A system and method for detecting film thickness and moisture content based on near-infrared transmission.

By synchronously acquiring and processing near-infrared transmission signals in the same measurement area, the interference of film thickness variation on moisture content detection was resolved, enabling synchronous detection of film thickness and moisture content, and improving the accuracy and stability of the detection.

CN123111102APending Publication Date: 2026-11-17CHANGZHOU REECHI PRECISION MEASURETECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611585630.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-14
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing methods for detecting film thickness and moisture content cannot be performed in the same measurement area and at the same sampling time, leading to shifts in moisture characteristics caused by changes in thickness, which affects the accuracy and stability of the detection.

Method used

The near-infrared transmission method is adopted, and multiple near-infrared acquisition channels are set up in the same measurement area to simultaneously acquire the transmission signals of the reference and moisture-sensitive bands. Through synchronous preprocessing and calibration, the offset of thickness changes on moisture characteristics is corrected, so as to achieve synchronous detection of thickness and moisture content.

Benefits of technology

It improves the accuracy and stability of film thickness and moisture content detection, reduces system complexity, and increases detection efficiency and integration.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123111102A_ABST
    Figure CN123111102A_ABST
Patent Text Reader

Abstract

This invention belongs to the field of near-infrared spectroscopy detection technology, specifically relating to a thin film thickness and moisture content detection system and method based on near-infrared transmission. This invention simultaneously acquires reference transmission signals and moisture transmission signals through multi-channel near-infrared transmission in the same measurement area. Thickness features are formed using the reference band, and moisture features are formed using the moisture-sensitive band. The moisture feature offset is determined based on the changes in thickness features obtained from synchronous sampling, and the moisture features are then specifically corrected. Thus, thin film thickness and moisture content detection results are obtained simultaneously in the same detection process, effectively reducing the interference of thin film thickness changes on moisture content detection. Simultaneously, it reduces the number of detection devices and the detection process, improving the real-time performance, stability, and integration of online thin film detection.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of near-infrared spectroscopy detection technology, specifically relating to a thin film thickness and moisture content detection system and method based on near-infrared transmission. Background Technology

[0002] Membrane materials are widely used in the production of lithium-ion battery separators, polymer packaging films, coated films, and functional polymer films. Film thickness and moisture content are crucial parameters affecting film quality and the performance of subsequent products. During continuous film production, film thickness can vary due to factors such as extrusion processes, stretch ratios, traction speeds, temperatures, and material distribution uniformity. The moisture content of the film's interior and surface is easily influenced by factors such as the moisture content of raw materials, ambient humidity, drying processes, and storage conditions. Therefore, rapid, continuous, and online monitoring of film thickness and moisture content during production is essential for timely detection of film quality fluctuations, adjustment of production processes, and ensuring product consistency.

[0003] Current methods for detecting film thickness and moisture content typically employ different detection devices, making it difficult to ensure that the two types of detection data correspond to the same measurement area and the same sampling time. When using near-infrared transmission for moisture content detection, changes in film thickness can cause changes in the transmission signal, which can easily lead to the additional response caused by thickness changes being misinterpreted as changes in moisture content. This results in deviations in the moisture content detection results. Furthermore, there is insufficient processing of the correspondence between changes in thickness and changes in moisture characteristics, making it difficult to determine and eliminate the moisture characteristic shift caused by changes in thickness in real time. Consequently, the accuracy, stability, and reliability of online synchronous detection of film thickness and moisture content are affected. Summary of the Invention

[0004] The purpose of this invention is to provide a thin film thickness and moisture content detection system and method based on near-infrared transmission, which can improve the accuracy, stability and online detection efficiency of moisture content detection by correcting the moisture characteristic shift caused by thickness changes.

[0005] The specific technical solution adopted by this invention is as follows: Methods for detecting film thickness and moisture content based on near-infrared transmission include: Acquire the film to be tested and use a near-infrared broadband light source to illuminate the film to be tested through transmission, forming at least two near-infrared acquisition channels corresponding to the same measurement area on the transmission side of the film to be tested; Based on at least two near-infrared acquisition channels, the near-infrared light transmitted through the thin film under test is acquired simultaneously to obtain the reference transmission signal corresponding to the reference band and the moisture transmission signal corresponding to the moisture-sensitive band. The reference transmission signal and the moisture transmission signal are preprocessed simultaneously to obtain the preprocessed reference transmission signal and the preprocessed moisture transmission signal. Thickness features are obtained based on the preprocessed reference transmission signal; Moisture characteristics are obtained based on the preprocessed moisture transmission signal; The thickness feature is correlated with the moisture feature to determine the moisture feature offset. The moisture characteristics are processed based on the moisture characteristic offset to obtain the water content characteristics corresponding to the thickness change of the film under test. The thickness detection results and moisture content detection results of the film under test are obtained based on the thickness characteristics and water content characteristics.

[0006] In a preferred embodiment, the thin film to be tested is acquired, and a near-infrared broadband light source is used to illuminate the thin film through transmission. At least two near-infrared acquisition channels corresponding to the same measurement area are formed on the transmission side of the thin film, including: The thin film to be tested is obtained and placed in the sample detection area between the near-infrared broadband light source and the receiver, so that the test area of ​​the thin film is located in the outgoing light path of the near-infrared broadband light source, and near-infrared radiation light is obtained. Near-infrared radiation is collimated and expanded to form a uniform illumination spot in the area to be tested. The uniform illumination spot covers the area to be tested and passes through the thin film to be tested, forming the corresponding transmitted light field of the area to be tested. At least two near-infrared acquisition channels are set on the transmission side of the transmitted light field, so that each near-infrared acquisition channel corresponds to the transmitted light field of the area to be measured. Different bands of near-infrared light in the transmitted light field are spectrally selected. The different bands of near-infrared light selected by each near-infrared acquisition channel are focused and converted into corresponding photoelectric signals, forming at least two near-infrared acquisition channels for the same area to be measured.

[0007] In a preferred embodiment, near-infrared light transmitted through at least two near-infrared acquisition channels is simultaneously acquired to obtain a reference transmission signal corresponding to a reference band and a moisture transmission signal corresponding to a moisture-sensitive band, including: Acquire the acquisition bands corresponding to at least two near-infrared acquisition channels, acquire the moisture absorption characteristics of each acquisition band, and determine at least one near-infrared acquisition channel as the reference acquisition channel and at least one near-infrared acquisition channel as the moisture acquisition channel. The reference band near-infrared light and the moisture-sensitive band near-infrared light transmitted from the same test area of ​​the thin film under test are obtained based on the reference acquisition channel and the moisture acquisition channel, and synchronous reception is performed at the same sampling time to obtain the reference band optical signal and the moisture-sensitive band optical signal. The reference band optical signal and the moisture-sensitive band optical signal are converted into photoelectric signals and simultaneously acquired to obtain the reference transmission signal and the moisture transmission signal.

[0008] In a preferred embodiment, the reference transmission signal and the moisture transmission signal are simultaneously preprocessed to obtain preprocessed reference transmission signal and preprocessed moisture transmission signal, including: Dark current correction is performed on the reference transmission signal and the moisture transmission signal to obtain the dark current corrected reference transmission signal and the moisture transmission signal. Based on a preset reference light intensity, the reference transmission signal and the moisture transmission signal after dark current correction are normalized to obtain a standardized transmission signal corresponding to the transmission characteristics of the film under test. The standardized transmission signal is synchronously filtered to obtain the preprocessed reference transmission signal and the preprocessed moisture transmission signal.

[0009] In a preferred embodiment, the thickness features are obtained based on the preprocessed reference transmission signal, including: The transmittance of the thin film under test in the reference band is determined based on the preprocessed reference transmission signal. The transmission attenuation characteristics of the thin film under test are determined based on the transmittance in the reference band, and the thickness parameters are determined based on the transmission attenuation characteristics. Obtain the preset thickness calibration data and combine it with the thickness parameters to obtain the thickness characteristics of the film under test.

[0010] In a preferred embodiment, moisture characteristics are obtained based on the preprocessed moisture transmission signal, including: The transmittance of the film under test in the moisture-sensitive band is determined based on the pre-processed moisture transmission signal, and the moisture-sensitive transmission characteristics are obtained. Determine the moisture absorption characteristics corresponding to the moisture-sensitive band based on the moisture-sensitive transmission characteristics. The initial moisture characteristics of the film under test are obtained based on the moisture absorption characteristics. Obtain the preset moisture calibration data and combine it with the initial moisture characteristics to obtain the moisture characteristics of the film under test.

[0011] In a preferred embodiment, the thickness feature is correlated with the moisture feature to determine the moisture feature offset, including: The sampling times corresponding to thickness and moisture features are obtained, and the thickness and moisture features corresponding to the same area to be measured are synchronously paired to obtain thickness and moisture sampling data. Based on thickness moisture sampling data, the thickness characteristic change and the corresponding moisture characteristic change are determined, and based on the thickness characteristic change and the moisture characteristic change, the additional change in moisture characteristics caused by the change in film thickness is determined. Based on the change in thickness characteristics and the additional change, the degree of shift of the moisture characteristics relative to the corresponding thickness state is determined, and the moisture characteristic shift amount is obtained.

[0012] In a preferred embodiment, the moisture characteristics are processed based on the moisture characteristic offset to obtain the water content characteristics corresponding to the thickness change of the film under test. The thickness detection results and moisture content detection results of the film under test are then obtained based on the thickness characteristics and water content characteristics, including: The moisture characteristics are corrected based on the moisture characteristic offset to obtain the corrected moisture characteristics; Obtain the change in moisture feature corresponding to the corrected moisture feature and the moisture feature offset; Based on the change in moisture characteristics, the corrected moisture characteristics are continuously sampled to obtain the moisture content characteristics corresponding to the change in the thickness of the film under test. Obtain preset thickness calibration data and preset moisture content calibration data. Determine the thickness detection result of the film to be tested based on the thickness characteristics and preset thickness calibration data, and determine the moisture content detection result of the film to be tested based on the moisture content characteristics and preset moisture content calibration data.

[0013] The present invention also provides a near-infrared transmission-based thin film thickness and moisture content detection system for use in the aforementioned near-infrared transmission-based thin film thickness and moisture content detection method, comprising: The acquisition channel module is used to acquire the film under test and uses a near-infrared broadband light source to illuminate the film under test through transmission, forming at least two near-infrared acquisition channels corresponding to the same measurement area on the transmission side of the film under test. The transmission signal acquisition module, based on at least two near-infrared acquisition channels, synchronously acquires the near-infrared light transmitted through the thin film under test, and obtains the reference transmission signal corresponding to the reference band and the moisture transmission signal corresponding to the moisture-sensitive band. The signal preprocessing module is used to simultaneously preprocess the reference transmission signal and the moisture transmission signal to obtain the preprocessed reference transmission signal and the preprocessed moisture transmission signal. The thickness feature acquisition module acquires thickness features based on the preprocessed reference transmission signal. The moisture feature acquisition module acquires moisture features based on the preprocessed moisture transmission signal. The feature offset module is used to associate thickness features with moisture features and determine the moisture feature offset. The detection results module processes the moisture characteristics based on the moisture characteristic offset to obtain the water content characteristics corresponding to the thickness change of the film under test, and obtains the thickness detection results and moisture content detection results of the film under test based on the thickness characteristics and water content characteristics.

[0014] And, a near-infrared transmission-based thin film thickness and moisture content detection terminal, including: One or more processors; A storage device on which one or more programs are stored; When one or more programs are executed by one or more processors, the one or more processors implement a method for detecting film thickness and moisture content based on near-infrared transmission.

[0015] The technical effects achieved by this invention are as follows: This invention, by employing multi-channel acquisition within the same measurement area, ensures that subsequently acquired signals of different wavelengths share the same spatial sampling basis, reducing detection discrepancies caused by varying measurement positions. Simultaneously, uniform illumination expands the effective detection area, minimizing the impact of localized pores, textures, and defects in the thin film on transmitted light intensity. Synchronous multi-wavelength acquisition ensures that the reference transmission signal and moisture transmission signal correspond to the same measurement time and the same measurement area, avoiding signal misalignment caused by thin film movement, vibration, or positional changes during traditional sequential acquisition of different wavelengths. Synchronous preprocessing of signals from different wavelengths enhances signal stability and consistency while preserving effective transmission information for each wavelength. It utilizes the transmission changes of near-infrared light after passing through the thin film to reflect changes in film thickness and equivalent optical path, obtaining thickness detection data without the need for additional mechanical thickness detection devices. This enables the synchronous acquisition of thickness and moisture content information within the same near-infrared detection system. This invention improves the consistency of moisture detection results for different thin film materials and under different detection conditions by converting optical detection features into moisture features that can be used for moisture content detection through calibration processing. By pairing thickness features and moisture features based on the same test area and the same sampling time, additional changes that occur synchronously with thickness changes are identified from changes in moisture features, avoiding the direct treatment of all such changes as moisture content changes. This reduces the interference of thin film thickness changes on the detection results of moisture-sensitive bands. By first correcting the moisture features using the moisture feature offset corresponding to the thickness change and then performing moisture content calibration, the impact of thickness changes on moisture content detection results can be reduced. This allows thickness detection and moisture content detection to be completed synchronously within the same near-infrared transmission detection system, thereby reducing the system complexity caused by independently configuring thickness detection equipment and moisture detection equipment, and improving the integration, detection efficiency, and detection stability of online thin film detection. Attached Figure Description

[0016] Figure 1 This is a flowchart of the method provided by the present invention; Figure 2 This is a system module diagram provided by the present invention. Detailed Implementation

[0017] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0018] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0019] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in a preferred embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that mutually excludes other embodiments.

[0020] Furthermore, the present invention will be described in detail with reference to the schematic diagrams. When describing the embodiments of the present invention in detail, the schematic diagrams are merely examples for ease of explanation and should not limit the scope of protection of the present invention.

[0021] Please see the appendix Figure 1 As shown, a method for detecting film thickness and moisture content based on near-infrared transmission is provided, including: S1. Obtain the film to be tested and use a near-infrared broadband light source to illuminate the film to be tested through transmission, forming at least two near-infrared acquisition channels corresponding to the same measurement area on the transmission side of the film to be tested. S2. Based on at least two near-infrared acquisition channels, the near-infrared light transmitted through the thin film under test is simultaneously acquired to obtain the reference transmission signal corresponding to the reference band and the moisture transmission signal corresponding to the moisture-sensitive band. S3. Simultaneously preprocess the reference transmission signal and the moisture transmission signal to obtain the preprocessed reference transmission signal and the preprocessed moisture transmission signal. S4. Obtain thickness features based on the preprocessed reference transmission signal; S5. Obtain moisture characteristics based on the preprocessed moisture transmission signal; S6. Associate the thickness feature with the moisture feature to determine the moisture feature offset; S7. Based on the moisture feature offset, the moisture feature is processed to obtain the water content feature corresponding to the change in the thickness of the film to be tested, and the thickness detection result and moisture content detection result of the film to be tested are obtained according to the thickness feature and the water content feature.

