A method and system for rapid detection of degradable plastic product ingredients based on near-infrared spectroscopy

Abstract

This invention relates to the field of spectral analysis technology, and discloses a rapid detection method and system for the components of biodegradable plastic products based on near-infrared spectroscopy. The method includes: acquiring a near-infrared full-band absorbance sequence as raw spectral data; obtaining corrected spectral data after baseline correction; determining the reference, polyester, and starch response regions based on the characteristic absorption peak positions and half-widths (WHMs); calculating the area enclosed by the absorbance curves and wavelength axes within the three regions to obtain the integrated absorbance of the reference, polyester, and starch; locating the relative position of the starch integrated absorbance within the reference and polyester integrated absorbance intervals to obtain a relative starch content index; calculating the ratio of the polyester to the reference integrated absorbance as a relative polyester content index; and combining the two indices to form a component feature vector. This invention can improve the efficiency of rapid detection of the components of biodegradable plastic products based on near-infrared spectroscopy.

Description

Technical Field

[0001] This invention relates to the field of spectral analysis technology, and in particular to a rapid detection method and system for the components of biodegradable plastic products based on near-infrared spectroscopy. Background Technology

[0002] Current methods for detecting the components of biodegradable plastic products mainly rely on chemical titration, thermogravimetric analysis, or gas chromatography-mass spectrometry. These methods typically require complex pretreatment procedures such as sample crushing, dissolution, extraction, or high-temperature pyrolysis, making the entire detection process time-consuming and unable to meet the needs of rapid sampling or real-time on-site testing in production lines or recycling and sorting. Furthermore, these destructive testing methods directly consume the samples, rendering them unusable and increasing testing costs and waste generation. In addition, the organic solvents and standard reagents used in chemical methods are expensive and require specialized personnel to operate; the repeatability of the test results is significantly affected by the operator's skill level, making standardization and automation difficult.

[0003] Existing non-destructive detection methods based on spectroscopy also have significant drawbacks. Conventional near-infrared spectroscopy relies on a large number of standard samples with known component contents to establish a multivariate calibration model. However, the preparation and calibration of standard samples are time-consuming and labor-intensive, and the established model is only applicable to samples with the same matrix type as the modeled sample. When the formulation or production process of biodegradable plastics changes, the model's predictive accuracy drops sharply. Another common method uses the peak height ratio or peak area ratio within a fixed wavelength range to estimate component content. However, the spectral baseline drift and scattering effects of different samples can cause large fluctuations in peak height and peak area. Furthermore, a fixed wavelength range cannot accommodate small shifts in the characteristic absorption peak positions between different samples, making it difficult to guarantee the accuracy and stability of the detection results. Therefore, existing technologies cannot achieve rapid, accurate, non-destructive, and model-maintenance-free quantitative detection of polyester and starch components in biodegradable plastic products. Summary of the Invention

[0004] This invention provides a rapid detection method and system for the components of biodegradable plastic products based on near-infrared spectroscopy, the main purpose of which is to solve the problem of low efficiency in the rapid detection of the components of biodegradable plastic products based on near-infrared spectroscopy.

[0005] To achieve the above objectives, this invention provides a rapid detection method for the components of biodegradable plastic products based on near-infrared spectroscopy, comprising: The absorbance sequence of the biodegradable plastic product to be tested in the near-infrared full band during the target process is obtained to obtain the original spectral data of the target process; Baseline correction is performed on the original spectral data to obtain the corrected spectral data for the target process; Based on the peak positions and half-widths of the characteristic absorption peak clusters in the corrected spectral data, the reference absorption region, polyester response region, and starch response region of the target process are determined respectively. The area enclosed by the absorbance curves in the reference absorption region, the polyester response region, and the starch response region and the wavelength axis is used as the reference integrated absorbance, polyester response integrated absorbance, and starch response integrated absorbance of the target process. Within the numerical range formed by the baseline integrated absorbance and the polyester response integrated absorbance, the relative position of the starch response integrated absorbance is located to obtain the relative starch content index of the target process. The relative polyester content index of the target process is determined based on the polyester response integrated absorbance and the baseline integrated absorbance, and the relative starch content index and the relative polyester content index are combined to obtain the component feature vector of the target process.

[0006] In a preferred embodiment, obtaining the absorbance sequence of the biodegradable plastic product under test in the near-infrared full-band during the target process, and obtaining the raw spectral data of the target process, includes: Receive the discrete wavelength point sequence collected in the near-infrared full-band range of the biodegradable plastic product to be tested; The absorbance values ​​corresponding to the discrete wavelength points in the discrete wavelength point sequence are arranged in ascending order of wavelength to obtain the original spectral data of the target process.

[0007] In a preferred embodiment, the baseline correction of the original spectral data to obtain the corrected spectral data for the target process includes: In the raw spectral data, identify wavelength pairs located outside the characteristic absorption peaks of the biodegradable plastic component in the biodegradable plastic product to be tested; The spectral background baseline of the target process is determined using the wavelength point pairs and their corresponding absorbance values ​​as endpoints. Based on the spectral background baseline, the absorbance value at the location with the same wavelength value as the wavelength point in the original spectral data is obtained to obtain the background absorbance value of the target process; The difference sequence between the original absorbance value at the wavelength point and the background absorbance value is used as the corrected spectral data for the target process.

[0008] In a preferred embodiment, determining the spectral background baseline of the target process using the wavelength point pairs and corresponding absorbance values ​​as endpoints includes: The wavelength value and corresponding absorbance value of the first wavelength point in the wavelength point pair are used as the starting wavelength and starting absorbance, and the wavelength value and corresponding absorbance value of the second wavelength point in the wavelength point pair are used as the ending wavelength and ending absorbance. The wavelength difference between the starting wavelength and the ending wavelength, and the absorbance difference between the starting absorbance and the ending absorbance are obtained. The wavelength offset between the wavelength value of the wavelength point in the original spectral data and the starting wavelength is obtained. The ratio of the wavelength offset to the wavelength difference is taken as the offset ratio, and the product of the offset ratio and the absorbance difference is taken as the absorbance offset share of the target process. The initial absorbance is added to the absorbance offset fraction to obtain the background absorbance value of the target process; By sequentially connecting all wavelength points and the points determined by their background absorbance values, the spectral background baseline of the target process is obtained.

[0009] In a preferred embodiment, determining the reference absorption region, polyester response region, and starch response region of the target process based on the peak positions and full width at half maximum (FWHM) of the characteristic absorption peak clusters in the corrected spectral data includes: The discrete data points of the corrected spectral data are arranged in order of wavelength from shortest to longest to obtain the discrete data point sequence of the target process; The data points in the discrete data point sequence with absorbance values ​​higher than the adjacent data points are taken as local maxima, and recorded as the reference peak position, polyester response peak position, and starch response peak position of the target process, respectively. Centered on the reference peak, the polyester response peak, and the starch response peak, respectively, extend towards the wavelength decreasing direction and the wavelength increasing direction until the local minimum value of the absorbance is located, thus obtaining the reference valley point pair, polyester valley point pair, and starch valley point pair of the target process; The wavelengths corresponding to the left and right valley points in the reference valley point pair, the polyester valley point pair, and the starch valley point pair are respectively used as the left boundary wavelength and the right boundary wavelength to obtain the reference absorption region, polyester response region, and starch response region of the target process.

[0010] In a preferred embodiment, the step of using the area enclosed by the absorbance curves in the reference absorption region, the polyester response region, and the starch response region and the wavelength axis as the reference integrated absorbance, polyester response integrated absorbance, and starch response integrated absorbance of the target process includes: Using the wavelength and absorbance values ​​of the corrected spectral data in the target process as the abscissa and ordinate, a wavelength absorbance plane for the target process is established, and the discrete data points in the reference absorption region, the polyester response region, and the starch response region are marked as coordinate points in the wavelength absorbance plane according to the wavelength and absorbance values. Connecting adjacent coordinate points in ascending order of wavelength in the wavelength absorbance plane yields the absorbance curve of the target process; Starting from the initial data point on the absorbance curve, adjacent data points are extracted sequentially to obtain the adjacent data point pairs of the target process and the corresponding left and right points; Using the absorbance value of the left point as the height of the left side of the trapezoid, the absorbance value of the right point as the height of the right side of the trapezoid, and the wavelength interval between the left point and the right point as the width of the trapezoid, a trapezoid for the target process is constructed. The vertices of the trapezoid are: the left point on the absorbance curve, the projection point of the left point on the wavelength axis of the wavelength absorbance plane, the projection point of the right point on the wavelength axis of the wavelength absorbance plane, and the right point on the absorbance curve. The areas of the trapezoids, the baseline integrated absorbance, the polyester response integrated absorbance, and the starch response integrated absorbance of the target process are summed up respectively.

[0011] In a preferred embodiment, the step of locating the relative position of the starch response integral absorbance within the numerical range formed by the reference integral absorbance and the polyester response integral absorbance to obtain the relative starch content index of the target process includes: The lower limit and upper limit of the interval for the target process are determined based on the ranking results of the baseline integrated absorbance and the polyester response integrated absorbance. Using the lower limit and the upper limit of the interval as endpoints, construct the closed interval of the target process; The relative offset of the starch response integral absorbance within the closed interval is used as the relative starch content index of the target process.

[0012] In a preferred embodiment, the formula for calculating the relative starch content index includes: in, This is an indicator of the relative starch content. The integral absorbance of the starch response. The baseline integrated absorbance, The integral absorbance of the polyester response.

[0013] In a preferred embodiment, determining the relative polyester content index of the target process based on the polyester response integrated absorbance and the reference integrated absorbance, and combining the relative starch content index and the relative polyester content index to obtain the component feature vector of the target process, includes: The ratio of the polyester response integrated absorbance to the baseline integrated absorbance is used as the relative polyester content index of the target process. Using the relative starch content index as the first component and the relative polyester content index as the second component, a component feature vector for the target process is constructed.

