are compared with a preset temperature threshold range and a uniformity standard threshold range. If the comparison result exceeds the threshold range, an early warning is triggered.

[0091] Specifically, when identifying abnormal coordinate points, the temperature threshold range can be understood as the allowable temperature fluctuation range for different locations or stages on the surface of the target workpiece. For example, for a specific area during the coating curing process, its temperature should be maintained within a certain temperature range. By comparing the real-time temperature value of each coordinate point in the temperature distribution map with the preset temperature threshold range, all discrete points exceeding the normal temperature range can be initially screened out; these points are then identified as abnormal coordinate points.

[0092] Spatial clustering of the anomalous coordinate points refers to grouping spatially adjacent anomalous coordinate points together to form one or more continuous anomalous regions. For example, density-based clustering algorithms (such as DBSCAN) or connectivity-based clustering algorithms can be used to merge adjacent anomalous points into a single anomalous connected region. The purpose is to distinguish between isolated anomalous points that may be caused by noise and regional anomalies that are of practical significance and are caused by process problems.

[0093] In practical applications, for each abnormal connected region, calculating the region's temperature statistical characteristics specifically involves calculating the average temperature, highest temperature, lowest temperature, temperature standard deviation, or temperature gradient of all abnormal points within that region. For example, the average temperature reflects the overall temperature level of the region, while the standard deviation reflects the temperature uniformity. These statistical characteristics are then compared with preset temperature threshold ranges and uniformity standard threshold ranges. The uniformity standard threshold range is used to assess the uniformity of temperature distribution within the abnormal region; for example, a maximum permissible temperature standard deviation can be set. If the comparison results show that the region's average temperature exceeds the preset temperature threshold range, or the temperature uniformity within the region (e.g., measured by standard deviation) exceeds the preset uniformity standard threshold range, then the region is determined to have a significant anomaly, and an early warning is triggered.

[0094] This invention first performs a preliminary temperature threshold comparison on each coordinate point in the temperature distribution map, enabling rapid identification of all potential anomalies. Furthermore, by spatially clustering these anomaly coordinate points, discrete anomalies potentially caused by measurement noise can be distinguished from actual, spatially defined anomaly regions. This clustering process allows the system to focus on regional temperature anomalies that are truly likely to affect product quality. Based on this, by calculating the temperature statistical characteristics of each connected anomaly region and comprehensively comparing them with preset temperature threshold ranges and uniformity standard threshold ranges, the severity and nature of the anomaly can be assessed more comprehensively and accurately—for example, whether it is overall overheating / undercooling or poor local uniformity. This multi-level judgment mechanism effectively avoids false alarms or missed alarms that may result from judging by a single indicator, thereby improving the accuracy and reliability of anomaly identification.

[0095] Through the above technical solution, this invention enables more refined and robust identification of temperature anomalies during the curing process of aluminum pigment coatings. Compared to simple point-to-point threshold judgment, this invention introduces spatial clustering to effectively filter out isolated noise points, avoiding unnecessary warnings and significantly reducing the false alarm rate. Simultaneously, by performing statistical feature analysis on abnormally connected regions and combining it with uniformity standards, the system can not only identify areas where the temperature deviates from the normal range but also assess the temperature uniformity within these regions, which is crucial for the curing quality of aluminum pigment coatings. Therefore, this invention can more accurately locate and diagnose potential process problems, providing a more reliable basis for timely adjustment of production parameters and ensuring product curing quality, thereby improving the intelligence and automation level of the production process.