[0022] As described in steps S1 to S7 above, the thin film under test is illuminated by a near-infrared broadband light source, and collimation, beam expansion, and homogenization are used to cover the test area with near-infrared light. At least two near-infrared acquisition channels are set on the transmission side of the thin film under test for the same measurement area, so that different acquisition channels can acquire transmitted near-infrared light of different bands from the same test area. By using the same measurement area for multi-channel acquisition, the signals of different bands obtained later can have the same spatial sampling basis, reducing the detection differences caused by different measurement positions. At the same time, uniform illumination is used to expand the effective detection area and reduce the influence of local pores, textures, and local defects of the thin film on the transmitted light intensity. Based on the moisture absorption characteristics of the corresponding bands of each near-infrared acquisition channel, at least one acquisition channel is used as the reference acquisition channel, and at least one acquisition channel is used as the moisture acquisition channel. The reference band near-infrared light and the moisture-sensitive band near-infrared light transmitted to the same area to be measured are synchronously received, photoelectrically converted, and synchronously acquired, thereby forming a reference transmission signal and a moisture transmission signal respectively. Through multi-band synchronous acquisition, the reference transmission signal and the moisture transmission signal correspond to the same measurement time and the same area to be measured, avoiding signal misalignment caused by film movement, vibration, or position changes when acquiring different bands sequentially in the traditional way. Dark current correction is performed on the reference transmission signal and the moisture transmission signal to eliminate the influence of the dark current of the photodetector itself and the initial bias on the detection signal. Then, normalization processing is performed based on the preset reference light intensity to convert the transmission signals obtained at different sampling times to a unified signal reference. The standardized transmission signal is further synchronously filtered to reduce short-term signal disturbances caused by light source output fluctuations, thin film transport vibrations and acquisition noise. Finally, the preprocessed reference transmission signal and the moisture transmission signal are obtained. By using a synchronous preprocessing process for signals of different bands, the signal stability and consistency can be improved while retaining the effective transmission information of each band. The transmittance of the film under test in the reference band is determined by using the pre-processed reference transmission signal, and the corresponding transmission attenuation characteristics are determined based on the transmittance. Then, the transmission attenuation characteristics are converted into thickness characteristics that can characterize the thickness change of the film under test by using preset thickness calibration data. Since the reference band is used to provide a relatively stable transmission reference, the transmission change of near-infrared light after passing through the film can reflect the change of film thickness and equivalent optical path. Thickness detection basis can be obtained without the need for additional mechanical thickness detection device, thereby realizing the synchronous acquisition of thickness information and water content information in the same near-infrared detection system. The transmittance of the film under test in the moisture-sensitive band is determined based on the pre-processed moisture transmission signal, and the moisture-sensitive transmission characteristics and corresponding moisture absorption characteristics are further obtained. Then, the moisture characteristics of the film under test are formed by combining the preset moisture calibration data. Since the moisture-sensitive band has a strong response to the moisture in the film, the moisture absorption characteristics can reflect the water content of the film under test. The optical detection characteristics are converted into moisture characteristics that can be used for moisture content detection through calibration processing, which improves the consistency of moisture detection results for different film materials and under different detection conditions. The sampling times corresponding to thickness and moisture features are obtained, and the thickness and moisture features corresponding to the same test area are synchronously paired to form thickness and moisture sampling data. Further, based on the change in thickness feature and the corresponding change in moisture feature, the additional change in moisture feature caused by the change in film thickness is determined, and the moisture feature offset is determined accordingly. By pairing thickness and moisture features based on the same test area and the same sampling time, the additional change that occurs synchronously with the change in thickness is identified from the change in moisture feature, avoiding directly treating all of this change as a change in moisture content, thereby reducing the interference of film thickness change on the detection results of moisture-sensitive bands. The moisture characteristics are corrected by using the moisture characteristic offset to obtain the corrected moisture characteristics. The corrected moisture characteristics are then continuously sampled based on the change in moisture characteristics to form a moisture content characteristic corresponding to the change in the thickness of the film under test. Subsequently, the thickness characteristics are converted into thickness detection results by combining the preset thickness calibration data and the preset moisture content calibration data, and the corrected moisture content characteristics are converted into moisture content detection results. By first correcting the moisture characteristics using the moisture characteristic offset corresponding to the thickness change and then performing moisture content calibration, the influence of thickness change on moisture content detection results can be reduced. This allows thickness detection and moisture content detection to be completed synchronously within the same near-infrared transmission detection system, thereby reducing the system complexity caused by independently configuring thickness detection equipment and moisture detection equipment, and improving the integration, detection efficiency and detection stability of online film detection. By synchronously acquiring multi-channel near-infrared transmission signals in the same measurement area, reference transmission signals and moisture transmission signals are obtained separately. Thickness features are formed using the reference band, and moisture features are formed using the moisture-sensitive band. Based on the changes in thickness features obtained from synchronous sampling, the moisture feature offset is determined, and the moisture features are corrected accordingly. In this way, film thickness detection results and moisture content detection results are obtained simultaneously in the same detection process, which can effectively reduce the interference of film thickness changes on moisture content detection, while reducing the number of detection devices and detection processes, and improving the real-time performance, stability and integration of online film detection.

[0023] In a preferred embodiment, a film to be tested is acquired, and a near-infrared broadband light source is used to illuminate the film through transmission. At least two near-infrared acquisition channels corresponding to the same measurement area are formed on the transmission side of the film to be tested, including: S101. Obtain the film to be tested and place it in the sample detection area between the near-infrared broadband light source and the receiving end, so that the test area of ​​the film to be tested is located in the outgoing light path of the near-infrared broadband light source, and obtain near-infrared radiation light. S102. Collimate and expand the near-infrared radiation light to form a uniform illumination spot in the area to be tested, so that the uniform illumination spot covers the area to be tested and passes through the thin film to be tested, forming the corresponding transmitted light field of the area to be tested. S103. Set at least two near-infrared acquisition channels on the transmission side of the transmitted light field, so that each near-infrared acquisition channel corresponds to the transmitted light field of the area to be measured, and perform spectral selection on the near-infrared light of different bands in the transmitted light field, and focus the near-infrared light of different bands selected by each near-infrared acquisition channel and convert it into corresponding photoelectric signals, forming at least two near-infrared acquisition channels for the same area to be measured.

[0024] As described in steps S101 to S103 above, the film material to be tested is obtained, and the sample testing area is determined according to the film's width, thickness, and actual conveying method. For example, for lithium-ion battery separators, continuous rolls are conveyed along the production line direction, and fixed testing windows are set between the conveying rollers, with the film area corresponding to the testing window serving as the testing area. For polymer films or coated films, conveying rollers, tension rollers, and flattening support structures can be used to keep the film relatively flat in the testing area, reducing the impact of film wrinkles, swaying, and local warping on the transmitted light path. Subsequently, a near-surface support is set on one side of the film to be tested. For a broadband infrared light source, a tungsten halogen lamp is preferred as the near-infrared broadband light source. When the tungsten halogen lamp is powered on, the filament temperature rises, generating continuous radiation covering the visible to near-infrared region. For example, a tungsten halogen lamp with a rated power of 50W-150W can output continuous broadband radiation under stable operating conditions. The actual selected power is determined based on the thin film transmittance, detection distance, and detector sensitivity. In implementation, a constant current drive circuit first provides a stable operating current to the tungsten halogen lamp, for example, setting the preheating time to 30s-120s, allowing the filament to reach a relatively stable thermal equilibrium state before the constant current drive... The filament operating current is kept relatively stable, and the filament generates near-infrared radiation. A quartz protective window can be installed at the light source outlet, and an aperture can be used to limit the initial emission range, for example, a circular aperture with a diameter of 10mm-30mm, to allow the near-infrared radiation generated by the light source to enter the subsequent collimating optical components. Taking a specific embodiment as an example, a 100W tungsten halogen lamp is used as a near-infrared broadband light source. After the lamp body is preheated for 60 seconds, it enters a stable working state, and the filament generates continuous radiation covering a range of approximately 800nm-2500nm. The subsequent actual detection mainly utilizes the range of approximately 1700nm-2000nm. Near-infrared light in the m range; the membrane to be tested is placed between the light source and the receiver, so that the area of ​​the membrane to be tested is located in the emission direction of the center of the light source. At this time, the broadband near-infrared radiation generated by the tungsten halogen lamp propagates from the light source outlet to the membrane to be tested, thereby obtaining the near-infrared radiation light for subsequent collimation and beam expansion processing; the use of a tungsten halogen lamp as the broadband light source in the embodiment is also consistent with the broadband light source configuration commonly used in near-infrared transmission detection; by providing a stable broadband near-infrared radiation source for multiple detection bands simultaneously, there is no need to configure multiple narrowband light sources separately, which is beneficial to simplify the optical structure and ensure that different detection bands are generated by the same light source; A collimating lens, a beam expander, and a homogenizing optical component are sequentially arranged between the tungsten halogen lamp and the thin film under test. The collimating lens is used to collect near-infrared radiation emitted by the filament in different directions and reduce the beam divergence angle. The beam expander is used to increase the lateral size of the collimated near-infrared beam. The homogenizing optical component is used to reduce the intensity difference between the center and edge of the light spot. Taking a specific embodiment as an example, an aspherical collimating lens with a focal length of 40mm-60mm is used to collect the near-infrared radiation generated by the tungsten halogen lamp, converting the original divergent light into an approximately parallel near-infrared beam. Then, a beam expander lens group with a focal length of 50mm-100mm is used to expand the beam. The beam diameter is expanded from approximately 10 mm to approximately 40 mm–60 mm, and a frosted glass diffuser or integrating diffuser structure of approximately 50 mm × 50 mm is placed in the beam-expanding optical path. This causes the light from different positions to be redistributed after multiple scatterings, thereby reducing the light intensity concentration in the central region of the beam spot. Subsequently, a focusing or shaping lens placed at the end of the beam-expanding optical path projects the homogenized near-infrared light onto the surface of the film under test, so that the final illumination spot covers the test area. For example, if the final illumination spot is set as a circular area with a diameter of approximately 50 mm, when the width of the test area is 30 mm, the 50 mm diameter illumination spot completely covers the 30 mm area. A test area of ​​m width is defined, with the thin film positioned at the center of the light spot to prevent the test area from falling out of the illumination range due to slight lateral movement of the thin film. When near-infrared radiation shines on the thin film, some light is absorbed by the film material, some is scattered due to the internal pores and material structure, and some continues to propagate through the film towards the transmission side. These rays from the transmission side of the thin film together form the transmitted light field corresponding to the test area. Taking a lithium-ion battery separator as an example, when a 50mm uniform illumination spot covers the test area of ​​the separator, the light passes through the separator substrate and its microporous structure, forming a light field covering the test area on the transmission side of the thin film. The transmitted light distribution is significant; since the illumination area is much larger than the size of a single micropore or local structure, the transmitted light obtained by the receiver does not come from a single local location, but rather contains comprehensive transmission information from multiple micropores and material regions within a certain area; using collimating optical components to convert the broadband light source output into an approximately parallel beam, and collecting the transmitted light through focusing optical components after sample transmission, is also a common optical path design in near-infrared transmission detection; it can reduce the impact of factors such as single pores, local thickness fluctuations, and micro-wrinkles on single-point detection results, improve the spatial stability of the detection signal, and reduce baseline changes caused by uneven illumination; A near-infrared transmission protection window is first set on the transmission side to isolate dust, moisture, and other contaminants in the production environment. Then, at least two narrowband filter components are set behind the protection window, such as a first narrowband filter and a second narrowband filter. The center wavelength of the first narrowband filter can be set to approximately 1720 nm, and the center wavelength of the second narrowband filter can be set to approximately 1940 nm. The area around 1720 nm serves as a reference band with weak moisture absorption, providing a transmission reference less affected by moisture changes. The area around 1940 nm serves as a moisture-sensitive band, acquiring transmission information more sensitive to the water content of the thin film. The actual filter bandwidth can be determined based on the light source power, detector response range, and detection... Sensitivity settings, for example, using a half-width at half maximum (WHM) of 10nm-30nm, such as a first narrowband filter with a center wavelength of 1720nm and a bandwidth of approximately 20nm, and a second narrowband filter with a center wavelength of 1940nm and a bandwidth of approximately 20nm; a focusing lens is set behind each narrowband filter, for example, a plano-convex lens with an effective aperture of approximately 20mm-30mm and a focal length of approximately 20mm-35mm, to focus the narrowband near-infrared light passing through the corresponding filter onto the photosensitive surface of the corresponding InGaAs photodetector; for example, the first channel uses the first plano-convex lens to focus near-infrared light near 1720nm onto the first InGaAs photodiode, and the second channel uses the second plano-convex lens to focus near-infrared light near 1940nm onto the photosensitive surface of the corresponding InGaAs photodetector; Near-infrared light is focused onto the second InGaAs photodiode. To ensure that both channels correspond to the same test area, during installation, visible auxiliary light or low-power near-infrared positioning light is first used to determine the center of the test area. Then, the lateral positions of the filters, focusing lenses, and detectors of the two channels are adjusted so that the centers of the beams received by both channels fall on the center area of ​​the 50mm uniform illumination spot, and the receiving areas of the two detectors correspond to the same transmitted light field. For example, the optical axis spacing of the two channels is set to 5mm-20mm, but through the large area of ​​the uniform illumination spot at the front end and the corresponding collection field of view, the information actually received by both channels comes from the same 50mm test area, rather than detecting two different positions separately. To avoid stray light interference between channels, a black light-shielding partition can be installed between the two channels. The height of the partition can be set to 10mm-30mm, and the inner wall of the acquisition cavity can be treated to remove light, so that the near-infrared light selected by the 1720nm channel mainly enters the first detector, and the near-infrared light selected by the 1940nm channel mainly enters the second detector. To ensure a stable time reference for subsequent synchronous sampling, the two photodetectors can be connected to the same synchronous acquisition circuit and triggered by the same sampling clock. For example, the sampling frequency can be set to 1000Hz, that is, the photoelectric signals of the first and second channels are acquired simultaneously every 1ms, forming a reference acquisition channel and a moisture acquisition channel corresponding to the same area to be measured.With the above setup, the broadband transmitted light generated from the same test area is distributed to at least two different wavelength acquisition channels. The 1720nm channel provides reference transmission information, and the 1940nm channel provides moisture-sensitive transmission information. Both channels spatially correspond to the same test area and can output corresponding photoelectric signals at the same sampling time. Therefore, the thickness and moisture features can be extracted directly using the synchronous changes between the two channels, avoiding the time misalignment caused by using moving filters or sequentially scanning different wavelengths. Using a fixed filter for wavelength selection and cooperating with a photodetector for near-infrared detection also has a high acquisition speed. Related near-infrared transmission detection schemes have already adopted fixed wavelength selection and photodetector structures to reduce the speed limitations caused by mechanical scanning.