[0014] To address the aforementioned problems, the present invention also provides a rapid detection system for the composition of biodegradable plastic products based on near-infrared spectroscopy, the system comprising: The raw spectral data module acquires the absorbance value sequence of the biodegradable plastic product to be tested in the near-infrared full band during the target process, thus obtaining the raw spectral data of the target process; The calibration spectral data module performs baseline correction on the original spectral data to obtain the calibration spectral data for the target process; The partitioning module determines the reference absorption region, polyester response region, and starch response region of the target process based on the peak position and half-peak width of the characteristic absorption peak clusters in the corrected spectral data. The three-zone integration module uses the area enclosed by the absorbance curves in the reference absorption zone, the polyester response zone, and the starch response zone and the wavelength axis as the reference integrated absorbance, polyester response integrated absorbance, and starch response integrated absorbance of the target process. The starch index module locates the relative position of the starch response integral absorbance within the numerical range formed by the baseline integral absorbance and the polyester response integral absorbance, thereby obtaining the relative starch content index of the target process. The component feature module determines the relative polyester content index of the target process based on the polyester response integrated absorbance and the benchmark integrated absorbance, and combines the starch relative content index with the polyester relative content index to obtain the component feature vector of the target process.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. This detection method adaptively determines the reference absorption region, polyester response region, and starch response region through baseline correction and the peak position and half-width of characteristic absorption peak clusters. It eliminates the need for pre-setting fixed wavelength ranges or relying on a large number of standard samples to establish a calibration model, significantly improving its adaptability to biodegradable plastic products with different formulations and production processes. The method uses a trapezoidal integration method to accumulate the area enclosed by the absorbance curves and wavelength axis within each absorption region, obtaining the reference integrated absorbance, polyester response integrated absorbance, and starch response integrated absorbance. The integration process constructs trapezoids segment by segment based on measured discrete data points and accumulates the areas, accurately restoring the actual absorption intensity of each absorption region and eliminating integration errors caused by peak shape asymmetry or peak position shift. The relative offset of the starch response integrated absorbance within the closed interval formed by the reference integrated absorbance and the polyester response integrated absorbance is used as an indicator of starch relative content, achieving a normalized expression of starch content. This indicator is automatically limited to the range of zero to one, eliminating the influence of sample thickness and optical path differences, and ensuring the stability and repeatability of detection results for different batches of samples.

[0016] 2. This detection method uses the ratio of the polyester response integrated absorbance to the reference integrated absorbance as an indicator of the relative polyester content. This ratio operation effectively suppresses multiplicative interference during spectral acquisition, ensuring that the polyester content estimation is independent of external calibration curves. A component feature vector is constructed using the starch relative content index as the first component and the polyester relative content index as the second component. This feature vector comprehensively characterizes the relative composition of the two main biodegradable components in biodegradable plastic products in a two-dimensional numerical form, providing a unified and calculable numerical basis for subsequent sample classification, quality judgment, and recycling sorting. The entire detection process only requires the acquisition of near-infrared full-band spectral data, requires no sample pretreatment, consumes no chemical reagents, is fast, and does not damage the sample, making it suitable for rapid sampling inspections on production lines and real-time on-site detection scenarios. Attached Figure Description

[0017] Figure 1 This is a schematic flowchart of a rapid detection method for the components of biodegradable plastic products based on near-infrared spectroscopy provided in an embodiment of the present invention; Figure 2 This is a functional block diagram of a rapid detection system for biodegradable plastic product components based on near-infrared spectroscopy, provided in an embodiment of the present invention. The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] This application provides a rapid detection method for the components of biodegradable plastic products based on near-infrared spectroscopy. The execution entity of this rapid detection method for the components of biodegradable plastic products based on near-infrared spectroscopy includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the rapid detection method for the components of biodegradable plastic products based on near-infrared spectroscopy can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0020] Reference Figure 1 The diagram shown is a flowchart illustrating a rapid detection method for the components of biodegradable plastic products based on near-infrared spectroscopy, according to an embodiment of the present invention. In this embodiment, the rapid detection method for the components of biodegradable plastic products based on near-infrared spectroscopy includes: In this embodiment of the invention, when obtaining the absorbance value sequence of the biodegradable plastic product under test in the near-infrared full band during the target process to obtain the original spectral data of the target process, it is specifically used for: Receive the discrete wavelength point sequence collected in the near-infrared full-band range of the biodegradable plastic product to be tested; The absorbance values ​​corresponding to the discrete wavelength points in the discrete wavelength point sequence are arranged in ascending order of wavelength to obtain the original spectral data of the target process.

[0021] Specifically, the operator places the biodegradable plastic product to be tested in the sample cell of the near-infrared spectrometer and starts the spectrometer to perform continuous scanning across the preset near-infrared full-band range. The light source inside the spectrometer emits incident light covering the entire near-infrared band. This incident light passes through the biodegradable plastic product and is received by the spectrometer's spectroscopic system and detector. The detector collects discrete wavelength points one by one across the entire near-infrared full-band range at fixed wavelength intervals, with each discrete wavelength point corresponding to a specific wavelength value.

[0022] Specifically, the system reads the stored discrete wavelength point sequence, which may be arranged in any order as output by the spectrometer, such as in descending order of wavelength or in chronological order of acquisition time. The data acquisition system extracts the wavelength value of each discrete wavelength point in the sequence and places all these wavelength values ​​into a temporary array for comparison.

[0023] Furthermore, for each discrete wavelength point, the detector measures the transmitted or reflected light intensity and converts the light intensity into an absorbance value according to the Lambert-Beer law. The spectrometer's data output port outputs a set of discrete wavelength points and their corresponding absorbance values ​​in ascending or descending order of wavelength. These discrete wavelength points are not arranged continuously but exist as a discrete sequence. The data acquisition system receives this set of discrete wavelength point sequences from the spectrometer in real time via data cable or wireless transmission. Each element in this sequence contains two attributes: a wavelength value and an absorbance value at that wavelength. The data acquisition system stores the received discrete wavelength point sequence in memory or on the hard drive as is, without any sorting or filtering, maintaining the original order of the spectrometer output.

[0024] Furthermore, the data acquisition system employs a bubble sort method to sort the wavelength values ​​in the temporary array from smallest to largest. The bubble sort method involves repeatedly traversing the temporary array, comparing adjacent wavelength values ​​one by one. If the preceding wavelength value is greater than the following wavelength value, their positions are swapped. In each round of traversal, the largest wavelength value in the currently unsorted portion is moved to the end. After multiple rounds of traversal, all wavelength values ​​are arranged in ascending order. Simultaneously with each swap of two wavelength values, the data acquisition system synchronously swaps the absorbance values ​​corresponding to these two wavelength values, thus ensuring that the pairing relationship between wavelength values ​​and absorbance values ​​remains unchanged. After completing all sorting operations, the data acquisition system obtains a new sequence in which the wavelength values ​​of discrete wavelength points are arranged in ascending order, with each wavelength value followed by its corresponding absorbance value. This sequence arranged in ascending order of wavelength is named the original spectral data.

[0025] In summary, directly acquiring the raw measurement data output by the spectrometer preserves all discrete wavelength points and their corresponding absorbance information across the entire near-infrared band, providing a complete data foundation for subsequent spectral preprocessing and component analysis. Since the near-infrared spectrometer outputs discrete wavelength points rather than continuous curves, receiving the complete sequence of discrete wavelength points avoids human errors introduced by interpolation or resampling, ensuring the originality and authenticity of the spectral data. This receiving operation does not alter the physical meaning of the data and directly interfaces with the spectrometer's data interface, making the entire detection process compatible with different models of spectrometers, improving the method's versatility and portability. Simultaneously, storing the data in the form of a discrete wavelength point sequence provides a clear data structure for subsequent sorting, baseline correction, and integration calculations, facilitating rapid access to the absorbance value of each wavelength point by the data processing system using an index.

[0026] In summary, rearranging the original discrete wavelength point sequence in ascending order of wavelength value ensures strict wavelength monotonicity in the spectral data, providing a unified and standardized wavelength coordinate axis for subsequent baseline correction, characteristic absorption peak identification, and integration region division. Without this sorting operation, subsequent steps would require frequent searching and comparison of wavelength order when processing spectral data, significantly increasing computational complexity and the probability of errors. The sorted original spectral data maintains consistent wavelength intervals between adjacent data points, facilitating the accumulation calculation of integrated absorbance using the trapezoidal rule. This sorting operation also ensures that the spectral data of different samples have the same wavelength arrangement order, supporting batch processing and comparative analysis. Furthermore, the trend of absorbance variation with wavelength is clearly visible in the spectral data arranged in ascending wavelength order, facilitating manual or automatic identification of the position and shape of characteristic absorption peaks, and laying a reliable data structure foundation for the subsequent accurate determination of the reference absorption region, polyester response region, and starch response region.

[0027] In this embodiment of the invention, when performing baseline correction on the original spectral data to obtain the corrected spectral data for the target process, the specific steps are as follows: In the raw spectral data, identify wavelength pairs located outside the characteristic absorption peaks of the biodegradable plastic component in the biodegradable plastic product to be tested; The spectral background baseline of the target process is determined using the wavelength point pairs and their corresponding absorbance values ​​as endpoints. Based on the spectral background baseline, the absorbance value at the location with the same wavelength value as the wavelength point in the original spectral data is obtained to obtain the background absorbance value of the target process; The difference sequence between the original absorbance value at the wavelength point and the background absorbance value is used as the corrected spectral data for the target process.

[0028] Specifically, the system retrieves the wavelength ranges covered by the characteristic absorption peaks of known degradable plastic components in the tested degradable plastic product from a pre-constructed database of characteristic absorption peaks of plastic components. This database records the location ranges of characteristic absorption peaks of common degradable plastic components in the near-infrared band. The data processing system then iterates through each wavelength point in the original spectral data, determining whether the wavelength value of that point falls within the wavelength range of the characteristic absorption peaks of any known degradable plastic component.

[0029] Specifically, the two out-of-peak points with the smallest and largest wavelength values ​​are selected from the identified wavelength point pairs as global endpoints. Alternatively, all wavelength point pairs can be sequentially connected to form a segmented baseline; here, the method of sequentially connecting all wavelength point pairs is used. The data processing system arranges all identified wavelength point pairs in ascending order of wavelength. Each wavelength point pair contains two points: a left endpoint and a right endpoint. Each endpoint has a wavelength value and an absorbance value. For two adjacent wavelength point pairs, the right endpoint of the first wavelength point pair and the left endpoint of the second wavelength point pair are adjacent on the wavelength axis or have a gap. The data processing system uses a linear interpolation method to fill the background absorbance value between these two endpoints.

[0030] Specifically, the spectral background baseline is a broken line composed of multiple discrete points, each corresponding to a wavelength value and its background absorbance value. The data processing system reads the wavelength value carried by each wavelength point in the original spectral data; for example, the wavelength value of the first wavelength point in the original spectral data is a specific value. The data processing system searches for the location on the spectral background baseline that is exactly the same as this wavelength value. Since the spectral background baseline has already calculated the background absorbance value for each integer wavelength or each wavelength value appearing in the original spectral data using a linear interpolation method, the data processing system directly retrieves the corresponding background absorbance value from the spectral background baseline data table based on the wavelength value.