[0096] In some preferred embodiments, it is assumed that during the curing process of the aluminum pigment coating, the surface temperature distribution map of the target workpiece shows multiple coordinate points with temperature values ​​slightly higher than a preset upper limit in a certain local area. If only point-to-point comparison is performed, these points may all be marked as abnormal, but these points may be scattered and do not constitute a substantial problem. However, according to the present invention, these coordinate points with temperature values ​​exceeding the preset temperature threshold range are first identified as abnormal coordinate points. Subsequently, the system performs spatial clustering on these abnormal coordinate points. For example, if 20 abnormal coordinate points are found to be spatially adjacent, forming a circular area with a diameter of approximately 5 cm, this area is identified as an abnormally connected area. Next, for this abnormally connected area, its average temperature, maximum temperature, and temperature standard deviation are calculated. Assuming the average temperature of this area is 220°C, and the preset upper limit of the temperature threshold is 210°C; at the same time, the temperature standard deviation of this area is 5°C, and the preset uniformity standard threshold (e.g., the maximum permissible standard deviation) is 3°C. Since the average temperature and temperature standard deviation of the abnormally connected area both exceed the preset threshold range, the system will immediately trigger an early warning, indicating that there is an anomaly of overheating and poor uniformity in this specific area. This approach can effectively distinguish between genuine regional anomalies caused by process issues and random noise, thus providing more instructive early warning information.

[0097] In some embodiments of the present invention described above, a method A is proposed to calculate the true temperature of the coating by solving the polarization radiative transfer equations. However, during its implementation, the calculation results of Method A may be affected by factors such as changes in environmental parameters, model assumption errors, or sensor noise, leading to potential uncertainties or unreliability in the output true temperature. If these problems are not addressed, it may affect the accurate monitoring and quality control of the aluminum pigment coating curing process.

[0098] To address this, the present invention further proposes a mode coordination and verification step to improve the reliability of the temperature results obtained by mode A. The method further includes a mode coordination and verification step, comprising: When obtaining the true temperature using Mode A, a transient micro-polarization feature vector for verification is simultaneously generated based on the radiation signal. The verification transient micro-polarization feature vector is input into the transient micro-polarization fingerprint database corresponding to mode B for matching, to obtain the temperature reference value and the corresponding confidence level; If the deviation between the actual temperature calculated by Mode A and the temperature reference value exceeds a preset tolerance and / or the confidence level is lower than a preset threshold, then the calculation result of Mode A is determined to be unreliable, and at least one of the following operations is triggered: enabling the output result of Mode B as the final actual temperature, triggering a remeasurement and calibration of the environmental spectral radiance, and outputting a warning message with a low confidence level flag.

[0099] Specifically, the mode coordination and verification step aims to achieve cross-verification and correction of temperature measurement results by combining the advantages of Mode A and Mode B. Specifically, while Mode A performs the real temperature inversion calculation, it simultaneously uses the radiation signal acquired by the multispectral infrared polarization sensor array to generate a transient micro-polarization feature vector for verification. The generation method of this verification transient micro-polarization feature vector is the same as that in Mode B, namely, it is formed by preprocessing the radiation signal, obtaining transient radiation intensity values, combining them into an initial feature vector, calculating the dynamic rate of change, and expanding the dimensions.

[0100] Furthermore, the generated transient micro-polarization feature vector for verification is input into a pre-constructed transient micro-polarization fingerprint database of Mode B for matching. This database is pre-constructed by conducting controlled heating experiments on standard coating samples and simultaneously acquiring fully polarized radiation data, real temperature data, and microstructure evolution image data. It contains correlation information between standard transient micro-polarization feature vectors and real temperatures under different microstructure states. By matching with the fingerprint database, a temperature reference value calculated by Mode B and the confidence level of the matching result can be obtained. The confidence level can be understood as the reliability of the match, for example, expressed as the reciprocal of the matching similarity or distance.

[0101] Based on this, the system compares the actual temperature calculated by Mode A with the temperature reference value derived by Mode B. If the deviation between the two exceeds a preset tolerance range, or if the confidence level obtained by Mode B is lower than a preset threshold, it indicates that the calculation result of Mode A may be unreliable. In this case, the system will trigger at least one preset operation to ensure the accuracy of the final output temperature data. These operations may include: directly using the output result of Mode B as the final actual temperature, because Mode B is based on fingerprint database matching and may have better robustness to certain environmental disturbances; or triggering a remeasurement and recalibration of the ambient spectral radiance to correct the input parameters of Mode A; or outputting a warning message with a low confidence flag to remind operators of the potential uncertainty of the current temperature data, which may require manual intervention or further inspection.

