Mixed plastics sorting system and method based on infrared spectroscopy

Abstract

The present invention discloses a mixed plastic sorting system and method based on infrared spectroscopy, relates to the technical field of plastic sorting, and solves the technical problems of low sorting accuracy, poor sorting effect of complex mixed plastics, complex process flow and high cost of existing sorting equipment; the present invention obtains standard infrared spectrum curves of various types of plastics; analyzes standard spectrum information of various types of plastics based on the standard infrared spectrum curves; identifies the absorption characteristic peaks of the plastics to be classified and obtains the spectrum information of the plastics to be classified; compares the spectrum information of the plastics to be classified with the standard spectrum information to determine the type of the plastics to be classified; collects images of the plastics to be classified and analyzes the quality of the plastics to be classified; and classifies the plastics to be classified according to the type and quality of the plastics to be classified; the present invention can efficiently and accurately realize the sorting of mixed plastics, and provides an advanced technical means for the plastic recycling industry.

Description

Technical Field

[0001] The present invention belongs to the field of plastic sorting, and in particular relates to a mixed plastic sorting system and method based on infrared spectroscopy. Background Art

[0002] In the plastic recycling industry, the sorting of mixed plastics has always been a key problem. Existing color sorters, sorting machines and other equipment mainly sort materials based on appearance characteristics such as color and shape. However, these methods have obvious limitations. On the one hand, for plastics with similar colors and appearances but different materials, existing sorting equipment is difficult to accurately distinguish, resulting in low sorting accuracy and low purity of recycled plastics, affecting the subsequent processing and utilization value. On the other hand, existing sorting equipment is unable to effectively identify the true material of some plastics whose appearance has changed after treatments such as dyeing and aging, resulting in sorting errors and reduced sorting efficiency. In addition, traditional sorting machines often require the cooperation of multiple equipment when processing complex mixed plastics, which results in complex process flows and high costs.

[0003] Therefore, the present invention provides a mixed plastic sorting system and method based on infrared spectroscopy. Summary of the Invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a mixed plastic sorting system and method based on infrared spectroscopy, which is used to solve the technical problems of low sorting accuracy, poor sorting effect of complex mixed plastics, complex process flow and high cost of existing sorting equipment.

[0005] To achieve the above objectives, the first aspect of the present invention provides a mixed plastics sorting method based on infrared spectroscopy, comprising:

[0006] Obtaining standard infrared spectrum curves for various types of plastics; analyzing standard spectrum information for various types of plastics based on the standard infrared spectrum curves; wherein the standard spectrum information includes standard peak value, standard area, and standard position information of characteristic absorption peaks;

[0007] Identify the absorption characteristic peak of the plastic to be classified and obtain spectral information of the plastic to be classified; compare the spectral information of the plastic to be classified with the standard spectral information to determine the type of the plastic to be classified; wherein the spectral information includes the peak value, area and position information of the characteristic absorption peak; and

[0008] Collect images of plastics to be classified and analyze the quality of the plastics to be classified;

[0009] The plastics to be sorted are sorted according to their type and quality.

[0010] Preferably, obtaining standard infrared spectrum curves of various types of plastics includes:

[0011] Extract a number of plastic samples of various types and obtain infrared spectrum curves of the plastic samples of various types;

[0012] The infrared spectrum curve is divided into several segments according to the variation trend, and several coordinate points are extracted from each segment of the curve; wherein the coordinate points extracted from the same type of infrared spectrum curve have the same horizontal coordinates;

[0013] Extract the ordinate value corresponding to the same abscissa from the same type of infrared spectrum curve, and mark the median of the ordinate value as the standard ordinate value; the abscissa and the standard ordinate value constitute the standard coordinate point; wherein the same type of infrared spectrum curve is the infrared spectrum curve of the same type of plastic;

[0014] Several standard coordinate points are fitted to obtain the standard infrared spectrum curves of various types of plastics.