[0025] In a preferred embodiment, based on at least two near-infrared acquisition channels, the near-infrared light transmitted through the thin film under test is simultaneously acquired to obtain a reference transmission signal corresponding to a reference band and a moisture transmission signal corresponding to a moisture-sensitive band, including: S201. Obtain the acquisition bands corresponding to at least two near-infrared acquisition channels, obtain the moisture absorption characteristics of each acquisition band, and determine at least one near-infrared acquisition channel as the reference acquisition channel and at least one near-infrared acquisition channel as the moisture acquisition channel. S202. Based on the reference acquisition channel and the moisture acquisition channel, acquire the reference band near-infrared light and the moisture-sensitive band near-infrared light transmitted through the same test area of ​​the film under test, and perform synchronous reception at the same sampling time to obtain the reference band optical signal and the moisture-sensitive band optical signal. S203. Convert the reference band optical signal and the moisture-sensitive band optical signal into photoelectric signals and collect them synchronously to obtain the reference transmission signal and the moisture transmission signal.

[0026] As described in steps S201 to S203 above, the center wavelength and bandwidth of the narrowband filters configured in each channel are read. For example, if the first acquisition channel is equipped with a narrowband filter with a center wavelength of 1720nm and a bandwidth of 20nm, then the acquisition band of the first acquisition channel is determined to be 1710nm-1730nm; if the second acquisition channel is equipped with a narrowband filter with a center wavelength of 1940nm and a bandwidth of 20nm, then the acquisition band of the second acquisition channel is determined to be 1930nm-1950nm. The center wavelength and bandwidth of the filter can be directly derived from the factory calibration parameters of the filter, or they can be determined after installation using a standard near-infrared light source and a standard... The wavelength detection equipment performs actual band calibration by inputting near-infrared light of known wavelengths into the corresponding acquisition channel. For example, within the range of 1700nm-2000nm, inputting light point-by-point at wavelength intervals of 5nm or 10nm, and recording the peak position and effective response range of the detector's output signal as a function of wavelength. The actual center wavelength is determined by the peak position, and the actual effective bandwidth is determined by the positions on both sides of the peak that reach a certain proportion of the peak value. For example, if actual testing shows that the first channel has the maximum output signal at 1720nm and maintains an effective response in the range of 1710nm-1730nm, then 1710nm-1730nm is determined as the acquisition band of the first acquisition channel; the second channel... The signal output is strongest at 1940nm, and it maintains an effective response in the 1930nm-1950nm range. Therefore, 1930nm-1950nm is determined as the acquisition band for the second acquisition channel. After obtaining each acquisition band, the moisture content of the film under test is determined by establishing a moisture calibration sample and performing near-infrared transmission testing, rather than directly using a general moisture absorption band as a fixed basis. First, several samples with basically the same thickness are selected from the same batch of films under test, such as lithium-ion battery separators with a thickness of 20μm-30μm. The moisture content of the samples is changed by using a constant temperature and humidity chamber, for example, setting the temperature to 25℃ and the relative humidity to 10% and 3%, respectively. The samples were dried at 0%, 50%, 70%, and 90% relative humidity for 12-24 hours each to achieve a relatively stable hygroscopic state. The actual moisture content was then determined by precise weighing. For example, the sample was first dried at 105℃ for 2-4 hours to obtain its dried mass. The mass of the sample under different humidity conditions was then weighed, and the actual moisture content was determined based on the difference between the wet and dry masses. For instance, if a sample had a dry mass of 10g and a stable mass of 10.05g under 50% relative humidity, the 0.05g increase in mass could be considered the corresponding absorbed moisture mass, and further, 0.05g divided by 10g would yield 0.A 5% water content was used. For each sample with different water content, transmission signals in the ranges of 1710nm-1730nm and 1930nm-1950nm were collected using the two acquisition channels mentioned above, and the transmission intensity of the corresponding bands under each water content was compared. For example, in the first acquisition band, when the water content increased from 0.2% to 1%, the transmission signal only changed slightly, while in the second acquisition band, the transmission signal decreased significantly. Therefore, the first acquisition band can be determined as the reference band for weak water absorption, and the second acquisition band can be determined as the water-sensitive band. This determination method has practical basis because water has significant absorption in the near-infrared region around 1450nm and 1940nm, with 1940nm usually showing a strong water response. However, the specific film substrate may produce additional absorption in the corresponding band. Therefore, determining the final band through actual measurement of the film under test can reduce the material... The influence of matrix differences; for example, for polymer membranes, 1720nm is used as a weak moisture absorption reference band, and 1940nm as a moisture-sensitive band; if actual calibration reveals a significant moisture response at 1720nm, the reference channel can be adjusted to around 1800nm, or a band with a weaker moisture response can be selected from the 1700nm-1850nm range; ultimately, based on the above measurement results, the acquisition channel with a weak moisture response that can stably reflect changes in film transmission is determined as the reference acquisition channel, and the acquisition channel with a significant moisture response that can produce stable transmission changes with changes in water content is determined as the moisture acquisition channel; in this way, the reference acquisition channel and the moisture acquisition channel are not arbitrarily designated, but are jointly determined by the actual channel band parameters and the moisture calibration test results of the film under test, which allows for adjustment of the detection band for different film materials and improves the applicability of the detection scheme. In a uniform illumination spot with a diameter of approximately 50 mm, the detection area of ​​the thin film under test is located at the center of this 50 mm illumination spot. A first and second acquisition channel are set on the transmission side, with the receiving field of view of both channels covering the central portion of the 50 mm area. For example, the actual effective receiving area of ​​both channels is set to 30 mm × 30 mm. By adjusting the optical axes of the two channels, the centers of the two effective receiving areas coincide with the center of the area under test, ensuring that the first and second acquisition channels acquire the transmitted light generated by the same area under test. After the thin film under test is irradiated by broadband near-infrared light, part of the light is absorbed by the thin film material, part is affected by the internal pore structure and material structure of the thin film, and the remaining part passes through the thin film. On the transmission side, a transmitted light field containing multiple wavelength components is formed on the thin film transmission side; the narrowband filter in the first acquisition channel only allows near-infrared light in the range of, for example, 1710nm-1730nm to pass through, thus extracting reference band near-infrared light from the common transmitted light field; the narrowband filter in the second acquisition channel only allows near-infrared light in the range of, for example, 1930nm-1950nm to pass through, thus extracting moisture-sensitive band near-infrared light from the same transmitted light field; to achieve synchronous reception at the same sampling moment, the photodetectors of the two channels are connected to the same synchronous sampling trigger circuit, and the two channels are simultaneously triggered to start acquisition by the same clock signal, for example, setting the sampling frequency to 1000Hz, that is, generating a sampling every 1ms. The system triggers a signal to simultaneously read the signals from the first and second channels at 1ms, and then simultaneously reads the signals from both channels again at 2ms, continuing this process continuously. For example, at the 100th sampling moment, the first channel acquires the transmitted light near 1720nm from the test area at that moment, while the second channel acquires the transmitted light near 1940nm from the same test area at the same moment. These two signals form a set of corresponding sampling data in time. The optical signal refers to the information on the change in light intensity that enters the corresponding detector after being selected by a narrowband filter. In actual hardware, this can be received by the photosensitive surface of the photodetector and manifested as light power or light intensity changing over time. For example, the first channel receives the transmitted light near 1720nm at a certain sampling moment. The near-transmitted light power is 0.8mW, and the second channel simultaneously receives a transmitted light power of 0.5mW near 1940nm. At this moment, a set of reference band optical signals and moisture-sensitive band optical signals are formed. Since the two signals come from the same test area and are obtained at the same time, the spatial changes in the two channels can be synchronously reflected when the film undergoes lateral movement, longitudinal movement, vibration, or local thickness changes. This avoids the band signal difference caused by the thickness of 25μm and 26μm at one location when using two independent locations for detection, and also avoids the problem that the two signals no longer correspond to the same film position when collecting 1720nm and 1940nm signals sequentially because the film has moved. InGaAs photodetectors are set up for two detection bands near 1720nm and 1940nm respectively, because InGaAs detectors can cover the aforementioned near-infrared band. When the first acquisition channel receives transmitted light near 1720nm, the first InGaAs detector generates a corresponding photocurrent based on the incident light power, for example, 80μA photocurrent at an incident light power of 0.8mW. When the second acquisition channel receives transmitted light near 1940nm, the second InGaAs detector generates, for example, 50μA photocurrent. Subsequently, the photocurrents are converted into voltage signals by transimpedance amplifier circuits, for example, in the first channel. Using a 10kΩ feedback resistor, the 80μA photocurrent is converted into a 0.8V voltage signal after transimpedance conversion; the second channel also uses a 10kΩ feedback resistor, so the 50μA photocurrent generates a 0.5V voltage signal. Since the actual responsivity, dark current, and amplifier gain of different detectors may vary, channel calibration can be performed under no-light and standard reference light conditions before formal testing: First, turn off the near-infrared broadband light source, and collect, for example, 1000 samples under no-light conditions. Calculate the average output value of each channel as the dark current reference for that channel; for example, the average output of the first channel for 1000 samples is 0.02V. The average output of the second channel is 0.03V. In subsequent sampling, the corresponding reference is subtracted from the original outputs of both channels. Then, a stable standard near-infrared reference light is input to both channels; for example, 1720nm and 1940nm standard light are used to generate 1mW of reference light power respectively, and the corresponding outputs of the two channels are recorded to form channel gain calibration data. After the above calibration, the output signals of the two channels are then passed through an analog low-pass filter for high-frequency noise suppression, for example, setting the filter cutoff frequency in the range of 300Hz-450Hz. If the system sampling frequency is 1000Hz, the cutoff frequency is set to 400Hz. High-frequency noise, significantly faster than the rate of change of the detection signal, is attenuated while retaining the rapid changes that need to be detected during continuous film transport. Then, the analog voltage signals from both channels are simultaneously input to a multi-channel synchronous analog-to-digital converter (ADC), such as a 16-bit or 24-bit ADC, and synchronously converted using the same sampling clock. Taking a 16-bit ADC with a 0V-5V input range as an example, its digital value range is 0-65535. When the first channel inputs 0.8V, the digital conversion result is approximately 0.8 divided by 5 and then multiplied by 65535, yielding approximately 10486. When the second channel inputs 0.5V, the corresponding digital conversion result is approximately 0.Dividing 5 by 5 and then multiplying by 65535 yields approximately 6554. At the same sampling moment, the digital signal 10486 corresponding to the first channel and the digital signal 6554 corresponding to the second channel are obtained and recorded as the reference transmission signal and moisture transmission signal at that sampling moment, respectively. Continuing this synchronous acquisition at a sampling frequency of 1000Hz, for example, acquiring signals continuously for 1 second, yields 1000 reference transmission signals and 1000 moisture transmission signals, which are then stored according to the same sampling sequence number. For example, the reference transmission signal and the moisture transmission signal at the first sampling moment correspond to the same test area, and the two signals at the second sampling moment also correspond to the next sampling state of the same test area. This stably converts the intensity changes of different wavelengths in the transmitted light field into digital transmission signals with a unified sampling time base. This not only completes the conversion of optical signals to digital detection signals but also maintains the temporal correspondence between the two detection wavelengths through synchronous sampling, which is beneficial for improving the spatial correspondence, temporal consistency, and detection stability during high-speed online detection of continuous thin films.

[0027] In a preferred embodiment, the reference transmission signal and the moisture transmission signal are simultaneously preprocessed to obtain preprocessed reference transmission signal and preprocessed moisture transmission signal, including: S301. Perform dark current correction on the reference transmission signal and the moisture transmission signal to obtain the dark current corrected reference transmission signal and the moisture transmission signal. S302. Based on the preset reference light intensity, the reference transmission signal and the moisture transmission signal after dark current correction are normalized to obtain the standardized transmission signal corresponding to the transmission characteristics of the film under test. S303. The standardized transmission signal is synchronously filtered to obtain the preprocessed reference transmission signal and the preprocessed moisture transmission signal.