[0102] This invention effectively addresses the limitations of single-mode measurements by introducing mode coordination and verification steps. Specifically, when Mode A performs inversion calculations based on a physical model, the accuracy of its results may be affected by environmental parameters, model parameters, or measurement noise. By synchronously generating a transient micro-polarization feature vector for verification and inputting it into the transient micro-polarization fingerprint database of Mode B for matching, an independent, data-driven temperature reference value and its confidence level can be obtained. The fingerprint database matching method of Mode B, because it directly correlates radiation characteristics with actual temperature and microstructure state, may exhibit stronger robustness under certain complex or rapidly changing conditions. When the calculation results of Mode A deviate significantly from the temperature reference value provided by Mode B, or when the matching confidence level of Mode B is low, the system can promptly identify the potential unreliability of the results from Mode A. Thus, by triggering corresponding corrective measures, such as switching to the results of Mode B, recalibrating environmental parameters, or issuing an early warning, this invention can dynamically adjust the measurement strategy, thereby avoiding erroneous temperature judgments due to errors in a single mode and significantly improving the reliability and accuracy of the entire measurement system.

[0103] Through the above technical solution, this invention significantly improves the reliability of the aluminum pigment coating curing temperature measurement method. Specifically, by introducing a mode synergy and verification mechanism, this invention can effectively detect and correct measurement errors or uncertainties that may occur in Mode A under specific conditions. This dual-mode cross-verification mechanism enables the system to provide more stable and accurate temperature data when facing complex and changing environments or workpiece conditions. When the result of Mode A is determined to be unreliable, the system can intelligently switch to the output of Mode B or prompt for calibration, thereby avoiding misjudgments caused by the limitations of a single measurement mode and greatly reducing quality risks in the production process. In addition, by outputting warning information with a low confidence flag, operators can obtain timely feedback on the reliability of the measurement data, thereby making more informed decisions and further ensuring the quality and efficiency of the aluminum pigment coating curing process.

[0104] In some preferred embodiments, it is assumed that the surface temperature of the target workpiece is 200°C during the curing process of the aluminum pigment coating. Mode A calculates a temperature of 205°C by solving the polarization radiative transfer equations. Simultaneously, the system generates a transient micro-polarization feature vector for verification and inputs it into the transient micro-polarization fingerprint database of Mode B for matching. Mode B matches a temperature reference value of 198°C with a confidence level of 0.95 (high). At this point, the deviation between the calculated result of Mode A (205°C) and the temperature reference value of Mode B (198°C) is 7°C. If the preset tolerance is 5°C, then the 7°C deviation exceeds the preset tolerance. The system will determine that the current calculation result of Mode A is unreliable. In this case, the system can trigger the use of the output result of Mode B (198°C) as the final true temperature, or trigger a remeasurement and calibration of the ambient spectral radiance to correct the input parameters of Mode A, or output a warning message with a low confidence flag to remind the operator of the potential uncertainty of the current temperature data.

[0105] For another example, suppose at another measurement moment, the temperature calculated by mode A is 210℃. The synchronously generated verification transient micro-polarization feature vector is input into the mode B fingerprint database for matching, yielding a temperature reference value of 208℃. However, the confidence level of this match is only 0.4 (low), while the preset confidence threshold is 0.6. Although the temperature deviation (2℃) between mode A and mode B is within the preset tolerance (5℃), the system will still determine that the calculation result of mode A is unreliable because the confidence level of mode B is lower than the preset threshold. In this case, the system may choose to output a warning message with a low confidence flag, prompting the operator that the reliability of the current measurement result is low and suggesting manual verification or further system diagnosis to avoid potential errors caused by low-confidence matching.