[0031] Specifically, the data processing system simultaneously reads the original spectral data and the already obtained background absorbance value sequence. These two sequences have the same number of data points and are in the exact same wavelength order. The system starts processing from the first data point, extracting the original absorbance value of the first wavelength point in the original spectral data, and then extracting the background absorbance value corresponding to the same wavelength position from the background absorbance value sequence. The system subtracts the background absorbance value from the original absorbance value to obtain the absorbance difference at that wavelength point. The system judges the difference obtained from this subtraction operation. If the difference is negative, it is directly retained without taking the absolute value or setting it to zero, because a negative difference indicates that the original absorbance at that wavelength point is lower than the background baseline, which is a possible physical phenomenon in actual spectra. The system sequentially performs the above subtraction operation on each wavelength point in the original spectral data, from the second wavelength point, the third wavelength point, until the last wavelength point, using the original absorbance value and the background absorbance value at the same wavelength position for each subtraction.

[0032] Furthermore, if the wavelength value falls within the characteristic absorption peak range, it is marked as an in-peak point; if the wavelength value falls outside all characteristic absorption peak ranges, it is marked as an out-of-peak point. The data processing system arranges all wavelengths marked as out-of-peak points into a candidate point sequence in ascending order of wavelength. The data processing system extracts two adjacent wavelengths from the candidate point sequence sequentially, forming a wavelength point pair between these two adjacent out-of-peak points. The data processing system continues to traverse the candidate point sequence, forming a wavelength point pair between every two adjacent out-of-peak points, until the entire candidate point sequence has been traversed, resulting in multiple wavelength point pairs. Each point in these wavelength point pairs is located outside the characteristic absorption peaks of all biodegradable plastic components; therefore, the absorbance values ​​of these wavelength point pairs only reflect the contribution of the spectral background and do not include the contribution of the component's characteristic absorption.

[0033] Furthermore, taking the right endpoint of the first wavelength pair as the starting point and the left endpoint of the second wavelength pair as the ending point, at each integer wavelength position between the starting and ending points, the data processing system calculates the total wavelength distance between the starting and ending points. Then, it calculates the wavelength offset from the starting point to the current wavelength position, divides the wavelength offset by the total wavelength distance to obtain a proportional value, multiplies this proportional value by the total absorbance difference between the starting and ending points, and finally adds the absorbance value of the starting point to the product to obtain the background absorbance value at that wavelength position. The data processing system performs the above linear interpolation operation on the intervals between all adjacent wavelength pairs, and also performs the same linear interpolation between the left and right endpoints within all wavelength pairs, thereby obtaining the background absorbance value corresponding to each wavelength point from the minimum to the maximum wavelength range. Connecting these background absorbance values ​​in wavelength order into a continuous broken line, this broken line is the spectral background baseline.

[0034] Furthermore, if a wavelength value in the original spectral data is exactly equal to the wavelength value at the midpoint of a certain wavelength point, the absorbance value at that midpoint is directly used as the background absorbance value. If the wavelength value in the original spectral data lies between two midpoints, the background absorbance value already calculated using linear interpolation at that location is used. The data processing system acquires the corresponding background absorbance value for each wavelength point in the original spectral data according to their order of appearance, and arranges these background absorbance values ​​into a new sequence according to the same wavelength order. Each element in this sequence corresponds one-to-one with a wavelength point in the original spectral data. The data processing system names this new sequence the background absorbance value sequence, and the length of this sequence is exactly the same as the length of the original spectral data, and the wavelength order is also completely consistent.

[0035] Furthermore, the data processing system arranges all calculated absorbance differences into a new sequence in ascending order of wavelength. The wavelength value of each data point in this sequence is consistent with the wavelength value in the original spectral data; only the absorbance value is replaced by the difference between the original value and the background value. The data processing system names this new sequence "corrected spectral data." This corrected spectral data eliminates the influence of the spectral background baseline, retaining only the net absorbance information contributed by the characteristic absorption of each component in the biodegradable plastic product. The data processing system outputs and stores the corrected spectral data for subsequent peak identification and partitioning steps.

[0036] In summary, this identification process accurately filters out wavelengths that contribute only to the spectral background and contain no characteristic absorptions of degradable plastic components, thus providing a clean background reference for subsequent background baseline construction. Since characteristic absorption peaks of degradable plastic components interfere with background estimation, directly fitting the background baseline using full-band data can lead to over- or under-subtraction of the background. By limiting the identified wavelength pairs to outside the characteristic absorption peaks, it ensures that the absorbance changes of these pairs are entirely caused by non-absorption factors such as light scattering and instrument drift, thereby improving the accuracy and physical plausibility of the background baseline. This identification process, based on a pre-built database of characteristic absorption peaks of plastic components, is adaptable to different types of degradable plastic products, eliminates the need for manual background point selection, and enhances the automation and repeatability of the method.

[0037] In summary, by using wavelength pairs outside the characteristic absorption peaks as endpoints, a continuous spectral background baseline covering the entire wavelength band is constructed through linear interpolation. This accurately describes the variation trend of background contribution caused by non-absorption factors in the original spectral data with wavelength. Since background interference usually changes slowly with wavelength, linear interpolation between adjacent peak pairs can approximate the true background curve with lower computational cost, avoiding overfitting or oscillations that may occur with high-order polynomial fitting. The background baseline determined by wavelength pairs as endpoints has a clear physical meaning; that is, the baseline is determined only by background points and is not affected by data from the characteristic absorption peak region, thus ensuring the objectivity and stability of background subtraction. This baseline construction method does not require pre-assuming the functional form of the background and is applicable to measured spectra with various complex background morphologies.

[0038] In summary, a background absorbance value was assigned to each wavelength point in the original spectral data, perfectly corresponding to its wavelength position, achieving precise alignment between the background baseline and the original spectral data in the wavelength dimension. Since the wavelength points in the original spectral data may be non-equally spaced or non-integer values, while the background baseline is a continuous polyline obtained through interpolation, the background absorbance values ​​were retrieved one by one according to the same wavelength, ensuring that the original absorbance and background absorbance strictly correspond to the same wavelength position in subsequent interpolation calculations. This point-by-point matching method avoids errors introduced by wavelength misalignment, ensuring that the background estimate for each wavelength point accurately reflects the interference intensity at that wavelength. The obtained background absorbance value sequence has the same length and wavelength order as the original spectral data, preparing the data for the next step of interpolation correction.

[0039] In summary, point-by-point subtraction effectively removes background interference from the original spectral data, yielding net absorbance information contributed solely by the characteristic absorption of each component in the biodegradable plastic product. The corrected spectral data eliminates additive interferences such as light scattering, sample thickness inhomogeneity, and instrument baseline drift, making spectral data comparable between different samples and batches. The difference sequence retains the positive and negative signs, accurately reflecting situations where the original absorbance is lower than the background baseline, thus avoiding information loss. This corrected spectral data is directly used for subsequent characteristic absorption peak localization and integrated absorbance calculation, improving the stability and accuracy of the integrated area. Compared to traditional derivative correction or standard normal transformation, this background subtraction method is based on physically interpretable out-of-peak linear interpolation, without altering the shape and position of characteristic absorption peaks, ensuring a linear response relationship for subsequent quantitative component analysis.

[0040] In this embodiment of the invention, when determining the spectral background baseline of the target process using the wavelength point pair and the corresponding absorbance value as endpoints, it is specifically used for: The wavelength value and corresponding absorbance value of the first wavelength point in the wavelength point pair are used as the starting wavelength and starting absorbance, and the wavelength value and corresponding absorbance value of the second wavelength point in the wavelength point pair are used as the ending wavelength and ending absorbance. The wavelength difference between the starting wavelength and the ending wavelength, and the absorbance difference between the starting absorbance and the ending absorbance are obtained. The wavelength offset between the wavelength value of the wavelength point in the original spectral data and the starting wavelength is obtained. The ratio of the wavelength offset to the wavelength difference is taken as the offset ratio, and the product of the offset ratio and the absorbance difference is taken as the absorbance offset share of the target process. The initial absorbance is added to the absorbance offset fraction to obtain the background absorbance value of the target process; By sequentially connecting all wavelength points and the points determined by their background absorbance values, the spectral background baseline of the target process is obtained.

[0041] Specifically, the data processing system selects one wavelength pair currently being processed from the multiple identified wavelength pairs. This wavelength pair contains two wavelength points arranged in ascending order of wavelength. The data processing system designates the point with the smaller wavelength value among these two wavelength points as the first wavelength point, extracts the wavelength value of the first wavelength point as the starting wavelength, and simultaneously extracts the absorbance value corresponding to the first wavelength point as the starting absorbance.

[0042] Specifically, the system reads the wavelength value of the wavelength point from the original spectral data where the background absorbance value needs to be calculated; this wavelength value is called the current wavelength. The data processing system subtracts the obtained starting wavelength value from the current wavelength value; the difference is called the wavelength offset, which represents the distance between the current wavelength and the starting wavelength. The data processing system divides the wavelength offset by the previously calculated wavelength difference; the result of this division is called the offset ratio, which is a value between zero and one. The offset ratio is zero when the current wavelength equals the starting wavelength, and one when the current wavelength equals the ending wavelength.

[0043] Specifically, the obtained initial absorbance value is added to the newly calculated absorbance offset value. When the absorbance offset is positive, the result is greater than the initial absorbance; when the absorbance offset is negative, the result is less than the initial absorbance.

[0044] Specifically, the data processing system has obtained the background absorbance value for every wavelength point from the minimum to the maximum wavelength range. These data points constitute a set of points on a two-dimensional plane, where the x-coordinate of each point is the wavelength value and the y-coordinate is the background absorbance value. The data processing system extracts these points in ascending order of wavelength value, starting from the minimum wavelength point and ending at the maximum wavelength point. The system connects adjacent points with straight line segments. Specifically, it extracts the first and second points and connects them with a straight line segment on the two-dimensional plane; then it extracts the second and third points and connects them with a straight line segment.

[0045] Furthermore, the point with the larger wavelength value among the two wavelength points is called the secondary wavelength point. The wavelength value of this secondary wavelength point is extracted as the termination wavelength, and the corresponding absorbance value is extracted as the termination absorbance. The data processing system performs a subtraction operation on the starting and termination wavelengths, subtracting the starting wavelength value from the termination wavelength value; the resulting difference is called the wavelength difference. The data processing system also performs a subtraction operation on the starting and termination absorbances, subtracting the starting absorbance value from the termination absorbance value; the resulting difference is called the absorbance difference. The data processing system temporarily stores these two differences in memory for subsequent offset ratio calculations.