[0028] As described in steps S301 to S303 above, before formally testing the film under test, the near-infrared broadband light source is first turned off, or a light-shielding mechanism placed between the light source and the receiver is used to block the near-infrared light entering the two near-infrared acquisition channels, so that both the reference acquisition channel and the moisture acquisition channel are in a state of no effective incident light; then, keeping the detector gain, integration time, amplification factor, sampling frequency, and analog-to-digital conversion parameters unchanged during formal testing, 1000 baseline data points from the reference channel and 1000 baseline data points from the moisture channel are continuously acquired at a sampling frequency of 1000Hz, and then the data are processed for 100... Summing the zero data points and dividing by 1000 yields the reference dark current values ​​for the baseline channel and the moisture channel. For example, in a single dark current acquisition, if the average of the 1000 sampled values ​​obtained from the analog-to-digital conversion of the baseline channel is 0.02V, and the average of the 1000 sampled values ​​of the moisture channel is 0.03V, then 0.02V is used as the reference dark current value for the baseline channel, and 0.03V is used as the reference dark current value for the moisture channel. During formal testing, for example, if the acquisition voltage corresponding to the reference transmission signal is 0.82V and the acquisition voltage corresponding to the moisture transmission signal is 0.53V at a certain sampling moment, then... Subtracting 0.02V from 0.82V yields a dark current-corrected reference transmission signal of 0.8V, and subtracting 0.03V from 0.53V yields a dark current-corrected moisture transmission signal of 0.5V. For continuous operation, the same subtraction process is performed at each sampling time, thus forming continuous sequences of dark current-corrected reference transmission signals and dark current-corrected moisture transmission signals. To avoid deviations caused by long-term fixation of the dark current reference value due to changes in ambient temperature, the dark current update cycle can be further set, for example, every 10 minutes, 30 minutes, or 60 minutes. It can perform a single dark current acquisition and then re-acquire 500-5000 dark current samples for averaging. During the detection period, it can also update according to a preset temperature range. For example, if the detector temperature changes by more than 2°C, dark current acquisition can be re-executed. For scenarios where detection cannot be interrupted on the production line, dark current acquisition can be performed before detection begins, after roll change, or after the equipment is started and preheated. It can remove the detector dark current and the zero-point bias of the acquisition circuit from the original signals of the two channels. At the same time, the two channels use the same sampling clock and the same dark current acquisition period to keep the reference transmission signal and the moisture transmission signal in time synchronized. Without placing the film under test, the near-infrared broadband light source is operated according to the formal detection state, for example, the light source is preheated for 60 seconds to stabilize the output. Then, keeping the optical position, filter, detector gain, and sampling frequency of the two acquisition channels unchanged, 1000 thin-film-free reference signals are continuously acquired. For example, if the reference acquisition channel obtains 1000 sample values ​​in the thin-film-free state, and its average acquisition voltage is 1V, then 1V is determined as the preset reference light intensity of the reference channel. If the moisture acquisition channel continuously acquires 1000 sample values ​​under the same conditions, and its average acquisition voltage is 0.8V, then 0.8V is determined as the preset reference light intensity of the moisture channel. The preset reference light intensity does not necessarily have to be expressed directly in physical units. Optical power can also be measured using a reference signal value obtained through photoelectric conversion and analog-to-digital conversion. This reference value must originate from data actually acquired by the same detection channel under specified reference conditions. To improve the stability of the reference value, 3-10 sets of reference acquisitions are performed consecutively, with 500-5000 samples acquired in each set. The average values ​​of each set are then further averaged. For example, if three reference values ​​are 0.998V, 1.002V, and 1V, 0.998V is first added to 1.002V to obtain 2V, then added to 1V to obtain 3V, and finally divided by 3 to obtain 1V. This 1V is used as the reference value for the reference channel. A similar approach is used to obtain a reference value of, for example, 0.8V for the moisture channel. After obtaining the reference value, the signal after dark current correction is... Normalization is performed by taking the dark current correction signal at each sampling moment as the current effective transmission signal and comparing it with the reference signal of the corresponding channel. For example, at a certain sampling moment, the dark current correction signal of the reference channel is 0.8V, and the reference value of the reference channel is 1V. Then, 0.8V is divided by 1V to get 0.8, which is the normalized reference transmission signal at that sampling moment. The dark current correction signal of the moisture channel is 0.5V, and the reference value of the moisture channel is 0.8V. Then, 0.5V is divided by 0.8V to get 0.625, which is the normalized moisture transmission signal at that sampling moment. Subsequent synchronous sampling points are processed in the same way. For example, at a sampling frequency of 1000Hz, 1000 sets of normalized signals are generated per second. The system includes a reference transmission signal and 1000 sets of standardized moisture transmission signals, with the 100th reference signal and the 100th moisture signal corresponding to the same test area and the same sampling time. The reacquisition period for the reference light intensity can be set to 5-60 minutes, for example, a thin-film-free reference acquisition can be performed every 30 minutes. When a change in the light source output is detected to exceed a preset range, such as a change in the average value of two adjacent reference acquisitions exceeding 2%-5%, or after the light source is replaced, the detector is replaced, or the optical path is readjusted, the reference light intensity is also reacquisitioned. The detector output voltage is converted to the transmission level relative to the thin-film-free reference state, for example, the reference channel is converted from 0.8V to a standardized transmission value of 0.8, and the moisture channel is converted from 0.5V to 0.A standardized transmittance value of 625 is used to reduce the impact of absolute light intensity variations at different detection times on subsequent thickness and moisture feature extraction. The filtering parameters are determined based on the transport speed of the film under test, the expected rate of change to be detected, and the sampling frequency, rather than directly using fixed parameters. For example, if the system uses 1000Hz sampling and the film under test moves at a speed of 1m / s along the transport direction, and if a thickness or moisture content change of more than 5mm is retained, the corresponding change time is approximately 5ms. Therefore, the range of signal changes that need to be retained can be set in the range of tens to hundreds of Hz, and electronic noise significantly higher than this range can be suppressed. In this embodiment, a low-pass filter with a cutoff frequency of 300-450Hz is used, for example, setting the cutoff frequency to 400Hz. z, while maintaining the same filtering cycle, the same number of sampling points, and the same trigger clock for both the reference channel and the moisture channel; alternatively, a 5-point, 7-point, or 11-point moving average can be used. For example, when using a 5-point moving average, the sum of the first to fifth consecutive standardized sample values ​​is divided by 5 to obtain the smoothed value corresponding to the third sampling position. For instance, if the standardized transmittance values ​​of the five consecutive reference channels are 0.8, 0.804, 0.798, 0.802, and 0.796, adding 0.8 and 0.804 gives 1.604, adding 0.798 gives 2.402, and adding 0.802 gives 3. 0.204, plus 0.796, equals 4. Dividing by 5 gives 0.8, thus obtaining a filtered reference transmission signal of 0.8. For the moisture channel, for example, five consecutive standardized transmission values ​​are 0.625, 0.629, 0.623, 0.627, and 0.626. Adding these five numbers sequentially gives 3.13, then dividing by 5 gives 0.626, resulting in a filtered moisture transmission signal of 0.626. In practical implementation, if a digital low-pass filter is used during equipment calibration, tests can be conducted using a standard light source and a film of known thickness, gradually setting the cutoff frequency to 100Hz, 200Hz, 300Hz, etc. The signal fluctuation amplitude before and after filtering, as well as the response change when a film of known thickness passes through the detection area, are recorded at Hz, 400Hz, 500Hz, and 600Hz, respectively. Then, a parameter that can reduce high-frequency noise without significantly weakening the actual thickness change response is selected. For example, after testing, the noise fluctuation at 400Hz is reduced from ±0.02 to ±0.006, while the effective transmission change generated by the film of known thickness still maintains more than 95% of the original change amplitude. Therefore, 400Hz is determined as the filtering cutoff frequency of this embodiment. For continuous production lines, the parameter switching range can also be set according to the conveying speed, for example, when the conveying speed is 0.At speeds of 2-2 m / s, the low-pass cutoff frequency is set to 100-500 Hz, and a specific value is selected through calibration experiments. The reference transmission signal and the moisture transmission signal are synchronously filtered using the same sampling clock. For example, the 100th reference signal and the 100th moisture signal simultaneously enter their respective filtering processes, and the 101st reference signal and the 101st moisture signal simultaneously enter the next processing position, thus avoiding time misalignment between the two channels due to different filtering delays. Finally, preprocessed reference transmission signals and preprocessed moisture transmission signals, corresponding one-to-one with the original sampling times, are obtained.

[0029] In a preferred embodiment, thickness features are obtained based on the preprocessed reference transmission signal, including: S401. Determine the transmittance of the thin film under test in the reference band based on the pre-processed reference transmission signal. S402. Determine the transmission attenuation characteristics of the thin film under test based on the transmittance under the reference band, and determine the thickness parameters based on the transmission attenuation characteristics. S403. Obtain the preset thickness calibration data and, in combination with the thickness parameters, obtain the thickness characteristics corresponding to the film to be tested.

[0030] As described in steps S401 to S403 above, a preset reference light intensity corresponding to the reference acquisition channel is obtained. This preset reference light intensity has already been actually acquired under the condition that the film under test is removed, and the near-infrared broadband light source, optical path, filter, detector gain, and sampling parameters remain unchanged. For example, the reference light intensity of the reference acquisition channel is averaged after 1000 samplings to obtain 1V. Then, the preprocessed reference transmission signal corresponding to the film under test is acquired. For example, 0.8V is obtained at a certain sampling moment. Since this 0.8V has already undergone dark current correction, reference light intensity normalization, and filtering, it can be directly compared with the reference state. The effective reference transmission signal under the test film state is mapped to the effective reference signal under the no-film reference state to obtain 0.8, which means that the effective light signal after passing through the test film in the reference band is about 80% of that in the reference state. If it needs to be expressed as a percentage, 0.8 is multiplied by 100 to get 80%, which means that the reference band transmittance of the test film at this sampling position is 80%. For continuous detection, the same processing is performed for each sampling time. For example, the preprocessed reference transmission signals corresponding to 5 consecutive sampling times are 0.8, ... With reference values ​​of 79, 0.78, 0.77, and 0.76, and a reference state of 1, the reference band transmittances are 80%, 79%, 78%, 77%, and 76%, respectively, forming transmittance data that varies continuously along the film transport direction. The reference band is selected to be insensitive or relatively insensitive to changes in film moisture content. For example, in the aforementioned embodiment, a channel with a center wavelength of approximately 1720 nm can be used. However, the specific band cannot be determined directly based on fixed values; it should be selected through band screening using actual film samples. For example, if film samples from the same batch but with different moisture contents are prepared, at, for example, 10%, 30%, and 50%... The film was equilibrated for 12–24 hours at 70% and 90% relative humidity, respectively. Then, the transmission signals of each candidate wavelength band were collected. If the transmittance near 1720 nm changed only slightly with the change in water content, while the change near 1940 nm was significant, then the channel corresponding to 1720 nm was used as the reference channel. The transmittance of the reference wavelength band obtained in this way is mainly used to characterize the film's own thickness and equivalent optical path change, rather than mainly to characterize the change in moisture content. The standardized electrical signal was further converted into comparable transmittance, so that the detection results at different detection times, different light source output levels, and different film positions have a uniform scale. When near-infrared light passes through a thin film, the material absorption of the film, the internal optical path, and the change in optical path caused by variations in film thickness will cause the transmitted light to attenuate relative to the reference state without a thin film. Therefore, the logarithm of the reciprocal of the transmittance can be used as the transmission attenuation characteristic; that is, the lower the transmittance, the greater the corresponding transmission attenuation characteristic. If the transmittance of the reference band at a certain location is 80%, first convert 80% to 0.8, then determine its reciprocal, obtaining 1.25, and then take the logarithm to base 10, obtaining approximately 0.097. This 0.097 can be used as the transmission attenuation characteristic corresponding to that location. If the transmittance at another location is... If the transmittance is 70%, first convert 70% to 0.7, obtaining its reciprocal approximately 1.429. Then take the logarithm to base 10, yielding approximately 0.155. Therefore, the transmission attenuation characteristic at the second position is greater than that at the first position, indicating a higher degree of optical attenuation at that position. In embodiments where the logarithmic form is not directly used, the difference between the reference transmittance and the current transmittance can also be used directly as the transmission attenuation characteristic. For example, if the reference transmittance is 100% and the current transmittance is 80%, then 100% minus 80% gives a 20 percentage point transmission attenuation. However, for scenarios with a large range of film thickness, the logarithmic transmission attenuation characteristic is preferred. This is because the decrease in transmitted light intensity with increasing optical path is a common phenomenon in near-infrared transmission measurements, and Beer-Lambert-type processing typically uses the logarithm of transmittance to characterize the degree of absorption. After obtaining the transmission attenuation characteristics, the thickness parameter is further determined based on these characteristics. The thickness parameter is not directly claimed as the final physical thickness, but rather as an intermediate parameter characterizing the thickness or equivalent optical path change corresponding to the current transmission attenuation. This avoids directly identifying the optical attenuation value as the actual thickness before actual thickness calibration. For example, through preliminary calibration experiments, it was found that a 10 μm thick film corresponds to a transmission attenuation... The characteristic value is approximately 0.035, 20 μm corresponds to approximately 0.07, 30 μm corresponds to approximately 0.105, 40 μm corresponds to approximately 0.14, and 50 μm corresponds to approximately 0.175. In this embodiment, if the transmission attenuation characteristic of the film under test is 0.105, the corresponding thickness parameter can be preliminarily determined to be 30 μm. If the transmission attenuation characteristic is 0.122, it lies between 0.105 and 0.14 and is interpolated according to the degree of change between two adjacent calibration points. For example, 0.122 is 0.017 more than 0.105, and the difference between 0.105 and 0.14 is 0.035, with 0.017 accounting for approximately 0.Since the thickness is half of 0.35, the thickness parameter can be determined to be approximately 35 μm. However, for actual thin films such as lithium-ion battery separators with porous structures, scattering effects, or surface structure variations, it cannot be simply assumed that transmission attenuation and thickness are always strictly linearly correlated. This is because the scattering medium may cause the transmission response to deviate from the ideal Beer-Lambert law. Therefore, setting actual thickness calibration data and calibrating the thickness parameter is more suitable for online detection of actual thin films than directly applying theoretical formulas. Further converting the original transmittance into a transmission attenuation characteristic directly corresponding to the degree of optical attenuation, and then using this transmission attenuation characteristic to obtain the thickness parameter, allows thickness information to be extracted step-by-step from the near-infrared transmission signal, avoiding the influence of the detector output scale when directly using the original voltage signal to determine the thickness. Prepare a set of standard film samples that are as consistent as possible with the material, manufacturing process, and surface condition of the film to be tested. For example, for lithium-ion battery polyolefin separators, select separators from the same batch on the actual production line as calibration samples. Obtain the actual thickness of each calibration sample using a calibrated contact thickness gauge, micrometer, or other traceable thickness measurement equipment. For example, prepare nine sets of calibration samples with thicknesses of approximately 10μm, 15μm, 20μm, 25μm, 30μm, 35μm, 40μm, 45μm, and 50μm. For each set, select at least 10 different locations for actual thickness measurement to obtain reference thickness data for each sample. For example, for a 30μm specification separator, the 10 measurement locations yielded 29.7μm, 30.1μm, and 50μm thicknesses, respectively. The thicknesses of the calibration sample were 10 μm, 30 μm, 29.9 μm, 30.2 μm, 30.1 μm, 29.8 μm, 30 μm, 30.1 μm, and 29.9 μm. These values ​​were added together to obtain 299.8 μm, and then divided by 10 to obtain 29.98 μm. This 29.98 μm was used as the actual reference thickness of the calibration sample. The calibration sample was then placed sequentially in the near-infrared detection area, and its reference transmission signal was acquired under the same light source, optical path, reference band, detector gain, and sampling parameters as the actual detection. The corresponding reference band transmittance and transmission attenuation characteristics were obtained. For example, the transmittance values ​​for the five calibration samples of 10 μm, 20 μm, 30 μm, 40 μm, and 50 μm were 0.035, 0.07, 0.105, and 0.07, respectively. The transmission attenuation characteristics of 14 and 0.175 are obtained, and each transmission attenuation characteristic is paired with the corresponding actual reference thickness to form thickness calibration data. To improve the reliability of the calibration data, 500-5000 continuous data points are collected at each thickness level, and the continuous data is divided into multiple effective sampling segments according to the detection position, for example, each segment contains 100 sampling points. Then, the average of each segment is calculated to obtain the representative transmission attenuation characteristics of each thickness level. If the transmission attenuation characteristics obtained at 10 positions in the same thickness level are 0.103, 0.105, 0.104, 0.106, 0.105, 0.104, 0.107, 0.105, 0.103, and 0.106, respectively, then these values ​​are added together. The result is 1.048, which is then divided by 10 to obtain 0.1048. This 0.1048 is paired with the actual thickness of approximately 30 μm to form a thickness calibration data point. After obtaining multiple sets of calibration data, a thickness calibration table is formed in ascending order of thickness. For example, a transmission attenuation characteristic of 0.035 corresponds to 10 μm, 0.07 to 20 μm, 0.105 to 30 μm, 0.14 to 40 μm, and 0.175 to 50 μm. For cases where the actual data is not perfectly linear, piecewise interpolation is performed between adjacent calibration points. For example, if the transmission attenuation characteristic corresponding to the thickness parameter of the film to be measured is 0.122, since 0.122 lies between 0.105 and 0.14, interpolation is performed between the two calibration points of 30 μm and 40 μm.The difference between 122 and 0.105 is 0.017, while the difference between 0.14 and 0.105 is 0.035. Dividing 0.017 by 0.035 yields approximately 0.486. Multiplying the interval increment of 10μm (between 30μm and 10μm) by 0.486 gives approximately 4.86μm. Finally, adding 30μm and 4.86μm gives approximately 34.86μm. Therefore, the thickness characteristic at this location is determined to be approximately 34.9μm. If the calibration data exhibits significant nonlinearity, simple linear interpolation is not performed. Instead, adjacent calibration points are divided into multiple thickness intervals, each using a mapping method determined by the actual calibration data. This ensures that the thickness characteristic better matches the optical response of the actual thin film. The thickness calibration data originates from data with actual thickness reference values. The standard thin film sample, the actual transmission signal acquired by the same near-infrared detection device, and the transmission attenuation characteristics obtained as described above are used, rather than a set of data without a source that is pre-entered manually. Furthermore, when the light source, filter, detector, optical path position, or the material of the thin film under test changes, the calibration data acquisition process should be re-executed. For example, the above nine sets of thickness samples should be re-measured after replacing the detector to avoid the original calibration data becoming invalid due to equipment changes. Existing research has shown that near-infrared transmission can be used for polymer thin film thickness measurement, and the correspondence between transmission response and thickness can be established through actual calibration. For online polymer thin film detection, actual calibration is also an important step in obtaining thickness results. Ultimately, the obtained thickness parameters serve as thickness characteristics with actual length units.