[0106] In some embodiments of the present invention, a method for measuring the curing temperature of aluminum pigment coatings is proposed, which can accurately measure the coating temperature on the surface of high-speed moving workpieces. However, in actual industrial production environments, especially on high-efficiency, continuous production lines, relying solely on the method description without a dedicated, integrated system for execution may lead to limitations such as complex operation, low efficiency, and difficulty in achieving real-time monitoring and automated control. If these problems are not addressed, the measurement process may become unstable, and data processing may be untimely, thus affecting the real-time feedback and adjustment of coating curing quality. Therefore, the present invention proposes an aluminum pigment coating curing temperature measurement system, aiming to achieve automated, efficient, and accurate execution of the above method through the synergistic effect of hardware and software, thereby ensuring quality control of the coating curing process.

[0107] An aluminum pigment coating curing temperature measurement system, the system being used to implement the above method, the system comprising: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the following modules: The radiation signal acquisition module is used to synchronously acquire radiation signals from the surface of a target workpiece with an aluminum pigment coating that is in a high-speed motion state, using a multispectral infrared polarization sensor array to collect radiation signals from the surface of the target workpiece in different infrared spectral bands and different polarization directions. The signal processing and temperature acquisition module is used to process the acquired radiation signal using at least one of the following two processing modes to obtain the true temperature of the coating: Mode A: Based on the radiation signal, the true temperature of the coating is calculated by solving the polarization radiation transmission equations. Mode B: Extract high-dimensional transient radiation features from the radiation signal, and calculate the true temperature of the coating by matching them with a pre-built transient micro-polarization fingerprint database; The temperature mapping distribution module is used to reconstruct and map all the acquired single-point temperature data, corresponding to the real-time motion information of the target workpiece, into a temperature distribution map covering the entire surface of the target workpiece through coordinate transformation and spatial interpolation. The anomaly identification and alert module is used to identify abnormal areas in the temperature distribution map based on preset temperature threshold range and uniformity standard threshold range, and to issue early warning alerts.

[0108] Specifically, the system is configured to include one or more processors and a memory. The memory stores one or more programs that, when executed by the one or more processors, enable the processors to perform the functions of the radiation signal acquisition module, signal processing and temperature acquisition module, temperature mapping distribution module, and anomaly identification and alerting module. The one or more processors can be understood as a central processing unit (CPU), graphics processing unit (GPU), application-specific integrated circuit (ASIC), field-programmable gate array (FPGA) or other programmable logic device, digital signal processor (DSP), or any hardware unit capable of executing instructions and processing data. The memory may include random access memory (RAM), read-only memory (ROM), flash memory, hard disk drive, or any other form of non-transitory computer-readable storage medium for storing the operating system, applications, and data.

[0109] The above description is merely an embodiment of the present invention and is not intended to limit the scope of protection of the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for measuring the curing temperature of an aluminum pigment coating, characterized in that, The method includes the following steps: S1. Acquiring radiation signals: For a target workpiece that is in a high-speed motion state and has an aluminum pigment coating on its surface, a multispectral infrared polarization sensor array is used to synchronously acquire radiation signals from the surface of the target workpiece in different infrared spectral bands and different polarization directions. S2. Signal Processing and Temperature Acquisition: The acquired radiation signal is processed using at least one of the following two processing modes to obtain the true temperature of the coating: Mode A: Based on the radiation signal, the true temperature of the coating is calculated by solving the polarization radiation transmission equations. Mode B: Extract high-dimensional transient radiation features from the radiation signal, and calculate the true temperature of the coating by matching them with a pre-built transient micro-polarization fingerprint database; S3. Temperature Mapping Distribution: All the acquired single-point temperature data, corresponding to the real-time motion information of the target workpiece, are reconstructed and mapped into a temperature distribution map covering the entire surface of the target workpiece through coordinate transformation and spatial interpolation. S4. Identify and alert to anomalies: Based on the preset temperature threshold range and uniformity standard threshold range, identify abnormal areas in the temperature distribution map and issue an early warning.