[0046] Furthermore, the offset ratio is multiplied by the previously calculated absorbance difference. The result of this multiplication is called the absorbance offset share, which represents the proportion of the total change in absorbance from the initial to the final wavelength, allocated according to the relative position of the current wavelength. When performing the multiplication, the data processing system directly multiplies the offset ratio by the absorbance difference, retaining the sign of the product. If the absorbance difference is negative, the absorbance offset share is also negative.

[0047] Furthermore, the value obtained from the addition operation is the absorbance value at the spectral background baseline at the current wavelength point, which the data processing system refers to as the background absorbance value. The data processing system repeatedly performs the calculations of wavelength offset, offset ratio, absorbance offset fraction, and addition of the initial absorbance and absorbance offset fraction for each wavelength point in the original spectral data, thus obtaining a corresponding background absorbance value for each wavelength point. The data processing system arranges these background absorbance values ​​into a sequence in ascending order of wavelength, with each background absorbance value in this sequence corresponding one-to-one with a wavelength point at the same wavelength position in the original spectral data.

[0048] Furthermore, this process is repeated until all adjacent point pairs are connected. These straight line segments, when joined end to end, form a continuous polygonal line covering every wavelength position across the entire near-infrared band. The data processing system names this continuous polygonal line the spectral background baseline, which reflects the overall trend of background contribution from factors such as light scattering and instrument response in the original spectral data as a function of wavelength. The data processing system stores the spectral background baseline in the form of a data table, where each row records a wavelength value and its corresponding background absorbance value, used for subsequent calculations of corrected spectral data.

[0049] In summary, endpoint information for each wavelength pair was extracted, providing precise starting and ending conditions for subsequent linear interpolation calculations. By calculating the wavelength and absorbance differences, the total variation between the two endpoints along the wavelength and absorbance axes was quantified, allowing the background absorbance value at any subsequent intermediate wavelength point to be proportionally determined based on these two total variations. This endpoint parameterization method avoids the use of complex curve fitting models, relying solely on actual out-of-peak data from the original spectrum. This ensures that the background baseline construction process is entirely driven by measured data, exhibiting high repeatability and physical transparency. Furthermore, pre-calculating and storing the wavelength and absorbance differences reduces repetitive calculations at each subsequent wavelength point, improving data processing efficiency.

[0050] In summary, the wavelength offset reflects the distance of the current wavelength point relative to the starting wavelength, while the offset ratio normalizes this distance to a range of zero to one, representing the relative position of the current wavelength point within the interval from the starting wavelength to the ending wavelength. Multiplying the offset ratio by the absorbance difference, i.e., extracting the corresponding share from the total absorbance change according to the relative position, ensures the linear gradual change of background absorbance with wavelength. This linear allocation method fully conforms to the physical law of the slow change of the spectral background baseline with wavelength, avoiding spurious fluctuations that may be introduced by nonlinear interpolation. Furthermore, this step does not rely on any empirical parameters, using only the geometric information of the current wavelength point pair, making the background estimation result objective and unique.

[0051] In summary, by directly adding the initial absorbance to the absorbance offset, an accurate estimate of the background absorbance for the current wavelength point is generated. When the wavelength offset is zero, the sum equals the initial absorbance; when the wavelength offset equals the wavelength difference, the sum equals the final absorbance; for any intermediate wavelength point, the sum lies strictly on the linear line between the initial and final absorbance. This accumulation method ensures that the background absorbance value changes continuously and monotonically throughout the entire wavelength range, without jumps or discontinuities. Since the background absorbance value for each wavelength point is calculated based on the two endpoints of the same wavelength pair using the same rules, the entire background baseline exhibits consistent smoothness both within and between wavelength pairs, providing a high-precision benchmark for subsequent background subtraction.

[0052] In summary, connecting the discretely calculated background absorbance values ​​at each wavelength point into a continuous broken line forms a complete spectral background baseline covering the entire near-infrared band. This broken line visually demonstrates the overall trend of background absorbance variation with wavelength, facilitating visualization and verification of the reasonableness of the background estimation. Since the background absorbance values ​​between adjacent wavelength points are obtained through linear interpolation, each segment of the broken line is a straight line, resulting in a piecewise linear continuous curve that approximates the true optical background with minimal computation. This background baseline is directly used for subsequent difference calculations between the original absorbance and the background absorbance, ensuring that the net absorbance at each wavelength point in the corrected spectral data is calculated based on the same baseline standard. Compared to using global polynomial fitting, this piecewise linear baseline is more adaptable to local background fluctuations and does not distort the overall baseline shape due to data from characteristic absorption peak regions far from the background points, thereby improving the accuracy of background subtraction.

[0053] In this embodiment of the invention, when determining the reference absorption region, polyester response region, and starch response region of the target process based on the peak positions and half-widths of the characteristic absorption peak clusters in the corrected spectral data, it is specifically used for: The discrete data points of the corrected spectral data are arranged in order of wavelength from shortest to longest to obtain the discrete data point sequence of the target process; The data points in the discrete data point sequence with absorbance values ​​higher than the adjacent data points are taken as local maxima, and recorded as the reference peak position, polyester response peak position, and starch response peak position of the target process, respectively. Centered on the reference peak, the polyester response peak, and the starch response peak, respectively, extend towards the wavelength decreasing direction and the wavelength increasing direction until the local minimum value of the absorbance is located, thus obtaining the reference valley point pair, polyester valley point pair, and starch valley point pair of the target process; The wavelengths corresponding to the left and right valley points in the reference valley point pair, the polyester valley point pair, and the starch valley point pair are respectively used as the left boundary wavelength and the right boundary wavelength to obtain the reference absorption region, polyester response region, and starch response region of the target process.

[0054] Specifically, the system reads the stored calibrated spectral data, which contains multiple discrete data points. Each discrete data point consists of a wavelength value and a calibrated absorbance value at that wavelength. The data processing system extracts the wavelength values ​​of all discrete data points and arranges these wavelength values ​​in ascending order.

[0055] Specifically, the system reads a pre-arranged sequence of discrete data points. Each data point in the sequence has a preceding and a following data point. The sequence begins with the following data point and ends with the preceding data point. The data processing system iterates from the second data point to the penultimate data point, performing the following comparison operations for each current data point: retrieve the absorbance value of the current data point, retrieve the absorbance value of the preceding data point, and retrieve the absorbance value of the following data point. The system then determines whether the absorbance value of the current data point is simultaneously greater than both the absorbance values ​​of the preceding and following data points. If both inequalities are true, the current data point is marked as a local maximum.

[0056] Specifically, the valley point positioning operation is first performed with the reference peak as the center. Starting from the data point corresponding to the reference peak, the operation moves to the left along the direction of decreasing wavelength. Each time the operation moves to the previous data point, the absorbance value of the current data point is compared with the absorbance value of the adjacent data point to the left. When the absorbance value of a data point is less than the absorbance values ​​of both the adjacent data point to the left and the adjacent data point to the right, the data point is marked as a local minimum and is taken as the left valley point.

[0057] Specifically, the left valley point in the reference valley point pair is extracted, and its wavelength value is used as the left boundary wavelength of the reference absorption region. The data processing system extracts the right valley point in the reference valley point pair, extracts its wavelength value, and uses this wavelength value as the right boundary wavelength of the reference absorption region. Using the left and right boundary wavelengths as endpoints, the data processing system defines a continuous wavelength interval containing all wavelength positions from the left boundary wavelength to the right boundary wavelength; this wavelength interval is named the reference absorption region.

[0058] Furthermore, a bubble sort method is used to perform the permutation operation. Specifically, it iterates through all wavelength values, comparing adjacent wavelength values ​​sequentially. If a preceding wavelength value is greater than a following wavelength value, the two wavelength values ​​are swapped, along with their corresponding absorbance values. In each iteration, the largest wavelength value in the currently unsorted portion is moved to the end. After multiple iterations, all wavelength values ​​are arranged in ascending order. After the permutation is complete, the data processing system obtains a new sequence in which the wavelength values ​​of the discrete data points strictly increase from shortest to longest, with each wavelength value followed by its corresponding corrected absorbance value. The data processing system names this sequence, arranged by wavelength from shortest to longest, the discrete data point sequence.

[0059] Furthermore, the data processing system continues to traverse the entire discrete data point sequence to find all local maxima. The system retrieves the standard wavelength positions corresponding to the characteristic absorption peaks of the reference component, polyester component, and starch component from a pre-built database of characteristic absorption peak positions for plastic components, and matches these standard wavelength positions with the wavelength values ​​of each found local maxima. Successfully matched local maxima are recorded as the reference peak position, polyester response peak position, and starch response peak position, respectively, with each peak position containing two attributes: wavelength value and absorbance value.

[0060] Furthermore, starting from the reference peak, the data processing system moves to the right along the direction of increasing wavelength. Each time it moves to the next data point, it compares the absorbance value of the current data point with the absorbance value of the adjacent data point to its right. When a data point's absorbance value is simultaneously less than the absorbance values ​​of both its left and right adjacent data points, this data point is marked as a local minimum and designated as the right valley point. The data processing system combines the left and right valley points into a valley point pair, named the reference valley point pair. The data processing system repeats the same operation for the polyester response peak and the starch response peak, obtaining polyester valley point pairs and starch valley point pairs respectively. Each valley point pair contains both wavelength and absorbance values ​​for its left and right valley points.

[0061] Furthermore, the same operation is performed on the polyester valley point pairs. The wavelength value of the left valley point in the polyester valley point pair is extracted as the left boundary wavelength of the polyester response region, and the wavelength value of the right valley point in the polyester valley point pair is extracted as the right boundary wavelength of the polyester response region. The wavelength interval defined by these two boundary wavelengths is named the polyester response region. The data processing system performs the same operation on the starch valley point pairs. The wavelength value of the left valley point in the starch valley point pair is extracted as the left boundary wavelength of the starch response region, and the wavelength value of the right valley point in the starch valley point pair is extracted as the right boundary wavelength of the starch response region. The wavelength interval defined by these two boundary wavelengths is named the starch response region. The data processing system stores these three wavelength intervals—the reference absorption region, the polyester response region, and the starch response region—for subsequent integrated absorbance calculations.