[0031] In a preferred embodiment, moisture characteristics are obtained based on the preprocessed moisture transmission signal, including: S501. Determine the transmittance of the film under test in the moisture-sensitive band based on the pre-processed moisture transmission signal, and obtain the moisture-sensitive transmission characteristics. S502. Determine the moisture absorption characteristics corresponding to the moisture-sensitive band based on the moisture-sensitive transmission characteristics. S503. Obtain the initial moisture characteristics of the film under test based on its moisture absorption characteristics; S504. Obtain the preset moisture calibration data and combine it with the initial moisture characteristics to obtain the moisture characteristics of the film to be tested.

[0032] As described in steps S501 to S504 above, the reference light intensity of the moisture acquisition channel is obtained. This reference light intensity should be actually acquired under the condition that the film to be tested is removed, and the light source, filter, detector gain, sampling frequency, and optical path position are all consistent with the actual detection state. For example, after the light source is preheated for 60 seconds, 1000 reference data points of the moisture channel are continuously acquired. The sum of the 1000 data points is divided by 1000 to obtain 0.8V, which is then used as the reference light intensity of the moisture acquisition channel. Then, the pre-processed water at a certain sampling time is acquired. For example, if the transmitted signal is 0.5V, since dark current correction, reference intensity normalization, and filtering have already been performed, this signal is correlated with the reference state of the moisture channel to obtain 0.5 divided by 0.8, which is 0.625. Multiplying 0.625 by 100% gives 62.5%, determining that the transmittance of the film under test in the moisture-sensitive band at this sampling time is 62.5%. For continuous sampling, for example, the moisture transmission signals obtained at five adjacent sampling times are 0.5V, 0.496V, 0.492V, 0.4V, 0.4V, 0.4V, 0.4V, 0.4V, 0.4V, 0.5 ... At 88V and 0.490V, with a reference light intensity of 0.8V, the transmittance of the moisture-sensitive band was obtained as 62.5%, 62.0%, 61.5%, 61%, and 61.25%, respectively. Then, based on continuous transmittance, the moisture-sensitive transmission characteristics were obtained. Instead of directly using a single transmittance as the final moisture characteristic, data that reduces the influence of random noise was extracted from the transmittance changes in the moisture-sensitive band. For example, 20 consecutive sampling points were used as a moisture-sensitive characteristic window, and the transmittance of the 20 sampling points was averaged. If the 20 transmittance values ​​are concentrated between 61% and 63.0%, add the 20 transmittance values ​​together and divide by 20. For example, if the sum of the 20 transmittance values ​​is 1245%, then dividing by 20 gives 62.25%. Use 62.25% as the moisture-sensitive transmittance characteristic corresponding to this feature window. To enhance the response to moisture changes, the transmittance change within this window can also be obtained simultaneously. For example, if the initial transmittance of the window is 63.0% and the transmittance at the end of the window is 61%, then subtracting 61% from 63.0% gives 2.0 percentage points.The moisture-sensitive transmission characteristic of this window is defined as 0 percentage points. This characteristic includes the current transmittance in the moisture-sensitive band and the degree of transmission change within the preset sampling window. The method used can be determined during the calibration phase using actual film samples. For example, prepare films from the same batch in dry, low-moisture, medium-moisture, and high-moisture states. Place them in environments with relative humidity of 10%, 30%, 50%, 70%, and 90% for 12–24 hours to reach a stable state. Collect at least 1000 moisture channel data points for each state, and then compare the transmittance changes in the moisture-sensitive band corresponding to each state. If the moisture content increases after 19... The transmittance near 40nm decreased from 70% to 60%, while near 1720nm it only decreased from 82% to 80%, indicating that the 1940nm channel has a more significant moisture response and is therefore identified as the moisture-sensitive channel. In actual studies, 1940nm typically exhibits strong water absorption characteristics and shows higher moisture sensitivity than 1450nm in some materials. Therefore, this embodiment can use approximately 1940nm as the preferred moisture-sensitive band, but the final result will be based on the actual thin film calibration. The preprocessed electrical signal is converted into a moisture-sensitive transmittance with a uniform scale, and single-point noise is further reduced through a time window. Since increased moisture content in the moisture-sensitive band typically weakens the near-infrared transmitted light in the corresponding band, moisture absorption characteristics can be determined based on the moisture-sensitive transmittance. To convert the moisture-sensitive transmittance into a transmission attenuation form, first convert the percentage to a decimal, for example, convert 62.5% to 0.625. Then determine the reciprocal of 0.625, which is 1.6. Take the common logarithm of this value, which is approximately 0.204. Use 0.204 as the moisture absorption characteristic at that sampling time. If the moisture-sensitive transmittance at another location is 55%, first convert 55% to 0.55. The reciprocal of 0.55 is approximately 1.818. Take the common logarithm, which is approximately 0.26. Therefore, the moisture absorption characteristic at the latter location is larger. The previous position indicates that the optical absorption of the corresponding moisture-sensitive band is stronger. The moisture absorption characteristic is determined by the difference between the moisture-sensitive transmission characteristic and the dry reference transmission characteristic, depending on the actual application. For example, a fully dried film from the same batch is first prepared, and its moisture content is determined to be, for example, 0.05% using an oven drying method. Then, its transmittance is collected in the 1940nm channel, obtaining a dry reference transmittance of 70%. The transmittance of the film under test is 62.5%, so 70% minus 62.5% equals 7.5 percentage points. This 7.5 percentage points is used as the moisture absorption change characteristic relative to the dry state. For implementations requiring improved stability, reference positions are set on both sides of the moisture-sensitive band, for example, the moisture-sensitive... The wavelength range used is 1930-1950 nm, with 1900-1920 nm selected as a local reference range. The transmission characteristics of the two ranges are differentially processed to reduce the impact of slow light source drift and overall thin film transmission variations. The moisture-sensitive transmission characteristics obtained above are used to determine the moisture absorption characteristics. The method for determining the moisture absorption characteristics can be selected through calibration experiments. For example, prepare at least five thin film samples with different water contents, such as 0.1%, 0.3%, 0.5%, 0.7%, and 1%, and obtain the true water content using a reference method. Then, collect the moisture-sensitive transmission characteristics of each group of samples and convert them into moisture absorption characteristics. If the obtained moisture absorption characteristics are 0.01, 0.0, ... 3. Values ​​of 0.051, 0.072, and 0.101 generally increase with increasing water content, confirming the effectiveness of this water absorption characteristic in characterizing changes in water content. Although the water absorption band near 1940 nm has high sensitivity, it may also be affected by the material matrix. Therefore, this step preferably uses an actual calibration sample with the same material as the film being tested to determine the water absorption characteristic, rather than directly setting a fixed response based solely on the pure water spectrum. Related NIR studies also indicate that water has significant absorption near approximately 1450 nm and 1940 nm, but the specific absorption location and response intensity are affected by the sample matrix, temperature, and hydrogen bonding environment. Converting a decrease in transmittance into an increase in water absorption allows for the determination of the initial water characteristic based on the degree of absorption. Obtain the moisture absorption characteristics of the samples from the initial moisture calibration and sort them according to their true moisture content from low to high. For example, prepare five groups of film samples with moisture contents of 0.1%, 0.3%, 0.5%, 0.7%, and 1%. Each group of films is first conditioned in a constant temperature and humidity environment, for example, maintaining a temperature of 25℃ and relative humidity of 20%, 40%, 60%, 80%, and 90%, respectively, for 12–24 hours each. After stabilization, weigh the initial wet mass and the dry mass of the samples after complete drying. Subtract the dry mass from the wet mass to obtain the moisture mass of the sample. Then, correlate the moisture mass with the dry mass to obtain the actual moisture content; for example, the mass of a certain group of films after conditioned is 10.05g. The sample was then dried at 105℃ until the difference between two consecutive weighings was less than 0.001g, yielding a dry mass of 10g. Subtracting 10g from 10.05g gives 0.05g of moisture. Dividing 0.05g by 10g gives 0.005, which is then converted to a percentage, resulting in 0.5%. Therefore, the true moisture content of the calibrated sample is 0.5%. The sample was then placed in a detection device, for example, measuring a moisture absorption characteristic of 0.051. This established a calibration data point corresponding to a 0.5% moisture content for the 0.051 moisture absorption characteristic. For the film to be tested, if the moisture absorption characteristic is 0.061, and the calibration data shows 0.051 corresponding to 0.5% and 0.072 corresponding to 0.7%, then the moisture content can be determined based on two adjacent calibration data points. Initial moisture characteristics are determined at fixed points. For example, the difference between 0.061 and 0.051 is 0.01, and the difference between 0.072 and 0.051 is 0.021. Therefore, 0.01 accounts for approximately 47.6% of 0.021, which corresponds to approximately 47.6% of the 0.2 percentage points from 0.5% to 0.7%, resulting in approximately 0.095 percentage points. Adding 0.5% and 0.095% gives approximately 0.595%, thus determining the initial moisture characteristic to be approximately 0.595%. To improve stability, the moisture absorption characteristics of multiple consecutive sampling points can also be averaged before determining the initial moisture characteristic. For example, 10 consecutive sampling points yield 0.059, 0.06, 0.061, 0.06, 0.062, 0.061, 0.05... 9, 0.06, 0.061, and 0.06 are summed to obtain 0.603, which is then divided by 10 to get 0.0603. 0.0603 is used as the representative moisture absorption characteristic for this detection location. The initial moisture characteristic is then determined based on the calibration data. For films with a wide moisture content range, the calibration range can be divided into multiple intervals, such as 0-0.2%, 0.2-0.5%, 0.5-1%, and 1-2%. The initial moisture characteristic is determined using the calibration data for each corresponding interval to avoid forcibly treating the entire moisture content range as a straight line. The division of each interval is based on the actual test results of the calibration sample. For example, if the rate of change in moisture absorption characteristics changes significantly around 0.5%, then 0.0603 can be used as the interval.5% is used as a dividing point between intervals; in this way, the water content is initially calculated from the water absorption characteristics, which can ensure the basis for hierarchical referencing in the data chain, and also leave room for subsequent processing using complete water calibration data. Prepare a set of calibration samples similar to the material, thickness range, manufacturing process, and surface condition of the film to be tested. For example, for lithium battery separators in the 20-40 μm range, select five thicknesses: 20 μm, 25 μm, 30 μm, 35 μm, and 40 μm. For each thickness, set five moisture contents: 0.1%, 0.3%, 0.5%, 0.7%, and 1%, forming 25 calibration states. Prepare at least 5-10 samples for each state to reduce the influence of individual sample differences. Each sample is first conditioned in a constant temperature and humidity environment, for example, with the temperature set at 25℃ and the relative humidity set at 20%, 40%, 60%, 80%, or 90%, for 12-24 hours. Then, perform actual moisture reference measurements, for example, using the aforementioned weighing drying method. The method obtains the true moisture content; then, the moisture transmission signal of the sample is acquired in a near-infrared detection device, and sequentially undergoes dark current correction, reference light intensity normalization, synchronous filtering, transmittance determination, moisture absorption characteristic determination, and initial moisture characteristic acquisition. Finally, the initial moisture characteristic is paired with the actual moisture content obtained by the reference method to obtain preset moisture calibration data; for example, in a batch of calibration data, the actual moisture contents of 0.3%, 0.5%, 0.7%, and 1% correspond to initial moisture characteristics of 0.03, 0.051, 0.072, and 0.103, respectively; for the film under test, the initial moisture characteristic is 0.061, which is between 0.051 and 0.072. Therefore, segmented conversion is performed between the two calibration points of 0.5% and 0.7%, 0. Subtracting 0.051 from 0.61 gives 0.01, and subtracting 0.051 from 0.072 gives 0.021. Dividing 0.01 by 0.021 gives approximately 0.476. Multiplying the 0.2 percentage point between 0.5% and 0.7% by 0.476 gives approximately 0.095 percentage points. Finally, adding 0.5% and 0.095% gives approximately 0.595%. Therefore, the final moisture characteristic can be determined to be approximately 0.595%. If the calibration experiment finds that the initial moisture characteristic is not strictly linear with the actual moisture content, the above linear interpolation is not mandatory. Instead, multiple moisture content intervals are divided according to the actual calibration data, such as 0-0.3%, 0.3-0.5%, 0.5-0.7%, and 0.7-1%, and corresponding intervals are established for each interval. The calibration mapping is as follows: For example, after measuring multiple calibration points within the 0.5-0.7% range, if the actual data shows that an increase of 0.01 in the initial moisture characteristic only corresponds to an increase of 0.08 percentage points in the moisture content, then the conversion should be performed according to the actual calibration results for that range, instead of continuing to use the example value of 0.01 corresponding to 0.095 percentage points; to avoid the calibration data being affected by a single sample, multiple samples at the same moisture state are repeatedly collected. For example, if 10 samples yield values ​​of 0.05, 0.052, 0.051, 0.053, 0.05, 0.051, 0.052, 0.05, 0.051, and 0.052 respectively, these values ​​are added together to obtain 0.512, then divided by 10 to obtain 0.0512, and 0.0512 is taken as 0.The representative initial moisture characteristics corresponding to a 5% moisture content state are determined. Simultaneously, independent validation samples can be used to verify the calibration results. For example, the calibration samples do not participate in the final validation. Five additional samples with unknown moisture content are prepared, and their true moisture content is measured using a reference method. This measured content is then compared with the near-infrared spectroscopy (NIRS) detection results. If the reference moisture contents of the five samples are 0.20%, 0.40%, 0.6%, 0.80%, and 1.00%, respectively, and the NIRS detection results are 0.21%, 0.39%, 0.61%, 0.79%, and 1.02%, respectively, then the maximum absolute deviation can be further calculated. For example, subtracting 1.00% from 1.02% yields 0.02 percentage points, which can be used to validate the calibration. The initial moisture content calibration data should be re-acquired when significant changes occur in the light source, filter, detector, optical path position, or thin film material, such as after replacing the 1940nm filter. For production equipment, calibration should be performed according to the equipment maintenance cycle, such as every 30 or 90 days. Alternatively, recalibration can be performed when the deviation from the reference sample exceeds a preset range, such as 0.05-0.1 percentage points. The obtained initial moisture characteristics are ultimately converted into moisture characteristics with actual moisture content significance; for example, an initial moisture characteristic of 0.061 may be ultimately determined to be 0.595%, which is suitable for the optical properties of specific thin film materials.