2. The method for measuring the curing temperature of aluminum pigment coating according to claim 1, characterized in that, In step S2, the step of calculating the true temperature of the coating by solving the polarization radiative transfer equations based on the radiation signal includes: Before the target workpiece enters the measurement area, a radiation measurement is performed on a preset reference point to obtain the spectral radiance of the reference point. Based on the known emissivity and known true temperature of the reference point, the ambient spectral radiance is calculated by back-calculating Planck's radiation law. Based on the collected radiation signals, a set of polarization radiation transmission equations based on Fresnel's equations is established, and the thermal emission radiation of the target workpiece itself and the reflected radiation from the environment are obtained by solving the equations. By pre-setting the emissivity ratio of the coating at two wavelengths, and based on the direct proportional relationship between the thermal emission radiation of the target workpiece itself and the Planck blackbody radiation function when ignoring ambient reflected radiation, an initial temperature estimate of the coating is obtained. Based on the functional relationship between the coating emissivity ratio and the actual temperature of the coating, and the ambient spectral radiance at the two wavelengths, an inversion equation set is established. Starting from the initial temperature estimate, the inversion equation set is solved using a numerical iterative solution method to obtain the actual temperature of the coating and the effective emissivity of the coating at the two wavelengths.

3. The method for measuring the curing temperature of aluminum pigment coating according to claim 2, characterized in that, The step of establishing an inversion equation set based on the functional relationship between the coating emissivity ratio and the coating's true temperature and the environmental spectral radiance at the two wavelengths includes: For the first wavelength Second wavelength ,satisfy: ； ； / =f( The functional relationship was obtained by conducting offline spectral radiometric calibration experiments on standard samples of the coating. In the formula, For coating at wavelength The measured spectral radiance, For coating at wavelength Effective emissivity at the following levels Let be the Planck blackbody radiation function. As a reference point at wavelength The measured environmental spectral radiance, This represents the actual temperature of the coating.

4. The method for measuring the curing temperature of aluminum pigment coating according to claim 2, characterized in that, The method of obtaining an initial temperature estimate of the coating by using the pre-set emissivity ratio of the coating at two wavelengths, based on the direct proportional relationship between the thermal emission radiation of the target workpiece itself and the Planck blackbody radiation function when ignoring ambient reflected radiation, includes: calculating the initial temperature estimate of the coating based on its own thermal emission radiation at the two wavelengths and in combination with the pre-set emissivity ratio of the coating at the two wavelengths using Wien's approximation law or a lookup table method.

5. The method for measuring the curing temperature of aluminum pigment coating according to claim 2, characterized in that, The numerical iterative solution method adopts either the Newton-Raphson method or the least squares method.

6. The method for measuring the curing temperature of aluminum pigment coating according to claim 1, characterized in that, In step S2, the step of extracting high-dimensional transient radiation features from the radiation signal and calculating the true temperature of the coating by matching them with a pre-constructed transient micro-polarization fingerprint database includes: The acquired radiation signal is preprocessed to obtain transient radiation intensity values. All the acquired transient radiation intensity values ​​are combined in a preset order to form an initial feature vector. The dynamic change rate of the initial feature vector in a continuous time series is calculated to obtain dynamic change information. The dynamic change information is used as a new feature dimension to expand the initial feature vector to form a higher-dimensional transient micro-polarization feature vector. In a controlled environment, a reference plate with a standard coating having an equivalent structure to the coating of the target workpiece is heated at an ultra-high heating rate simulating the actual production process. During the heating process, the following data are collected simultaneously: full polarization radiation data of the surface of the standard coating, the actual temperature data of the standard coating obtained by a contact sensor, and the microstructure evolution image data of the standard coating obtained by a microscopic imaging system. Based on the full polarization radiation data, a standard transient micro-polarization feature vector is generated. Based on the microstructure evolution image data, the microstructure state is determined. The standard transient micro-polarization feature vector, the actual temperature, and the microstructure state at the same time stamp are associated to form entries in the transient micro-polarization fingerprint database. The transient micro-polarization feature vectors acquired in real time are compared with the transient micro-polarization fingerprint database, and the true temperature of the current measurement point is calculated based on the comparison results.