[0062] In summary, reorganizing the discrete data points in the calibrated spectral data according to their wavelength values ​​in ascending order ensures that the entire spectral data has strictly monotonically increasing wavelength coordinates. This provides a unified and standardized sequential structure for subsequent searches of local extrema and the division of absorption intervals. Since the discrete wavelength points acquired by the spectrometer may not be output in strictly ascending order, this forced sorting eliminates potential order disorder or reverse order issues in the original data, avoiding errors in peak position identification due to wavelength order confusion. In the sorted discrete data point sequence, the wavelength intervals between adjacent data points are clear, facilitating the location of local maxima and minima using a simple adjacent comparison method. This sorting operation also ensures that the calibrated spectral data of different samples in the same batch have a completely consistent wavelength arrangement, supporting batch automated processing and comparative analysis, significantly improving the standardization and computational efficiency of data processing.

[0063] In summary, by comparing the absorbance values ​​of adjacent data points, the highest points of each characteristic absorption peak in the calibrated spectral data are accurately identified, providing precise central reference points for subsequent determination of the baseline absorption region, polyester response region, and starch response region. Since the characteristic absorption peaks of various components of biodegradable plastics in the near-infrared spectrum exhibit a Gaussian or Lorentzian distribution, with the highest absorbance values ​​at the peak positions, a local maximum point can be uniquely determined by judging whether the absorbance value of a data point is simultaneously greater than the absorbance values ​​of its left and right adjacent points. This method is simple, intuitive, and computationally inefficient. By matching with a pre-constructed database of characteristic absorption peak positions of plastic components, successfully matched local maximum points are recorded as the baseline peak position, polyester response peak position, and starch response peak position, respectively, achieving automatic assignment of the three types of peak positions without manual intervention. This peak position identification method is unaffected by the overall intensity scaling of the spectrum, relying only on the relative levels of absorbance values, thus exhibiting robustness and providing a reliable starting center for subsequent positioning of valley points extending from the peak position to both sides.

[0064] In summary, starting from the center of each characteristic absorption peak, the search proceeds point-by-point along both decreasing and increasing wavelength directions until a local minimum absorbance value is found, thus accurately determining the valley boundary points on both sides of each characteristic absorption peak. This method, extending from the peak to both sides, fully considers the independent shape and width of each characteristic absorption peak, adaptively determining the wavelength range covered by the peak and avoiding the range mismatch problem caused by using a fixed wavelength window. Since the width of characteristic absorption peaks in different samples or batches may vary due to factors such as crystallinity and blending degree, dynamically searching for local minimum points ensures that each sample has an absorption range that best matches its actual spectral shape. The criterion for judging a local minimum point is that the absorbance value of the current data point is simultaneously less than the absorbance values ​​of its two adjacent data points on the left and right. This criterion is symmetrical to the criterion for a local maximum point, making the peak-valley localization method inherently consistent and repeatable. The obtained baseline valley point pairs, polyester valley point pairs, and starch valley point pairs record the left and right boundary positions of each absorption range, defining an accurate wavelength range for the next step of integral area calculation.

[0065] In summary, the wavelengths of the left and right valley points in each valley point pair are directly extracted and used as the left and right boundary wavelengths of the corresponding absorption regions, thus constructing three non-overlapping or partially overlapping continuous wavelength intervals. These three absorption regions correspond to the characteristic absorption spectral regions of the reference component, polyester component, and starch component, respectively. The boundary of each absorption region is entirely determined by the valley point position of the measured spectrum of the sample, exhibiting sample-adaptive characteristics. Compared with using globally fixed wavelength intervals, this method based on valley point pair boundaries can effectively eliminate interference from adjacent absorption peaks, ensuring that each integration interval mainly contains the characteristic absorption contribution of the target component, while minimizing crosstalk from other components or background. Since the boundary wavelengths are directly extracted from the calibration spectral data, this process does not rely on any external parameters or manual settings, ensuring the consistency and objectivity of absorption intervals obtained by different operators and at different times. The clearly defined reference absorption region, polyester response region, and starch response region provide precise wavelength start and end ranges for subsequent trapezoidal integration, thereby improving the accuracy and repeatability of integrated absorbance calculations, ultimately making the starch and polyester relative content indices calculated based on these integrated absorbances more reliable.

[0066] In this embodiment of the invention, when the area enclosed by the absorbance curves in the reference absorption region, the polyester response region, and the starch response region and the wavelength axis is used as the reference integrated absorbance, polyester response integrated absorbance, and starch response integrated absorbance of the target process, it is specifically used for: Using the wavelength and absorbance values ​​of the corrected spectral data in the target process as the abscissa and ordinate, a wavelength absorbance plane for the target process is established, and the discrete data points in the reference absorption region, the polyester response region, and the starch response region are marked as coordinate points in the wavelength absorbance plane according to the wavelength and absorbance values. Connecting adjacent coordinate points in ascending order of wavelength in the wavelength absorbance plane yields the absorbance curve of the target process; Starting from the initial data point on the absorbance curve, adjacent data points are extracted sequentially to obtain the adjacent data point pairs of the target process and the corresponding left and right points; Using the absorbance value of the left point as the height of the left side of the trapezoid, the absorbance value of the right point as the height of the right side of the trapezoid, and the wavelength interval between the left point and the right point as the width of the trapezoid, a trapezoid for the target process is constructed. The vertices of the trapezoid are: the left point on the absorbance curve, the projection point of the left point on the wavelength axis of the wavelength absorbance plane, the projection point of the right point on the wavelength axis of the wavelength absorbance plane, and the right point on the absorbance curve. The areas of the trapezoids, the baseline integrated absorbance, the polyester response integrated absorbance, and the starch response integrated absorbance of the target process are summed up respectively.

[0067] Specifically, the wavelength and absorbance values ​​of each discrete data point in the calibrated spectral data are read, with the wavelength value used as the x-axis and the absorbance value as the y-axis. The data processing system creates a two-dimensional coordinate system in memory, where the horizontal axis represents wavelength and the vertical axis represents absorbance; this coordinate system is named the wavelength-absorbance plane.

[0068] Specifically, all marked coordinate points within the reference absorption region are extracted on the wavelength absorbance plane. These coordinate points are naturally arranged in ascending order of wavelength value. The data processing system starts with the coordinate point with the smallest wavelength value and connects it to the coordinate point with the second smallest wavelength value using a straight line segment. The data processing system then connects the coordinate point with the second smallest wavelength value to the coordinate point with the third smallest wavelength value using a straight line segment, and so on, until the coordinate point with the largest wavelength value is connected to the coordinate point preceding it.

[0069] Specifically, taking the absorbance curve within the reference absorption region as the processing object, the data point with the smallest wavelength value on the absorbance curve is found and used as the starting data point. Starting from the starting data point, the data processing system sequentially takes two adjacent data points along the direction of increasing wavelength. The first data point is called the left point, and the second data point is called the right point. These two points together form an adjacent data point pair.

[0070] Specifically, the first pair of adjacent data points within the reference absorption region is extracted. The absorbance value of the left point in this pair is extracted and used as the left height of the trapezoid. The data processing system extracts the absorbance value of the right point from the same pair and uses this value as the right height of the trapezoid. The data processing system calculates the difference between the wavelength values ​​of the left and right points and uses this difference as the width of the trapezoid. The data processing system constructs a quadrilateral in the wavelength absorbance plane. The first vertex of this quadrilateral is the position of the left point on the absorbance curve, possessing both the wavelength and absorbance values ​​of the left point. The second vertex of this quadrilateral is the projection of the left point onto the wavelength axis of the wavelength absorbance plane; the wavelength value of this projection is equal to the wavelength value of the left point, and the absorbance value is zero.

[0071] Specifically, the area of ​​the first trapezoid within the reference absorption region is calculated. The trapezoid's area is calculated by adding the height of the left side and the height of the right side of the trapezoid, multiplying the result by the width of the trapezoid, and then dividing the product by two. The data processing system adds the calculated area of ​​the first trapezoid to an accumulator with an initial value of zero. The data processing system then continues to calculate the area of ​​the second trapezoid within the reference absorption region and adds this area to the same accumulator as well.

[0072] Furthermore, from the calibration spectral data, all discrete data points whose wavelength values ​​fall between the left and right boundary wavelengths of the reference absorption region are selected. For each such data point, a position is found on the wavelength absorbance plane where the horizontal coordinate equals the wavelength value and the vertical coordinate equals the absorbance value. This position is then plotted as a coordinate point. The data processing system performs the same operation on the discrete data points within the polyester response region, marking each data point in the polyester response region as a coordinate point on the wavelength absorbance plane according to its wavelength and absorbance values. The data processing system also performs the same operation on the discrete data points within the starch response region, marking each data point in the starch response region as a coordinate point on the wavelength absorbance plane according to its wavelength and absorbance values. After the above marking, three sets of coordinate points appear on the wavelength absorbance plane, corresponding to the calibration spectral data in the reference absorption region, polyester response region, and starch response region, respectively.

[0073] Furthermore, these straight line segments, when connected end-to-end, form a continuous broken line, which the data processing system refers to as the absorbance curve. The data processing system performs the same connection operation on the coordinate points within the polyester response region, connecting adjacent coordinate points in ascending wavelength order to obtain the absorbance curve for the polyester response region. The data processing system also performs the same connection operation on the coordinate points within the starch response region, connecting adjacent coordinate points in ascending wavelength order to obtain the absorbance curve for the starch response region. The three absorbance curves are located within their respective wavelength ranges, and each curve reflects the continuous change in absorbance with wavelength.

[0074] Furthermore, the data processing system records the wavelength and absorbance values ​​of the left and right points in this pair of adjacent data points. The system continues moving forward, using the previously recorded right point as the new left point and the next data point after that as the new right point, forming a second pair of adjacent data points, and records the wavelength and absorbance values ​​of these two points. This process is repeated until all data points on the absorbance curve have been traversed, with the last pair of adjacent data points consisting of the second-to-last and the last points. The system ultimately obtains all pairs of adjacent data points on the absorbance curve within the reference absorption region, each pair containing a left and a right point. The system performs the same extraction operation on the absorbance curve within the polyester response region, obtaining all pairs of adjacent data points and their corresponding left and right points. The system also performs the same extraction operation on the absorbance curve within the starch response region, obtaining all pairs of adjacent data points and their corresponding left and right points.

[0075] Furthermore, the third vertex of the quadrilateral is the projection of the right point onto the wavelength axis of the wavelength absorbance plane. The wavelength value of this vertex is equal to the wavelength value of the right point, and the absorbance value is zero. The fourth vertex of the quadrilateral is the position of the right point on the absorbance curve, possessing both the wavelength and absorbance values ​​of the right point. These four vertices are connected sequentially to form a trapezoid, where the height of the left side is the left height, the height of the right side is the right height, the base extends from the left projection point to the right projection point on the wavelength axis, and the top side is a line segment on the absorbance curve connecting the left and right points. The data processing system performs the above construction operation on each adjacent data point pair within the reference absorption region, resulting in a series of trapezoids. The data processing system performs the exact same construction operation on adjacent data point pairs within the polyester response region and the starch response region, respectively, resulting in a series of trapezoids for each.