[0033] In a preferred embodiment, the thickness feature is correlated with the moisture feature to determine the moisture feature offset, including: S601. Obtain the sampling time corresponding to the thickness feature and moisture feature, and synchronously pair the thickness feature and moisture feature corresponding to the same area to be measured to obtain thickness and moisture sampling data. S602. Based on the thickness moisture sampling data, determine the thickness characteristic change and the corresponding moisture characteristic change, and based on the thickness characteristic change and the moisture characteristic change, determine the additional change in moisture characteristics caused by the change in film thickness. S603. Based on the change in thickness characteristics and the additional change, determine the degree of offset of the moisture characteristics relative to the corresponding thickness state, and obtain the moisture characteristic offset.

[0034] As described in steps S601 to S603 above, the thickness feature and its corresponding sampling time are acquired, as are the moisture feature and its corresponding sampling time. The sampling time can be generated by a shared sampling clock for the two near-infrared acquisition channels. For example, if the system performs synchronous sampling at 1000Hz, a sampling time is formed every 1ms, and a consecutive sampling sequence number is assigned to each sampling time. For instance, the first sampling point corresponds to 1ms, the second to 2ms, the third to 3ms, and so on. Since the reference acquisition channel and the moisture acquisition channel use the same trigger signal for acquisition, the same sampling sequence number... The reference transmission signal and moisture transmission signal come from the same area to be measured. Since the thickness and moisture characteristics are obtained from the data of their respective channels, they can be matched according to the sampling sequence. For example, if the thickness characteristic of 30 μm and the moisture characteristic of 0.6% are obtained at the 1000th sampling time, then the sampling time 1000, the thickness characteristic of 30 μm, and the moisture characteristic of 0.6% are taken as a set of thickness and moisture sampling data; if the thickness characteristic of 30.2 μm and the moisture characteristic of 0.61% are obtained at the 1001st sampling time, then the second set of thickness and moisture sampling data is formed; if the thickness characteristic of 30.5 μm and the moisture characteristic of 0.61% are obtained at the 1002nd sampling time, then the second set of thickness and moisture sampling data is formed.63% results in the formation of the third set of thickness moisture sampling data. For scenarios involving continuous film transport, the correspondence between sampling time and the actual spatial position of the film needs to be considered. Therefore, the film transport speed can be further obtained. For example, if the transport speed is 1 m / s obtained through an encoder, the film movement distance between adjacent sampling points at 1000 Hz sampling is 1 mm, meaning the film position corresponding to the 1000th sampling point differs from the 1001st sampling point by approximately 1 mm. If the system has a 30 mm × 30 mm effective detection area, each sampling point can be considered as a continuous detection position of that effective detection area in the transport direction. To avoid misalignment between the two channels due to data processing delays, the same hardware timestamp is written to the reference channel and moisture channel at the data acquisition end. For example, the 1000th reference signal and the 1000th moisture signal are both marked with a timestamp of 1 second. Even if the two channels pass through different amplification and filtering circuits, they are rearranged according to the same sampling sequence number. If a certain... If a sampling point is lost in a channel, for example, if the reference channel has four sampling points (1000, 1001, 1002, and 1003), but the moisture channel is missing point 1002, the data between 1001 and 1003 is not forcibly paired. Instead, the sampling position is marked as invalid, or synchronization data is re-established using adjacent valid sampling data to avoid incorrect pairing. In practice, 1000-10000 sets of thickness and moisture sampling data can be continuously generated, and then segmented according to a preset time window, for example, every 100 sampling points form a local data segment, with each data segment corresponding to a 100ms detection time. This allows for the observation of local changes in thickness and moisture separately. The final thickness and moisture sampling data includes at least the sampling time, thickness characteristics, and moisture characteristics corresponding to that sampling time. A strict data basis of the same location and time is established, making subsequent changes in thickness and moisture comparable, thus reducing errors caused by different locations and sampling times from the source. From the thickness moisture sampling data, two representative positions within adjacent sampling points or a preset time window are selected as comparison objects. For example, at the 1000th sampling time, the thickness feature is 30μm and the moisture feature is 0.6%. At the 1100th sampling time, the thickness feature is 35μm and the moisture feature is 0.7%. Since the difference between the two is 100 sampling points and the sampling frequency is 1000Hz, the time interval is 100ms. At this time, 30μm is subtracted from 35μm to obtain a thickness feature change of 5μm, and 0.6% is subtracted from 0.7% to obtain a moisture feature change of 0.1 percentage points. This results in a thickness increase of 5μm and a moisture feature change. Add 0.1 percentage points of synchronous change data; to avoid misjudgments caused by noise from a single sampling point, multiple sampling points can be used to determine the amount of change. For example, set 100 sampling points as a comparison window, and average the thickness and moisture features within the window respectively; assuming the average value of the 100 thickness features in the first window is 30 μm and the average value of the 100 thickness features in the second window is 35 μm, then the change in thickness feature is 5 μm; the average value of the 100 moisture features in the first window is 0.6% and the average value of the 100 moisture features in the second window is 0.7%, then the change in moisture feature is 0.1 percentage points; subsequently, it is necessary to confirm... To determine how much of the 0.1 percentage point change in moisture content is due to additional changes in thickness, a set of calibration films with essentially the same water content but different thicknesses were prepared. For example, films of the same material with thicknesses of 20 μm, 25 μm, 30 μm, 35 μm, and 40 μm were prepared and stabilized in a constant temperature and humidity environment, such as 25℃ and 50% relative humidity for 24 hours. Then, the moisture characteristics of samples of each thickness were used. Assuming that the actual water content of the 20 μm, 25 μm, 30 μm, 35 μm, and 40 μm samples is all around 0.5%, but the corresponding detected moisture characteristics are 0.5%, 0.52%, and 0.54%, respectively. If the moisture content is 0.56% and 0.58%, then the additional change in moisture content due to thickness variation can be determined from adjacent thickness ranges. For example, when the thickness increases by 5μm from 30μm to 35μm, the moisture content increases from 0.54% to 0.56%, which is an increase of 0.02 percentage points. Therefore, the additional change in moisture content corresponding to each 5μm increase in thickness can be taken as the actual calibration result for that thickness range. If the film under test detects a 5μm increase in thickness and a 0.1 percentage point increase in moisture content, then based on the calibration data of the corresponding thickness range, approximately 0.02 percentage points are determined to be the additional change caused by the thickness variation, and the remaining 0.A change of 0.08 percentage points represents a change in moisture characteristics that cannot be explained by the thickness variation. To improve calibration reliability, 5-10 samples are set up for each thickness range, and 500-5000 sampling points are collected for each. Then, the data for the same thickness range are averaged. For example, the moisture characteristics of 5 replicate samples of a 35μm thickness are 0.55%, 0.56%, 0.55%, 0.57%, and 0.56%, respectively. Adding these values ​​gives 2.79%, and dividing by 5 gives 0.558%. 0.558% is then taken as the representative moisture characteristic for a 35μm thickness. The representative moisture characteristic for a 30μm thickness is, for example, 0.538%. Subtracting the two gives 0.02 percentage points, meaning that the increase in moisture characteristic from 30μm to 35μm corresponds to an additional 0.02 percentage points. For thickness variations with a large range, the thickness variation is divided into multiple intervals, such as 20-25μm, 25-30μm, and 30-35μm. Instead of using a fixed value for the entire 20-40μm range, additional variation data should be established for the 35-40μm and 35-35μm intervals. If the actual variation is 0.02 percentage points for the 30-35μm interval and 0.025 percentage points for the 35-40μm interval, then when the thickness to be measured increases from 35μm to 40μm, 0.025 percentage points should be used as the corresponding additional variation. When the actual thickness variation is 7μm but the calibration data only includes the 30-35μm and 35-40μm intervals, the calibration results of adjacent intervals should be used for segmented conversion. For example, if 0.02 percentage points correspond to 30-35μm and 0.025 percentage points correspond to 35-40μm, then first determine that 5μm in the 30-35μm interval corresponds to 0.02 percentage points, then determine that 2μm in the 35-37μm interval corresponds to 0.02 percentage points based on the variation rate of the 35-40μm interval, resulting in approximately 0.01 percentage points, and finally obtaining an additional variation of approximately 0.03 percentage points. The change in thickness characteristic and the corresponding additional change in moisture characteristic are obtained. Then, the direction of the additional change is determined according to the direction of the thickness change. For example, when the thickness increases from 30 μm to 35 μm, the calibration data shows that the moisture characteristic will increase by 0.02 percentage points due to the increase in thickness. Therefore, 0.02 percentage points is taken as the additional change corresponding to this thickness change. If the actual detected moisture characteristic increases from 0.6% to 0.7%, the actual change in moisture characteristic is 0.1 percentage points. Then, 0.02 percentage points are subtracted from 0.1 percentage points to obtain 0.08 percentage points. This 0.08 percentage points is determined as the moisture characteristic relative to the thickness change. Feature offset; after considering the additional 0.02 percentage points of response resulting from the thickness change from 30μm to 35μm, there is still a 0.08 percentage point change in moisture characteristic that cannot be explained by the thickness change. Therefore, this 0.08 percentage point can be used as the basis for subsequent moisture characteristic correction. In order to make the offset more stable, the above process is repeated for multiple sampling points within a continuous detection window. For example, a 100ms window contains 100 synchronous sampling points. The thickness change, moisture change, and corresponding additional thickness change are obtained for each sampling point. Then, the moisture characteristic offset is obtained point by point. Assume that the total offset obtained from 100 sampling points is 7. 8 percentage points. Dividing 7.8 percentage points by 100 gives 0.078 percentage points. This 0.078 percentage point is taken as the representative moisture characteristic offset for this 100ms window. The direction of the offset is retained (positive or negative). For example, if the actual moisture characteristic change is 0.1 percentage points, and the thickness-related additional change is 0.02 percentage points, the offset is positive 0.08 percentage points. If the actual moisture characteristic change is only 0.01 percentage points, and the additional change caused by the increase in thickness is 0.02 percentage points, then subtracting 0.02 percentage points from 0.01 percentage points gives a negative 0.01 percentage point, indicating that the actual moisture characteristic change is less than the thickness change. The expected additional response; set an effective offset range, for example, based on repeated tests of no-load detection and stable water-containing samples, determine the normal offset noise range as ±0.02 percentage points. If a standard film with constant water content is prepared, continuously collect 10,000 moisture characteristic data points with constant thickness, calculate the offset of adjacent windows, for example, if the offset is mostly between -0.018 and 0.019 percentage points, then ±0.02 percentage points can be set as the normal fluctuation range; when the offset is 0.08 percentage points, exceeding ±0.02 percentage points is considered to have a significant offset that requires moisture characteristic correction; while when the offset is only 0.If the deviation is 0.01 percentage points, it can be considered within the normal detection fluctuation range. In subsequent processing, the original moisture characteristic can be maintained or only slightly corrected. This deviation threshold is not set arbitrarily, but is obtained by repeated detection of stable standard samples. For example, after continuously detecting 10,000 samples, the maximum deviation range under normal conditions is statistically analyzed, and then a certain safety margin is added. For example, if the actual maximum fluctuation is ±0.016 percentage points, then ±0.02 percentage points can be used as the judgment threshold. The final result is a clear data output. For example, if the thickness increases from 30μm to 35μm, the thickness change is 5μm, and the synchronous moisture characteristic increases from 0.6% to 0.7%, the moisture characteristic change is 0.1 percentage points. After calibration, it is determined that the thickness change causes an additional change of 0.02 percentage points, so a moisture characteristic deviation of positive 0.08 percentage points is obtained. Instead of directly misjudging the optical response caused by the change in film thickness as a change in water content, the thickness characteristic obtained by the independent thickness channel is used as a synchronous reference, and the additional moisture response corresponding to the thickness change is obtained through actual calibration. Then, this additional response is subtracted from the actual moisture change, thereby improving the accuracy and stability of film moisture content detection. .