7. The method for measuring the curing temperature of aluminum pigment coating according to claim 6, characterized in that, The steps of comparing the real-time acquired transient micro-polarization feature vector with the transient micro-polarization fingerprint database and calculating the true temperature of the current measurement point based on the comparison results include: Calculate the cosine similarity or Euclidean distance between the transient micro-polarization feature vectors acquired in real time and the feature vectors stored in the transient micro-polarization fingerprint database; Based on the cosine similarity being higher than a preset threshold or the Euclidean distance being lower than a preset threshold, one or more entries matching the transient micro-polarization feature vectors acquired in real time are selected from the transient micro-polarization fingerprint database. The actual temperature is calculated by using a weighted average or interpolation method to calculate the temperature values ​​corresponding to all matching entries.

8. The method for measuring the curing temperature of aluminum pigment coating according to claim 1, characterized in that, The step of identifying abnormal areas in the temperature distribution map and issuing an early warning based on a preset temperature threshold range and a uniformity standard threshold range includes: Based on the temperature distribution map, the real-time temperature value of each coordinate point in the temperature distribution map is compared with the preset temperature threshold range for the corresponding coordinate point to identify abnormal coordinate points; Spatial clustering is performed on the abnormal coordinate points to form one or more abnormal connected regions; For each abnormally connected region, the region's temperature statistical characteristics are calculated, and these characteristics are compared with a preset temperature threshold range and a uniformity standard threshold range. If the comparison result exceeds the threshold range, an early warning is triggered.

9. The method for measuring the curing temperature of aluminum pigment coating according to claim 2, characterized in that, The method further includes pattern coordination and verification steps, including: When obtaining the true temperature using Mode A, a transient micro-polarization feature vector for verification is simultaneously generated based on the radiation signal. The verification transient micro-polarization feature vector is input into the transient micro-polarization fingerprint database corresponding to mode B for matching, to obtain the temperature reference value and the corresponding confidence level; If the deviation between the actual temperature calculated by Mode A and the temperature reference value exceeds a preset tolerance and / or the confidence level is lower than a preset threshold, then the calculation result of Mode A is determined to be unreliable, and at least one of the following operations is triggered: enabling the output result of Mode B as the final actual temperature, triggering a remeasurement and calibration of the environmental spectral radiance, and outputting a warning message with a low confidence level flag.

10. A system for measuring the curing temperature of an aluminum pigment coating, the system being used to implement the method as described in any one of claims 1 to 9, characterized in that, The system includes: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the following modules: The radiation signal acquisition module is used to synchronously acquire radiation signals from the surface of a target workpiece with an aluminum pigment coating that is in a high-speed motion state, using a multispectral infrared polarization sensor array to collect radiation signals from the surface of the target workpiece in different infrared spectral bands and different polarization directions. The signal processing and temperature acquisition module is used to process the acquired radiation signal using at least one of the following two processing modes to obtain the true temperature of the coating: Mode A: Based on the radiation signal, the true temperature of the coating is calculated by solving the polarization radiation transmission equations. Mode B: Extract high-dimensional transient radiation features from the radiation signal, and calculate the true temperature of the coating by matching them with a pre-built transient micro-polarization fingerprint database; The temperature mapping distribution module is used to reconstruct and map all the acquired single-point temperature data, corresponding to the real-time motion information of the target workpiece, into a temperature distribution map covering the entire surface of the target workpiece through coordinate transformation and spatial interpolation. The anomaly identification and alert module is used to identify abnormal areas in the temperature distribution map based on preset temperature threshold range and uniformity standard threshold range, and to issue early warning alerts.

Images