[0076] Furthermore, the area of ​​each trapezoid within the reference absorption region is calculated sequentially, and the area is accumulated in the accumulator after each calculation. Once the areas of all trapezoids within the reference absorption region have been accumulated, the value stored in the accumulator is the total area enclosed by the absorbance curve and the wavelength axis within the reference absorption region. The data processing system names this value the reference integrated absorbance. The data processing system performs the same accumulation operation on all trapezoids constructed within the polyester response region, calculating and accumulating the area sequentially starting from the first trapezoid. The sum obtained is named the polyester response integrated absorbance. The data processing system also performs the same accumulation operation on all trapezoids constructed within the starch response region, and the sum obtained is named the starch response integrated absorbance.

[0077] In summary, converting the numerical information in the calibrated spectral data into a set of geometric points on a two-dimensional plane provides an intuitive and operational graphical foundation for subsequent absorbance curve plotting and trapezoidal area calculation. By separately labeling the coordinate points within the reference absorption region, polyester response region, and starch response region, data from three different wavelength ranges are presented independently on the same plane, avoiding data confusion between regions. The establishment of the wavelength absorbance plane transforms the absorbance variation with wavelength into a visualized curve, facilitating the understanding of the spectral characteristics of each absorption region. Simultaneously, storing discrete data in the form of coordinate points provides explicit geometric elements for subsequent connection of adjacent points and trapezoidal construction. Each coordinate point carries both wavelength and absorbance values, ensuring a one-to-one correspondence between numerical values ​​and geometric positions during integration.

[0078] In summary, by connecting discrete coordinate points sequentially with straight line segments to form a continuous polygonal line, i.e., the absorbance curve, the discrete calibrated spectral data is transformed into a continuous function graph. Since absorbance in near-infrared spectroscopy typically varies continuously with wavelength, linear connections between adjacent points can approximate the true spectral curve with minimal error, while avoiding spurious oscillations that may be introduced by higher-order interpolation. This absorbance curve retains all the information of the original discrete data; the inflection points between adjacent line segments are the original data points and do not alter the original values. After the connection operation, the region enclosed by the absorbance curve and the wavelength axis is clearly defined, providing a direct geometric object for subsequent calculation of the area under the curve using the trapezoidal rule. Independent absorbance curves are obtained for the reference absorption region, polyester response region, and starch response region, allowing the integral area of ​​each component to be calculated independently without interference.

[0079] In summary, decomposing continuous data points on the absorbance curve into a series of adjacent point pairs, each consisting of a left and a right point, prepares the data for segment-by-segment calculation of the area under the curve using the trapezoidal method. Since the absorbance curve is formed by connecting adjacent points with straight line segments, the total area under the curve equals the sum of the areas of the trapezoids between all adjacent point pairs; therefore, extracting adjacent point pairs is a prerequisite for area accumulation. Extraction is performed sequentially starting from the initial data point, ensuring that each data point is used exactly once. The naming of the left and right points clearly distinguishes the height sources of the left and right sides of each trapezoid. This extraction process does not rely on any additional parameters, but is based solely on the increasing wavelength order, exhibiting complete determinism and repeatability. The same extraction operation is performed on the absorbance curves of the reference absorption region, polyester response region, and starch response region, ensuring that the integral calculations for the three regions use a completely consistent processing flow, guaranteeing the fairness and comparability of the calculation results.

[0080] In summary, by precisely constructing a right trapezoid for the region under the curve between each pair of adjacent points, the calculation of the integral area is transformed into the calculation of the trapezoid area, greatly simplifying the complexity of numerical integration. The height of the left side of the trapezoid is taken from the absorbance value of the left point, the height of the right side is taken from the absorbance value of the right point, and the width of the trapezoid is taken from the wavelength interval between the left and right points. These three geometric parameters are entirely determined by the original data points, requiring no approximation or assumptions. Of the four vertices of the trapezoid, two vertices lie on the absorbance curve, and two vertices lie on the wavelength axis. The height (width) of the trapezoid is along the wavelength axis, making the formula for calculating the area of ​​the trapezoid unique and definite. This construction method preserves the actual shape of the absorbance curve within each small interval. When the absorbance change between the left and right points is linear, the trapezoid area is exactly equal to the true area under the curve; when the absorbance change is non-linear, the trapezoid area is a reasonable approximation of the true area, and the approximation error decreases with increasing data point density. For each pair of adjacent points in the reference absorption region, polyester response region, and starch response region, trapezoids are constructed according to the same rules, ensuring the consistency of the integration method.

[0081] In summary, by summing the areas of all trapezoids within each absorption region, the total area enclosed by the absorbance curve and the wavelength axis within that region is obtained. This total area represents the integrated absorbance of the corresponding component, directly reflecting the total characteristic absorption intensity of that component in the sample. The trapezoidal surface accumulation method is a numerical integration method that does not require knowledge of the analytical expression of the absorbance curve; it can accurately calculate the area under the curve based solely on discrete data points. Since the area of ​​each trapezoid is directly calculated based on measured data points, the accumulation process does not introduce additional errors, and the accumulation result has a definite numerical meaning. After obtaining the baseline integrated absorbance, polyester response integrated absorbance, and starch response integrated absorbance, these three integral values ​​serve as the basic input for subsequent calculations of the relative starch content and polyester content indicators. Their accuracy directly determines the reliability of the final component detection results. Compared with the peak height method or peak area fitting method, the trapezoidal accumulation method does not assume a peak shape function, is applicable to various non-ideal peak shapes, and has better robustness to noise and baseline residuals. Meanwhile, this accumulation operation can automatically adapt to the differences in the width of different absorption regions. Wide absorption regions contain more trapezoids, while narrow absorption regions contain fewer trapezoids. The integration result naturally reflects the contribution of the absorption region width to the total absorption intensity, without the need for additional normalization processing.

[0082] In this embodiment of the invention, when determining the relative position of the starch response integral absorbance within the numerical range formed by the reference integral absorbance and the polyester response integral absorbance to obtain the relative starch content index of the target process, it is specifically used for: The lower limit and upper limit of the interval for the target process are determined based on the ranking results of the baseline integrated absorbance and the polyester response integrated absorbance. Using the lower limit and the upper limit of the interval as endpoints, construct the closed interval of the target process; The relative offset of the starch response integral absorbance within the closed interval is used as the relative starch content index of the target process.

[0083] Specifically, the system reads the calculated baseline integrated absorbance and polyester response integrated absorbance values ​​and compares them. The data processing system performs a numerical comparison operation, determining whether the baseline integrated absorbance is less than the polyester response integrated absorbance. If the comparison result is true, the baseline integrated absorbance is used as the lower limit of the interval, and the polyester response integrated absorbance is used as the upper limit. If the comparison result is false, meaning the baseline integrated absorbance is greater than the polyester response integrated absorbance, the polyester response integrated absorbance is used as the lower limit of the interval, and the baseline integrated absorbance is used as the upper limit. If the two values ​​are equal, both the lower limit and the upper limit of the interval are equal to the equal value.

[0084] Specifically, the already determined lower and upper limits of the interval are used as the two endpoints of the closed interval. The lower limit is the endpoint with the smaller value within the closed interval, and the upper limit is the endpoint with the larger value within the closed interval. The data processing system constructs a continuous range of values ​​on the numerical axis, starting from the lower limit and ending at the upper limit, and including both the lower and upper limits themselves within this range.

[0085] Specifically, the system reads the calculated starch response integral absorbance value, along with the lower and upper limits of the constructed closed interval. The data processing system first determines if the starch response integral absorbance value is less than the lower limit of the interval; if so, it directly assigns a value of zero to the starch relative content index. The system then determines if the starch response integral absorbance value is greater than the upper limit of the interval; if so, it directly assigns a value of one to the starch relative content index.

[0086] Furthermore, the assignment operation corresponding to this sorting result is recorded. The lower limit and upper limit of the interval are stored in two different variables. These two values ​​together determine the endpoint positions of the subsequent closed interval. The sorting result directly determines the specific numerical allocation of the lower limit and upper limit of the interval.

[0087] Furthermore, the data processing system names this range of values, from the lower limit to the upper limit and including both endpoints, as the closed interval. Every value within this closed interval is greater than or equal to the lower limit and less than or equal to the upper limit. The length of the closed interval is equal to the difference between the upper and lower limits. The data processing system stores this closed interval in memory as a data structure that records the lower and upper endpoints of the interval, used in the next step to calculate the relative offset of the starch response integral absorbance within this closed interval. This closed interval corresponds to the range of values ​​involved in the denominator of the subsequent calculation formula.

[0088] Furthermore, if the starch response integrated absorbance value lies within a closed interval, the data processing system performs a relative offset calculation: subtracting the lower limit of the interval from the starch response integrated absorbance value yields a numerator difference; subtracting the lower limit of the interval from the upper limit yields a denominator difference; dividing the numerator difference by the denominator difference gives the quotient, which is the relative offset of the starch response integrated absorbance within the closed interval. This relative offset is a value between zero and one, representing the proportion by which the starch response integrated absorbance shifts from the lower limit to the upper limit of the interval. The data processing system uses this relative offset as an indicator of the relative starch content and stores it for subsequent construction of component feature vectors.

[0089] In summary, by comparing the baseline integrated absorbance and the polyester response integrated absorbance, the two endpoints for constructing the closed interval are automatically determined. This ensures that the allocation of the lower and upper limits of the interval is entirely driven by measured data, avoiding subjective bias caused by manually specifying the reference interval. Since the starch response integrated absorbance always lies between or near the baseline and polyester response integrated absorbance, the smaller value is set as the lower limit and the larger value as the upper limit, ensuring that the subsequently calculated relative starch content always falls within the range of zero to one, achieving a normalized expression of starch content. This ranking operation does not rely on any external standard curve or prior knowledge, making it applicable to different types and ratios of biodegradable plastic products, thus improving the method's adaptability and versatility. Furthermore, the clarity of the ranking results ensures complete repeatability in determining the lower and upper limits of the interval; different operators processing the same set of data will obtain completely consistent results.