[0035] In a preferred embodiment, the moisture characteristics are processed based on the moisture characteristic offset to obtain the water content characteristics corresponding to the thickness change of the film under test. The thickness detection results and moisture content detection results of the film under test are then obtained based on the thickness characteristics and water content characteristics, including: S701. Correct the moisture characteristics based on the moisture characteristic offset to obtain the corrected moisture characteristics; S702. Obtain the change in moisture feature corresponding to the corrected moisture feature and the moisture feature offset; S703. Based on the change in moisture characteristics, the corrected moisture characteristics are continuously sampled to obtain the moisture content characteristics corresponding to the change in the thickness of the film to be tested. S704. Obtain preset thickness calibration data and preset moisture content calibration data. Determine the thickness detection result of the film to be tested based on the thickness characteristics and preset thickness calibration data, and determine the moisture content detection result of the film to be tested based on the moisture content characteristics and preset moisture content calibration data.

[0036] As described in steps S701 to S704 above, the moisture characteristic and its offset are acquired, and the direction of change corresponding to the offset is read. For example, if the moisture characteristic is 0.7% at a certain sampling moment, and the offset is determined to be positive 0.08 percentage points based on the additional change caused by the thickness increasing from 30μm to 35μm, it indicates that the 0.08 percentage point change in the currently detected 0.7% cannot be directly explained by the actual water content change, but is an additional effect caused by the thickness change. Therefore, 0.08 percentage points are subtracted from 0.7% to obtain 0.62%, which is determined as the corrected moisture characteristic. If the result is negative 0.05 percentage points, for example, the moisture characteristic decreases by 0.05 percentage points relative to the corresponding thickness state after the actual thickness changes, then the moisture characteristic is reversed, that is, the current moisture characteristic of 0.6% is subtracted from negative 0.05 percentage points to obtain... The moisture content is reduced to 0.65%, resulting in a corrected moisture characteristic of 0.65%. The direction and correction range of the offset are determined by obtaining the normal fluctuation range in previous stable sample tests. For example, 10,000 data points are continuously collected on a standard film with a moisture content of 0.5% and a thickness of 30 μm. Under normal conditions, the moisture characteristic offset is mainly between -0.02 and +0.02 percentage points. Therefore, ±0.02 percentage points can be considered the normal offset range. When the offset is +0.08 percentage points, it exceeds this normal range, so a substantial correction is performed. When the offset is only +0.01 percentage points, it can be considered a normal measurement fluctuation, maintaining the original moisture characteristic or performing the same correction without significant impact. The offset range should be obtained from actual stable sample testing, not arbitrarily set. This allows for the separation of moisture characteristics from thickness effects, thereby reducing the interference of film thickness changes on moisture detection results. The process involves acquiring the moisture characteristic offset and the corrected moisture characteristic, then matching them at the same sampling time to determine the difference between the original and corrected moisture characteristics. For example, if the moisture characteristic is 0.7% at a certain sampling time, a moisture characteristic offset of +0.08 percentage points is obtained. Subtracting 0.08 percentage points from 0.7% yields a corrected moisture characteristic of 0.62%. The difference between 0.7% and 0.62% is 0.08 percentage points, which is then determined as the change in moisture characteristic at that sampling time. This change in moisture characteristic reflects the degree of correction made by the thickness compensation. For cases where the offset is negative, such as an original moisture characteristic of 0.6% and a moisture characteristic offset of -0.05 percentage points, resulting in a corrected moisture characteristic of 0.65%, subtracting 0.6% from 0.65% yields a change in moisture characteristic of +0.05 percentage points. This indicates that the correction has increased the moisture characteristic by 0.05 percentage points. The corrected absolute moisture characteristic and the change in moisture characteristic during the correction process are saved separately to avoid difficulty in determining the magnitude of changes in moisture characteristic during subsequent continuous sampling. Maintaining the same 1000Hz synchronous sampling frequency, that is, acquiring the corrected moisture characteristic corresponding to a sampling position every 1ms, and simultaneously recording the change in moisture characteristic at that sampling position; for example, if the corrected moisture characteristics obtained at 5 consecutive sampling times are 0.6%, 0.61%, 0.62%, 0.62%, and 0.63%, and the corresponding changes in moisture characteristic are 0.05, 0.06, 0.07, 0.07, and 0.08 percentage points, then a continuous data sequence is first formed according to the sampling order; if the conveying speed of the film under test is 1m / s If adjacent 1ms sampling positions differ by approximately 1mm, then the aforementioned 5 sampling points correspond to a continuous detection range of approximately 4mm along the film conveying direction. To reduce noise at individual sampling points, continuous sampling windows are set up, for example, forming a moisture content characteristic window every 20 sampling points. Assuming that the corrected moisture content characteristics of a certain 20 sampling points are concentrated between 0.6% and 0.64%, the 20 values ​​are added together and divided by 20. For example, the sum is 12.40%, resulting in 0.62%. This 0.62% is used as the representative moisture content characteristic of this detection window. To preserve the effects of thickness variations... The local changes are recorded simultaneously with the thickness characteristics of the 20 sampling points. For example, if the thickness gradually changes from 30 μm to 35 μm, then 0.62% is taken as the representative water content characteristic corresponding to this thickness change interval. If the average thickness of the next 20-point window is 35 μm and the average corrected water content characteristic is 0.65%, then continuous water content data corresponding to 0.62% for 30 μm and 0.65% for 35 μm are formed. Through continuous sampling, a set of water content characteristics that continuously change along the film running direction is obtained, for example, the thickness changes from 30 μm to 31 μm. At thicknesses of 32μm, 33μm, 34μm, and 35μm, the corresponding water content characteristics are 0.6%, 0.61%, 0.61%, 0.62%, 0.63%, and 0.63%, respectively. The resulting water content characteristics are not the moisture values ​​at a single time point, but rather the water content data that has been corrected for thickness offset and corresponds to the continuous detection positions of the film. By continuously sampling, the influence of single-point noise on the detection results is reduced, and the synchronous correspondence between thickness changes and water content characteristic changes can be maintained, enabling the final obtained water content characteristics to be used for online detection in continuous film production processes. The preset thickness calibration data and preset moisture content calibration data are not fixed values ​​set directly by humans, but are obtained through calibration using actual standard samples. For the thickness calibration data, standard films with the same or similar material, structure, and production process as the film to be tested are prepared. For example, nine thickness ranges are prepared: 10μm, 15μm, 20μm, 25μm, 30μm, 35μm, 40μm, 45μm, and 50μm. At least 5-10 samples are prepared for each range, and the actual thickness is measured using a calibrated contact thickness gauge, micrometer, interferometric thickness gauge, or existing thickness reference equipment on the production line. For example, the actual measured values ​​of the 10 samples in the 30μm range are 29.8, 30.1, 30, 29.9, 30.2, and 30, respectively. 1. The thicknesses are 29.8, 30, 30.1, and 29.9 μm. Adding these values ​​together gives 299.9 μm, then dividing by 10 gives 29.99 μm. Therefore, approximately 30 μm is used as the actual reference thickness for this range. Subsequently, the same batch of samples is placed in a near-infrared detector, and the corresponding thickness features are extracted. For example, a 30 μm sample yields a thickness feature of 30.1 μm, and a 35 μm sample yields 35.2 μm, forming thickness calibration data for the thickness features and the actual reference thickness. For moisture content calibration data, films from the same batch with different actual moisture contents need to be prepared, for example, setting six moisture content levels: 0.1%, 0.3%, 0.5%, 0.7%, 1%, and 1.5%. Humidity is adjusted using a constant temperature and humidity environment, for example, setting the temperature to 25℃. The relative humidity was set to 20%, 40%, 60%, 80%, and 90%, and each state was maintained for 12–24 hours. The actual moisture content was then obtained by weighing. For example, if the sample mass after humidification was 10.06g and the mass after complete drying was 10g, then 10.06g minus 10g yielded 0.06g of moisture. Dividing 0.06g by 10g gave 0.006, which was converted to a percentage, resulting in 0.6%. Therefore, the actual moisture content of the sample was 0.6%. Subsequently, the initial moisture characteristic was extracted, and the moisture content characteristic after thickness correction was obtained. For example, a sample with an actual moisture content of 0.6% yielded a moisture content characteristic of 0.061, forming moisture content calibration data corresponding to an actual moisture content of 0.6% for moisture content characteristic 0.061. Each moisture content... The sampling mode can prepare 5-10 samples and perform multiple repeated samplings. For example, the water content characteristics of 10 samples corresponding to a true water content of 0.5% are 0.05, 0.051, 0.052, 0.05, 0.051, 0.053, 0.051, 0.05, 0.052, and 0.051, respectively. Add these values ​​together to get 0.511, then divide by 10 to get 0.0511. Use 0.0511 as the representative water content characteristic corresponding to 0.5% water content. Finally, calibrated water content data are formed, for example, 0.03 corresponds to 0.3%, 0.051 corresponds to 0.5%, 0.061 corresponds to 0.6%, 0.072 corresponds to 0.7%, and 0.103 corresponds to 1%. For the water content characteristic of the membrane to be tested, for example, 0.If the value is 0.066, then it falls between 0.061 and 0.072. Therefore, a conversion is performed between the two calibration points of 0.6% and 0.7%. For example, 0.066 minus 0.061 gives 0.005, 0.072 minus 0.061 gives 0.011, 0.005 divided by 0.011 gives approximately 0.455, then multiplying the 0.1 percentage point between 0.6% and 0.7% by 0.455 gives approximately 0.0455 percentage points, and finally adding 0.6% and 0.0455% gives approximately 0.6455%. Therefore, the moisture content of the film to be tested is determined to be approximately 0.646%. To verify the validity of the preset calibration data, additional verification samples that were not used in the calibration should be prepared, such as 10 sets of unknown samples. The true thickness and true moisture content are obtained through a reference measurement method, and then the near-infrared detection results are compared with the reference results. If the thickness... If the maximum deviation is controlled within, for example, ±0.5 μm, and the maximum deviation of moisture content is controlled within, for example, ±0.05 percentage points, then the calibration data set can be considered to meet the current detection accuracy requirements. The actual threshold should be determined based on the film specifications and production quality requirements, and should not be used as a fixed standard for all films. Finally, two final detection results are output. For example, the thickness feature of 34.0 μm is converted to a thickness detection result of approximately 33.8 μm using thickness calibration data, and the moisture content feature of 0.066 is converted to a moisture content detection result of approximately 0.646% using moisture content calibration data, thus achieving simultaneous detection of thickness and moisture content in the same test area. The optical signal processing, thickness influence correction, and continuous sampling results are ultimately converted into thickness and moisture content values ​​with practical metrological significance. Simultaneously, independent reference samples are used to establish calibration data, ensuring that the detection results have a clear source and traceability.

[0037] Please see the appendix Figure 2 As shown, the present invention also provides a near-infrared transmission-based thin film thickness and moisture content detection system for the aforementioned near-infrared transmission-based thin film thickness and moisture content detection method, comprising: The acquisition channel module is used to acquire the film under test and uses a near-infrared broadband light source to illuminate the film under test through transmission, forming at least two near-infrared acquisition channels corresponding to the same measurement area on the transmission side of the film under test. The transmission signal acquisition module, based on at least two near-infrared acquisition channels, synchronously acquires the near-infrared light transmitted through the thin film under test, and obtains the reference transmission signal corresponding to the reference band and the moisture transmission signal corresponding to the moisture-sensitive band. The signal preprocessing module is used to simultaneously preprocess the reference transmission signal and the moisture transmission signal to obtain the preprocessed reference transmission signal and the preprocessed moisture transmission signal. The thickness feature acquisition module acquires thickness features based on the preprocessed reference transmission signal. The moisture feature acquisition module acquires moisture features based on the preprocessed moisture transmission signal. The feature offset module is used to associate thickness features with moisture features and determine the moisture feature offset. The detection results module processes the moisture characteristics based on the moisture characteristic offset to obtain the water content characteristics corresponding to the thickness change of the film under test, and obtains the thickness detection results and moisture content detection results of the film under test based on the thickness characteristics and water content characteristics.

[0038] The above describes a method where the film under test is positioned between a near-infrared broadband light source and a receiver, ensuring that the test area of ​​the film is located on the outgoing light path of the near-infrared broadband light source. The near-infrared broadband light source provides continuous near-infrared radiation to the test area. The acquisition channel module performs collimation, beam expansion, and homogenization to ensure that the illumination range incident on the film under test covers the test area and reduces the impact of light intensity differences at different locations on the transmission detection results. At least two near-infrared acquisition channels are set on the transmission side of the film under test. Each near-infrared acquisition channel selects different wavelengths in the transmitted light and focuses and performs photoelectric conversion on the selected near-infrared light, thereby forming multiple acquisition channels corresponding to the same test area. At least one acquisition channel is used to acquire the reference wavelength transmitted light, and at least one acquisition channel is used to acquire the moisture-sensitive wavelength transmitted light. By making multiple acquisition channels correspond to the same test area, the detection data obtained from different wavelengths have a unified spatial reference, avoiding detection deviations caused by differences in film thickness or water content at different detection locations. The transmission signal acquisition module determines the reference band for characterizing film thickness variation and the moisture-sensitive band for characterizing film water content based on the actual acquisition band of each near-infrared acquisition channel and the transmission characteristics of the film under test in different bands. It then selects the corresponding near-infrared band from the transmitted light using a corresponding filter structure. Each acquisition channel receives the transmitted light of the corresponding band under the same sampling sequence and converts the received light signal into a corresponding photoelectric signal. These signals are then synchronously acquired according to a unified sampling clock to obtain the reference transmission signal and the moisture transmission signal. Since both the reference transmission signal and the moisture transmission signal originate from the same test area and are obtained through synchronous sampling, the temporal and spatial correspondence between the two signals is guaranteed. This allows the subsequently obtained thickness and moisture characteristics to be processed for the same film location. By using the reference band and the moisture-sensitive band to acquire different detection information, thickness detection and moisture content detection are completed within the same optical acquisition structure, which improves the consistency and efficiency of thickness and moisture content detection. The signal preprocessing module acquires the dark current signal when the near-infrared light source is off or the transmitted light is blocked, and removes the corresponding dark current components from the reference transmission signal and the moisture transmission signal respectively to reduce the impact of the detector's background response on the detection results. It also acquires the reference light intensity of the corresponding acquisition channel when there is no film to be tested, and uses the reference light intensity to normalize the dark current-corrected transmission signal, converting the transmission signals obtained at different times and under different light source output states to a unified standardized range. After normalization, the standardized transmission signals of the two channels are synchronously filtered to reduce random noise and high-frequency disturbances, while maintaining the time correspondence between the reference transmission signal and the moisture transmission signal. This results in preprocessed reference and moisture transmission signals, reducing the impact of light source output fluctuations, detector dark current, electronic noise, and transient disturbances on subsequent feature extraction, and avoiding time shifts between the two detection channels due to different processing sequences. This improves the detection repeatability and stability of the system during long-term operation. The thickness feature acquisition module determines the transmittance of the film under test in the reference band based on the preprocessed reference transmission signal and the corresponding reference light intensity. Then, it determines the transmission attenuation feature corresponding to the change in optical path of the film under test based on the transmittance. Based on the pre-acquired thickness calibration data, it converts the transmission attenuation feature into a thickness feature that can characterize the film thickness. The thickness calibration data is obtained by detecting standard films with different actual thicknesses. The actual thickness of the standard films is obtained by a calibrated thickness measurement device. At the same time, the transmission response of each standard film in the reference band is recorded, thereby establishing the correspondence between the transmission response and the actual thickness. For the thickness feature obtained in the actual detection process, the corresponding thickness state is determined according to its calibration range, thereby obtaining a continuously changing thickness characterization result. By using the reference band to extract the thickness feature, the thickness information is obtained relatively independently from the moisture-sensitive detection signal. The near-infrared transmission response is converted into a thickness feature to realize online detection of film thickness, reducing the impact of traditional contact thickness measurement methods on the film's operating state and improving the data synchronization between thickness detection and moisture content detection. The moisture characteristic acquisition module determines the transmittance of the film under test in the moisture-sensitive band based on the moisture transmission signal and the corresponding reference light intensity. It then determines the moisture absorption characteristics corresponding to this band based on the transmittance. Next, it converts the moisture absorption characteristics into corresponding moisture characteristics based on pre-acquired moisture calibration data. This moisture calibration data is obtained by near-infrared sampling of films from the same batch or material system under different actual moisture states. Simultaneously, an independent reference detection method is used to determine the actual moisture content of each standard sample, and the actual moisture content is correlated with the corresponding near-infrared moisture response, thus forming the basis for moisture characteristic calibration. After obtaining the moisture characteristics, they are not directly used as the final moisture content detection result, but rather as the basis for subsequent thickness influence correction. This allows the moisture characteristics to be further processed in conjunction with thickness characteristics. By utilizing the moisture-sensitive band to acquire moisture characteristics, the system's responsiveness to changes in the film's moisture state is enhanced. Furthermore, by acquiring moisture characteristics separately from thickness characteristics, subsequent identification of additional changes in moisture characteristics caused by film thickness variations is possible, thereby reducing the interference of thickness variations on moisture content detection and improving the reliability of the moisture content detection results. The feature offset module acquires the sampling time corresponding to the thickness feature and moisture feature, and synchronously pairs the thickness feature and moisture feature corresponding to the same test area according to the same sampling time to form thickness and moisture sampling data. Based on the thickness feature change between adjacent sampling times or adjacent sampling windows, the corresponding thickness feature change amount is determined. At the same time, the moisture feature change amount within the same sampling range is determined. Using the pre-acquired moisture feature calibration data under different thickness states, the additional change amount of moisture feature corresponding to the current thickness change is determined. This additional change amount is compared with the actual detected moisture feature change amount to determine the degree of offset in the moisture feature relative to the corresponding thickness state, and the moisture feature offset amount is obtained. The moisture feature calibration data under different thickness states are obtained by detecting standard samples with basically consistent film moisture content and different film thicknesses, so that the additional change of moisture feature caused by thickness change can be identified from the actual moisture feature change. It can distinguish the additional change of moisture feature caused by film thickness change from the actual moisture detection response, and avoid misjudging the change in transmission response caused by thickness change as a change in moisture content when film thickness changes. The detection result module corrects the moisture feature based on the moisture feature offset obtained from the feature offset module, resulting in a corrected moisture feature to reduce the additional impact of film thickness variations on the moisture feature. It determines the corresponding change in moisture feature based on the moisture feature before and after correction, and continuously processes the corrected moisture feature according to the continuous sampling sequence. This allows the moisture response at multiple continuous sampling locations to form a moisture content feature corresponding to the film thickness variation. The thickness feature is then matched or converted with pre-acquired thickness calibration data to obtain the actual thickness detection result of the film under test. Simultaneously, the moisture content feature is matched or converted with pre-acquired moisture content calibration data to obtain the actual moisture content detection result of the film under test. The thickness calibration data comes from near-infrared detection results of standard films with different actual thicknesses. Based on the thickness test results, the moisture content calibration data comes from the near-infrared test results of films with different actual moisture content standards and the reference moisture content test results. Therefore, the final output thickness test results and moisture content test results have a clear calibration basis, forming a complete detection process from near-infrared transmission signal acquisition, signal preprocessing, thickness feature extraction, moisture feature extraction, thickness influence correction to the final result output. This allows the thickness test results and moisture content test results to be obtained synchronously based on the same test area. By correcting the moisture feature by the moisture feature offset, the interference of film thickness changes on moisture content detection is reduced, improving the accuracy and stability of moisture content detection. At the same time, it reduces the dependence on independent thickness test equipment, which is conducive to realizing online, synchronous and integrated quality inspection in the film production process.

[0039] And, a near-infrared transmission-based thin film thickness and moisture content detection terminal, including: One or more processors; A storage device on which one or more programs are stored; When one or more programs are executed by one or more processors, the one or more processors implement a method for detecting film thickness and moisture content based on near-infrared transmission.

[0040] The above description is merely a preferred embodiment of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention. Structures, devices, and operating methods not specifically described or explained in this invention are implemented according to conventional methods in the art unless otherwise specified or limited.

Claims

1. A method for detecting film thickness and moisture content based on near-infrared transmission, characterized in that, include: The film to be tested is acquired and illuminated by a near-infrared broadband light source, forming at least two near-infrared acquisition channels corresponding to the same measurement area on the transmission side of the film to be tested. Based on at least two near-infrared acquisition channels, the near-infrared light transmitted through the thin film under test is acquired simultaneously to obtain the reference transmission signal corresponding to the reference band and the moisture transmission signal corresponding to the moisture-sensitive band. The reference transmission signal and the moisture transmission signal are preprocessed simultaneously to obtain the preprocessed reference transmission signal and the preprocessed moisture transmission signal. Thickness features are obtained based on the preprocessed reference transmission signal; Moisture characteristics are obtained based on the preprocessed moisture transmission signal; The thickness feature is correlated with the moisture feature to determine the moisture feature offset. The moisture characteristics are processed based on the moisture characteristic offset to obtain the water content characteristics corresponding to the thickness change of the film under test. The thickness detection results and moisture content detection results of the film under test are obtained based on the thickness characteristics and water content characteristics.

2. The method for detecting film thickness and moisture content based on near-infrared transmission according to claim 1, characterized in that, Acquire the thin film to be tested and illuminate it using a near-infrared broadband light source. Form at least two near-infrared acquisition channels corresponding to the same measurement area on the transmission side of the thin film, including: The thin film to be tested is obtained and placed in the sample detection area between the near-infrared broadband light source and the receiver, so that the test area of ​​the thin film is located in the outgoing light path of the near-infrared broadband light source, and near-infrared radiation light is obtained. Near-infrared radiation is collimated and expanded to form a uniform illumination spot in the area to be tested. The uniform illumination spot covers the area to be tested and passes through the thin film to be tested, forming the corresponding transmitted light field of the area to be tested. At least two near-infrared acquisition channels are set on the transmission side of the transmitted light field, so that each near-infrared acquisition channel corresponds to the transmitted light field of the area to be measured. Different bands of near-infrared light in the transmitted light field are spectrally selected. The different bands of near-infrared light selected by each near-infrared acquisition channel are focused and converted into corresponding photoelectric signals, forming at least two near-infrared acquisition channels for the same area to be measured.

3. The method for detecting film thickness and moisture content based on near-infrared transmission according to claim 1, characterized in that, Based on at least two near-infrared acquisition channels, the near-infrared light transmitted through the thin film under test is simultaneously acquired to obtain the reference transmission signal corresponding to the reference band and the moisture transmission signal corresponding to the moisture-sensitive band, including: Acquire the acquisition bands corresponding to at least two near-infrared acquisition channels, acquire the moisture absorption characteristics of each acquisition band, and determine at least one near-infrared acquisition channel as the reference acquisition channel and at least one near-infrared acquisition channel as the moisture acquisition channel. The reference band near-infrared light and the moisture-sensitive band near-infrared light transmitted from the same test area of ​​the thin film under test are obtained based on the reference acquisition channel and the moisture acquisition channel, and synchronous reception is performed at the same sampling time to obtain the reference band optical signal and the moisture-sensitive band optical signal. The reference band optical signal and the moisture-sensitive band optical signal are converted into photoelectric signals and simultaneously acquired to obtain the reference transmission signal and the moisture transmission signal.

4. The method for detecting film thickness and moisture content based on near-infrared transmission according to claim 1, characterized in that, The reference transmission signal and the moisture transmission signal are simultaneously preprocessed to obtain the preprocessed reference transmission signal and the preprocessed moisture transmission signal, including: Dark current correction is performed on the reference transmission signal and the moisture transmission signal to obtain the dark current corrected reference transmission signal and the moisture transmission signal. Based on a preset reference light intensity, the reference transmission signal and the moisture transmission signal after dark current correction are normalized to obtain a standardized transmission signal corresponding to the transmission characteristics of the film under test. The standardized transmission signal is synchronously filtered to obtain the preprocessed reference transmission signal and the preprocessed moisture transmission signal.

5. The method for detecting film thickness and moisture content based on near-infrared transmission according to claim 1, characterized in that, Thickness features are obtained based on the preprocessed reference transmission signal, including: The transmittance of the thin film under test in the reference band is determined based on the preprocessed reference transmission signal. The transmission attenuation characteristics of the thin film under test are determined based on the transmittance in the reference band, and the thickness parameters are determined based on the transmission attenuation characteristics. Obtain the preset thickness calibration data and combine it with the thickness parameters to obtain the thickness characteristics of the film under test.

6. The method for detecting film thickness and moisture content based on near-infrared transmission according to claim 1, characterized in that, Moisture characteristics are obtained based on the preprocessed moisture transmission signal, including: The transmittance of the film under test in the moisture-sensitive band is determined based on the pre-processed moisture transmission signal, and the moisture-sensitive transmission characteristics are obtained. Determine the moisture absorption characteristics corresponding to the moisture-sensitive band based on the moisture-sensitive transmission characteristics. The initial moisture characteristics of the film under test are obtained based on the moisture absorption characteristics. Obtain the preset moisture calibration data and combine it with the initial moisture characteristics to obtain the moisture characteristics of the film under test.

7. The method for detecting film thickness and moisture content based on near-infrared transmission according to claim 1, characterized in that, Correlating thickness features with moisture features to determine moisture feature offset includes: The sampling times corresponding to thickness and moisture features are obtained, and the thickness and moisture features corresponding to the same area to be measured are synchronously paired to obtain thickness and moisture sampling data. Based on thickness moisture sampling data, the thickness characteristic change and the corresponding moisture characteristic change are determined, and based on the thickness characteristic change and the moisture characteristic change, the additional change in moisture characteristics caused by the change in film thickness is determined. Based on the change in thickness characteristics and the additional change, the degree of shift of the moisture characteristics relative to the corresponding thickness state is determined, and the moisture characteristic shift amount is obtained.

8. The method for detecting film thickness and moisture content based on near-infrared transmission according to claim 1, characterized in that, Moisture characteristics are processed based on moisture feature offset to obtain water content characteristics corresponding to changes in the thickness of the film under test. Based on the thickness and water content characteristics, the thickness and moisture content detection results of the film under test are obtained, including: The moisture characteristics are corrected based on the moisture characteristic offset to obtain the corrected moisture characteristics; Obtain the change in moisture feature corresponding to the corrected moisture feature and the moisture feature offset; Based on the change in moisture characteristics, the corrected moisture characteristics are continuously sampled to obtain the moisture content characteristics corresponding to the change in the thickness of the film under test. Obtain preset thickness calibration data and preset moisture content calibration data. Determine the thickness detection result of the film to be tested based on the thickness characteristics and preset thickness calibration data, and determine the moisture content detection result of the film to be tested based on the moisture content characteristics and preset moisture content calibration data.

9. A thin film thickness and moisture content detection system based on near-infrared transmission, applied to the thin film thickness and moisture content detection method based on near-infrared transmission as described in any one of claims 1 to 8, characterized in that, include: The acquisition channel module is used to acquire the film under test and uses a near-infrared broadband light source to illuminate the film under test through transmission, forming at least two near-infrared acquisition channels corresponding to the same measurement area on the transmission side of the film under test. The transmission signal acquisition module, based on at least two near-infrared acquisition channels, synchronously acquires the near-infrared light transmitted through the thin film under test, and obtains the reference transmission signal corresponding to the reference band and the moisture transmission signal corresponding to the moisture-sensitive band. The signal preprocessing module is used to simultaneously preprocess the reference transmission signal and the moisture transmission signal to obtain the preprocessed reference transmission signal and the preprocessed moisture transmission signal. The thickness feature acquisition module acquires thickness features based on the preprocessed reference transmission signal. The moisture feature acquisition module acquires moisture features based on the preprocessed moisture transmission signal. The feature offset module is used to associate thickness features with moisture features and determine the moisture feature offset. The detection results module processes the moisture characteristics based on the moisture characteristic offset to obtain the water content characteristics corresponding to the thickness change of the film under test, and obtains the thickness detection results and moisture content detection results of the film under test based on the thickness characteristics and water content characteristics.

10. A thin film thickness and moisture content detection terminal based on near-infrared transmission, characterized in that, include: One or more processors; A storage device on which one or more programs are stored; When one or more programs are executed by one or more processors, the one or more processors implement the near-infrared transmission-based thin film thickness and moisture content detection method according to any one of claims 1 to 8.