[0090] In summary, connecting the two endpoint values ​​obtained from the sorting into a continuous closed interval provides a standard reference range for starch response integrated absorbance, giving the calculation of the starch relative content index a clear geometric meaning. The closed interval includes all possible values ​​from the lower limit to the upper limit. The starch response integrated absorbance, as the metric, is placed within this interval for relative position evaluation, eliminating the dimensional influence caused by differences in overall absorbance levels between different samples. The length of the closed interval reflects the degree of difference in integrated absorbance between the reference component and the polyester component. This length, as the denominator, participates in subsequent relative offset calculations, acting as an automatic scaling mechanism to ensure the comparability of the starch relative content index among different samples. The operation of constructing the closed interval uses only the lower and upper limits of the interval, without introducing any additional parameters, ensuring the simplicity of the interval definition and mathematical rigor.

[0091] In summary, a normalized relative starch content index is obtained by calculating the ratio of the offset of the starch response integrated absorbance relative to the lower limit of the closed interval to the total length of the closed interval. This index accurately reflects the relative content level of starch components in biodegradable plastic products. The relative offset transforms the starch response integrated absorbance from its original dimensionality into a dimensionless proportional value, eliminating the multiplicative interference from sample thickness, optical path length, and instrument response, allowing direct comparison of results measured under different batches and conditions. When the starch response integrated absorbance equals the lower limit of the interval, the relative offset is zero, indicating extremely low starch content; when it equals the upper limit, the relative offset is one, indicating extremely high starch content; when it falls between these two values, it is output linearly, resulting in a monotonically positive correlation between the starch relative content index and the actual starch mass fraction. For cases where the starch response integrated absorbance falls outside the closed interval, a truncation assignment operation forces the index to be limited to the range of zero to one, avoiding meaningless values ​​outside the range caused by spectral anomalies or noise, thus ensuring the stability and interpretability of the index. The method for calculating the relative offset corresponds exactly to the formula for calculating the relative starch content index given in subsequent patent documents, ensuring the consistency and feasibility of the technical solution.

[0092] In this embodiment of the invention, the formula for calculating the relative starch content index is specifically used for: in, This is an indicator of the relative starch content. The integral absorbance of the starch response. The baseline integrated absorbance, The integral absorbance of the polyester response.

[0093] Specifically, the starch response integrated absorbance is obtained by summing the areas of the trapezoids enclosed by the absorbance curves and the wavelength axis within the starch response region. This result reflects the total intensity of the characteristic absorption of the starch component in the near-infrared spectrum. The reference integrated absorbance is obtained by summing the areas of the trapezoids enclosed by the absorbance curves and the wavelength axis within the reference absorption region. This result reflects the total absorption intensity of the reference component in the biodegradable plastic product, which does not change with the component ratio. The polyester response integrated absorbance is obtained by summing the areas of the trapezoids enclosed by the absorbance curves and the wavelength axis within the polyester response region. This result reflects the total intensity of the characteristic absorption of the polyester component in the near-infrared spectrum. These three integrated absorbance values ​​are calculated from the calibrated spectral data using the trapezoidal method and are stored in the memory of the data processing system as the basic input for calculating the relative starch content index.

[0094] Furthermore, the significance of the starch relative content index calculation formula lies in mapping the starch response integrated absorbance to a numerical range determined by the baseline integrated absorbance and the polyester response integrated absorbance, thus obtaining a normalized relative content value. This calculation process first compares the magnitudes of the baseline integrated absorbance and the polyester response integrated absorbance, using the smaller value as the lower limit of the interval and the larger value as the upper limit, thereby constructing a closed interval. Then, the offset of the starch response integrated absorbance relative to the lower limit of this closed interval is calculated, i.e., subtracting the lower limit from the starch response integrated absorbance. Next, the total length of the entire closed interval is calculated, i.e., subtracting the lower limit from the upper limit. Finally, the offset is divided by the total length of the interval, and the quotient is the relative position of the starch response integrated absorbance within the closed interval. This relative position value is between zero and one; it is zero when the starch response integrated absorbance equals the lower limit and one when it equals the upper limit. This index eliminates the influence of differences in overall absorbance intensity between different samples, allowing for cross-sectional comparisons of starch relative content between samples from different batches and of different thicknesses.

[0095] In general, when the starch response integrated absorbance value approaches the smaller of the baseline integrated absorbance and the polyester response integrated absorbance, the starch relative content index value approaches zero, indicating an extremely low relative starch content in the sample. When the starch response integrated absorbance value approaches the larger of the baseline integrated absorbance and the polyester response integrated absorbance, the starch relative content index value approaches one, indicating an extremely high relative starch content in the sample. When the starch response integrated absorbance is exactly equal to the average of the baseline integrated absorbance and the polyester response integrated absorbance, the starch relative content index value is 0.5, indicating an intermediate level of starch relative content. When the starch response integrated absorbance is less than the lower limit of the interval, the starch relative content index is forcibly assigned a value of zero; when the starch response integrated absorbance is greater than the upper limit of the interval, the starch relative content index is forcibly assigned a value of one. Therefore, this index monotonically increases with the increase of the starch response integrated absorbance, showing a linear positive correlation; that is, the higher the starch response integrated absorbance, the closer the starch relative content index value is to one, and vice versa.

[0096] In this embodiment of the invention, when determining the relative polyester content index of the target process based on the polyester response integrated absorbance and the reference integrated absorbance, and combining the relative starch content index and the relative polyester content index to obtain the component feature vector of the target process, it is specifically used for: The ratio of the polyester response integrated absorbance to the baseline integrated absorbance is used as the relative polyester content index of the target process. Using the relative starch content index as the first component and the relative polyester content index as the second component, a component feature vector for the target process is constructed.

[0097] Specifically, the system reads the calculated polyester response integrated absorbance and the reference integrated absorbance. The polyester response integrated absorbance is used as the dividend in the division operation, and the reference integrated absorbance is used as the divisor. The data processing system first checks if the reference integrated absorbance is zero. If it is, the relative polyester content index is directly assigned a preset maximum value, because the division operation cannot be performed correctly when the divisor is zero.

[0098] Specifically, the system reads the calculated relative starch content and relative polyester content values, which represent the relative content of starch and polyester components in the biodegradable plastic product being tested, respectively. The data processing system creates a two-position data structure in memory to store a two-dimensional vector. The first position stores the first component of the vector, and the second position stores the second component. The system then fills the first position of this data structure with the read relative starch content value, using this value as the first component of the component feature vector.

[0099] Furthermore, if the baseline integrated absorbance is not zero, the data processing system divides the value of the polyester response integrated absorbance by the value of the baseline integrated absorbance, and the quotient is the ratio of the polyester response integrated absorbance to the baseline integrated absorbance. The data processing system names this ratio the polyester relative content index and stores it in memory as a floating-point number. This polyester relative content index reflects the proportion of polyester component in the tested biodegradable plastic product relative to the reference component; a higher value indicates a higher relative content of polyester component.

[0100] Furthermore, the data processing system fills the second position of the data structure with the value of the relative polyester content index, using this value as the second component of the component feature vector. After filling, the data processing system names the entire data structure the component feature vector. This component feature vector is a vector containing two elements: the first element is the relative starch content index, and the second element is the relative polyester content index. The data processing system uses the component feature vector as the final output for subsequent identification or classification of biodegradable plastic product components.

[0101] In summary, this ratio eliminates the influence of sample thickness, optical path variation, and instrument response fluctuations on the absolute intensity of absorbance, making the relative polyester content measured at different batches and times comparable. Since the component corresponding to the reference absorption region has a constant content or absorbance in biodegradable plastic products that is unaffected by changes in polyester and starch content, using its integrated absorbance as the denominator effectively corrects for multiplicative interference during spectral acquisition. The ratio of the polyester response integrated absorbance to the reference integrated absorbance directly reflects the content ratio of the polyester component relative to the reference component, and this ratio is monotonically positively correlated with the actual mass fraction of polyester.

[0102] In summary, this component feature vector integrates the relative content information of starch and polyester, the two main biodegradable components, into a two-dimensional numerical space, providing a unified numerical representation for subsequent sample classification, similarity retrieval, or quality assessment. The first component, the relative starch content index, is restricted to a closed interval between zero and one, while the second component, the relative polyester content index, is a non-negative real number. The combination of these two components uniquely characterizes the compositional features of the biodegradable plastic product being tested. This feature vector exhibits translation invariance and scaling invariance, meaning that the same sample under different spectral intensities will map to similar positions in the feature space, thereby improving the stability and repeatability of the detection method. By calculating the Euclidean distance or cosine similarity between the component feature vectors of different samples, the compositional consistency between unknown samples and standard samples can be quickly determined, making it suitable for scenarios such as plastic recycling and sorting, and product quality sampling. The construction process of this feature vector is entirely based on the ratio and relative position of spectral integral absorbance, without relying on external training sets or standard curves, thus exhibiting good generalization ability and interpretability.

[0103] Compared with the prior art, the present invention has the following beneficial effects: 1. This detection method adaptively determines the reference absorption region, polyester response region, and starch response region through baseline correction and the peak position and half-width of characteristic absorption peak clusters. It eliminates the need for pre-setting fixed wavelength ranges or relying on a large number of standard samples to establish a calibration model, significantly improving its adaptability to biodegradable plastic products with different formulations and production processes. The method uses a trapezoidal integration method to accumulate the area enclosed by the absorbance curves and wavelength axis within each absorption region, obtaining the reference integrated absorbance, polyester response integrated absorbance, and starch response integrated absorbance. The integration process constructs trapezoids segment by segment based on measured discrete data points and accumulates the areas, accurately restoring the actual absorption intensity of each absorption region and eliminating integration errors caused by peak shape asymmetry or peak position shift. The relative offset of the starch response integrated absorbance within the closed interval formed by the reference integrated absorbance and the polyester response integrated absorbance is used as an indicator of starch relative content, achieving a normalized expression of starch content. This indicator is automatically limited to the range of zero to one, eliminating the influence of sample thickness and optical path differences, and ensuring the stability and repeatability of detection results for different batches of samples.

[0104] 2. This detection method uses the ratio of the polyester response integrated absorbance to the reference integrated absorbance as an indicator of the relative polyester content. This ratio operation effectively suppresses multiplicative interference during spectral acquisition, ensuring that the polyester content estimation is independent of external calibration curves. A component feature vector is constructed using the starch relative content index as the first component and the polyester relative content index as the second component. This feature vector comprehensively characterizes the relative composition of the two main biodegradable components in biodegradable plastic products in a two-dimensional numerical form, providing a unified and calculable numerical basis for subsequent sample classification, quality judgment, and recycling sorting. The entire detection process only requires the acquisition of near-infrared full-band spectral data, requires no sample pretreatment, consumes no chemical reagents, is fast, and does not damage the sample, making it suitable for rapid sampling inspections on production lines and real-time on-site detection scenarios.

[0105] like Figure 2 The diagram shown is a functional block diagram of a rapid detection system for the composition of biodegradable plastic products based on near-infrared spectroscopy, provided in an embodiment of the present invention.

[0106] The rapid detection system 100 for biodegradable plastic product components based on near-infrared spectroscopy described in this invention can be installed in an electronic device. Depending on the functions implemented, the rapid detection system 100 may include a raw spectral data module 101, a calibrated spectral data module 102, a partitioning module 103, a three-region integration module 104, a starch index module 105, and a component characteristic module 106. The module described in this invention can also be referred to as a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and perform a fixed function, and is stored in the memory of the electronic device.

[0107] In this embodiment, the functions of each module / unit are as follows: The raw spectral data module acquires the absorbance value sequence of the biodegradable plastic product to be tested in the near-infrared full band during the target process, thus obtaining the raw spectral data of the target process; The calibration spectral data module performs baseline correction on the original spectral data to obtain the calibration spectral data for the target process; The partitioning module determines the reference absorption region, polyester response region, and starch response region of the target process based on the peak position and half-peak width of the characteristic absorption peak clusters in the corrected spectral data. The three-zone integration module uses the area enclosed by the absorbance curves in the reference absorption zone, the polyester response zone, and the starch response zone and the wavelength axis as the reference integrated absorbance, polyester response integrated absorbance, and starch response integrated absorbance of the target process. The starch index module locates the relative position of the starch response integral absorbance within the numerical range formed by the baseline integral absorbance and the polyester response integral absorbance, thereby obtaining the relative starch content index of the target process. The component feature module determines the relative polyester content index of the target process based on the polyester response integrated absorbance and the benchmark integrated absorbance, and combines the starch relative content index with the polyester relative content index to obtain the component feature vector of the target process.

[0108] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0109] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0110] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0111] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0112] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0113] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A rapid detection method for the components of biodegradable plastic products based on near-infrared spectroscopy, characterized in that, The method includes: The absorbance sequence of the biodegradable plastic product to be tested in the near-infrared full band during the target process is obtained to obtain the original spectral data of the target process; Baseline correction is performed on the original spectral data to obtain the corrected spectral data for the target process; Based on the peak positions and half-widths of the characteristic absorption peak clusters in the corrected spectral data, the reference absorption region, polyester response region, and starch response region of the target process are determined respectively. The area enclosed by the absorbance curves in the reference absorption region, the polyester response region, and the starch response region and the wavelength axis is used as the reference integrated absorbance, polyester response integrated absorbance, and starch response integrated absorbance of the target process. Within the numerical range formed by the baseline integrated absorbance and the polyester response integrated absorbance, the relative position of the starch response integrated absorbance is located to obtain the relative starch content index of the target process. The relative polyester content index of the target process is determined based on the polyester response integrated absorbance and the baseline integrated absorbance, and the relative starch content index and the relative polyester content index are combined to obtain the component feature vector of the target process.

2. The rapid detection method for the components of biodegradable plastic products based on near-infrared spectroscopy as described in claim 1, characterized in that, The process of acquiring the absorbance sequence of the biodegradable plastic product under test in the near-infrared full-band during the target process yields the raw spectral data of the target process, including: Receive the discrete wavelength point sequence collected in the near-infrared full-band range of the biodegradable plastic product to be tested; The absorbance values ​​corresponding to the discrete wavelength points in the discrete wavelength point sequence are arranged in ascending order of wavelength to obtain the original spectral data of the target process.

3. The rapid detection method for the components of biodegradable plastic products based on near-infrared spectroscopy as described in claim 2, characterized in that, The baseline correction of the original spectral data to obtain the corrected spectral data for the target process includes: In the raw spectral data, identify wavelength pairs located outside the characteristic absorption peaks of the biodegradable plastic component in the biodegradable plastic product to be tested; The spectral background baseline of the target process is determined using the wavelength point pairs and their corresponding absorbance values ​​as endpoints. Based on the spectral background baseline, the absorbance value at the location with the same wavelength value as the wavelength point in the original spectral data is obtained to obtain the background absorbance value of the target process; The difference sequence between the original absorbance value at the wavelength point and the background absorbance value is used as the corrected spectral data for the target process.

4. The rapid detection method for the components of biodegradable plastic products based on near-infrared spectroscopy as described in claim 3, characterized in that, The step of determining the spectral background baseline of the target process using the wavelength point pairs and corresponding absorbance values ​​as endpoints includes: The wavelength value and corresponding absorbance value of the first wavelength point in the wavelength point pair are used as the starting wavelength and starting absorbance, and the wavelength value and corresponding absorbance value of the second wavelength point in the wavelength point pair are used as the ending wavelength and ending absorbance. The wavelength difference between the starting wavelength and the ending wavelength, and the absorbance difference between the starting absorbance and the ending absorbance are obtained. The wavelength offset between the wavelength value of the wavelength point in the original spectral data and the starting wavelength is obtained. The ratio of the wavelength offset to the wavelength difference is taken as the offset ratio, and the product of the offset ratio and the absorbance difference is taken as the absorbance offset share of the target process. The initial absorbance is added to the absorbance offset fraction to obtain the background absorbance value of the target process; By sequentially connecting all wavelength points and the points determined by their background absorbance values, the spectral background baseline of the target process is obtained.

5. The rapid detection method for the components of biodegradable plastic products based on near-infrared spectroscopy as described in claim 1, characterized in that, The determination of the reference absorption region, polyester response region, and starch response region of the target process based on the peak positions and half-widths of the characteristic absorption peak clusters in the corrected spectral data includes: The discrete data points of the corrected spectral data are arranged in order of wavelength from shortest to longest to obtain the discrete data point sequence of the target process; The data points in the discrete data point sequence with absorbance values ​​higher than the adjacent data points are taken as local maxima, and recorded as the reference peak position, polyester response peak position, and starch response peak position of the target process, respectively. Centered on the reference peak, the polyester response peak, and the starch response peak, respectively, extend towards the wavelength decreasing direction and the wavelength increasing direction until the local minimum value of the absorbance is located, thus obtaining the reference valley point pair, polyester valley point pair, and starch valley point pair of the target process; The wavelengths corresponding to the left and right valley points in the reference valley point pair, the polyester valley point pair, and the starch valley point pair are respectively used as the left boundary wavelength and the right boundary wavelength to obtain the reference absorption region, polyester response region, and starch response region of the target process.

6. The rapid detection method for the components of biodegradable plastic products based on near-infrared spectroscopy as described in claim 1, characterized in that, The step of using the area enclosed by the absorbance curves in the reference absorption region, the polyester response region, and the starch response region and the wavelength axis as the reference integrated absorbance, polyester response integrated absorbance, and starch response integrated absorbance of the target process includes: Using the wavelength and absorbance values ​​of the corrected spectral data in the target process as the abscissa and ordinate, a wavelength absorbance plane for the target process is established, and the discrete data points in the reference absorption region, the polyester response region, and the starch response region are marked as coordinate points in the wavelength absorbance plane according to the wavelength and absorbance values. Connecting adjacent coordinate points in ascending order of wavelength in the wavelength absorbance plane yields the absorbance curve of the target process; Starting from the initial data point on the absorbance curve, adjacent data points are extracted sequentially to obtain the adjacent data point pairs of the target process and the corresponding left and right points; Using the absorbance value of the left point as the height of the left side of the trapezoid, the absorbance value of the right point as the height of the right side of the trapezoid, and the wavelength interval between the left point and the right point as the width of the trapezoid, a trapezoid for the target process is constructed. The vertices of the trapezoid are: the left point on the absorbance curve, the projection point of the left point on the wavelength axis of the wavelength absorbance plane, the projection point of the right point on the wavelength axis of the wavelength absorbance plane, and the right point on the absorbance curve. The areas of the trapezoids, the baseline integrated absorbance, the polyester response integrated absorbance, and the starch response integrated absorbance of the target process are summed up respectively.

7. The rapid detection method for the components of biodegradable plastic products based on near-infrared spectroscopy as described in claim 1, characterized in that, The step of locating the relative position of the starch response integral absorbance within the numerical range formed by the baseline integrated absorbance and the polyester response integral absorbance to obtain the relative starch content index of the target process includes: The lower limit and upper limit of the interval for the target process are determined based on the ranking results of the baseline integrated absorbance and the polyester response integrated absorbance. Using the lower limit and the upper limit of the interval as endpoints, construct the closed interval of the target process; The relative offset of the starch response integral absorbance within the closed interval is used as the relative starch content index of the target process.

8. The rapid detection method for the components of biodegradable plastic products based on near-infrared spectroscopy as described in claim 7, characterized in that, The formula for calculating the relative starch content index includes: in, This is an indicator of the relative starch content. The integral absorbance of the starch response. The baseline integrated absorbance, The integral absorbance of the polyester response.

9. The rapid detection method for the components of biodegradable plastic products based on near-infrared spectroscopy as described in claim 1, characterized in that, The process involves determining the relative polyester content index of the target process based on the polyester response integrated absorbance and the baseline integrated absorbance, and combining the relative starch content index with the relative polyester content index to obtain the component feature vector of the target process, including: The ratio of the polyester response integrated absorbance to the baseline integrated absorbance is used as the relative polyester content index of the target process. Using the relative starch content index as the first component and the relative polyester content index as the second component, a component feature vector for the target process is constructed.

10. A rapid detection system for the components of biodegradable plastic products based on near-infrared spectroscopy, used to implement the rapid detection method for the components of biodegradable plastic products based on near-infrared spectroscopy as described in any one of claims 1-9, characterized in that, The system includes: The raw spectral data module acquires the absorbance value sequence of the biodegradable plastic product to be tested in the near-infrared full band during the target process, thus obtaining the raw spectral data of the target process; The calibration spectral data module performs baseline correction on the original spectral data to obtain the calibration spectral data for the target process; The partitioning module determines the reference absorption region, polyester response region, and starch response region of the target process based on the peak position and half-peak width of the characteristic absorption peak clusters in the corrected spectral data. The three-zone integration module uses the area enclosed by the absorbance curves in the reference absorption zone, the polyester response zone, and the starch response zone and the wavelength axis as the reference integrated absorbance, polyester response integrated absorbance, and starch response integrated absorbance of the target process. The starch index module locates the relative position of the starch response integral absorbance within the numerical range formed by the baseline integral absorbance and the polyester response integral absorbance, thereby obtaining the relative starch content index of the target process. The component feature module determines the relative polyester content index of the target process based on the polyester response integrated absorbance and the benchmark integrated absorbance, and combines the starch relative content index with the polyester relative content index to obtain the component feature vector of the target process.

Images