[0015] For infrared spectral curves of the same type, the present invention extracts coordinate points with the same horizontal coordinate and calculates the median of the vertical coordinate values ​​of these points as the standard vertical value to form standard coordinate points. This process unifies the original infrared spectral data of different samples of the same type under the same standard, eliminating the measurement value differences between different samples due to factors such as experimental conditions and instrument response, making the data of different samples comparable, and providing a high-quality, standardized data foundation for subsequent standard infrared spectral curve fitting. Standard infrared spectral curves of various types of plastics are obtained by fitting based on the standard coordinate points. Since the standard coordinate points have undergone data standardization processing, the interference of outliers and random errors is eliminated, and the common characteristics and variation patterns of the infrared spectra of the same type of plastics can be more accurately reflected, thereby making the fitted standard infrared spectral curve closer to the actual situation and improving the accuracy and reliability of the curve fitting.

[0016] Preferably, the analysis of standard spectral information of various types of plastics includes:

[0017] Based on the standard infrared spectrum curve, the peak value, area and position information of the characteristic absorption peak are obtained; the position information of the characteristic absorption peak includes the peak position, starting position, ending position and half-height width; the starting position is the position where the characteristic absorption peak starts, and the ending position is the position where the characteristic absorption peak ends;

[0018] The area of ​​the characteristic absorption peak is calculated as follows:

[0019] A number of coordinate points are extracted from the characteristic absorption peak, and the curve equation of the characteristic absorption peak is constructed based on the curve equation fitting; based on the starting position and the ending position of the absorption characteristic peak, the curve equation is integrated to obtain the standard area of ​​the characteristic absorption peak.

[0020] The present invention determines the highest point position, starting position, ending position and half-height width of the characteristic absorption peak, and the characteristic absorption peak area obtained by curve equation fitting and integral calculation takes into account the shape and intensity of the absorption peak, and can more accurately reflect the total amount of chemical bonds or functional groups represented by the characteristic absorption peak. Compared with using only the peak value, the area calculation can better reflect the overall information of the absorption peak and reduce the error caused by the irregular shape of the absorption peak. A number of coordinate points are extracted from the characteristic absorption peak, and the curve equation of the characteristic absorption peak is constructed based on the curve equation fitting, which can more accurately describe the shape and change trend of the absorption peak. Compared with simple linear fitting or polynomial fitting, the use of a more suitable curve equation can better fit the actual infrared spectrum absorption peak and improve the accuracy of the calculation.

[0021] Preferably, the identifying characteristic absorption peaks of the plastics to be classified comprises:

[0022] Based on the standard spectrum information, the peak values ​​of the characteristic absorption peaks of each type of plastic are extracted to obtain a peak sequence; wherein the peak values ​​in the peak sequence are different;

[0023] Extract the peak to be identified from the infrared spectrum curve of the plastic to be classified, calculate the difference between the peak value of the peak to be identified and each peak value in the peak sequence, and determine whether the difference is less than the difference threshold; if yes, mark the peak to be identified as a characteristic absorption peak; if not, do not mark it; wherein the peak to be identified is the maximum point in the infrared spectrum curve of the plastic to be classified.

[0024] The standard spectral information of the present invention has been verified and scientifically analyzed on a large number of samples, and the peak value of its characteristic absorption peak is representative and stable. By extracting the peak value in the standard spectral information to form a peak sequence, a reliable reference standard is provided for the identification of the characteristic absorption peak of the plastic to be classified, and slight differences between the peak value of the peak to be identified and the peak values ​​in the peak sequence are allowed, effectively avoiding misjudgment caused by measurement errors, instrument fluctuations and other factors.

[0025] Preferably, the step of determining the type of plastic to be classified includes:

[0026] Count the number of peaks that are identical between the peak of the plastic to be classified and the standard peak of each type of plastic. The plastic type with the number of identical peaks exceeding the quantity threshold is the candidate type of the plastic to be classified;

[0027] Calculate the absolute value of the area difference between the area of ​​the characteristic absorption peak of the plastic to be classified and the standard area of ​​each identical absorption characteristic peak in the candidate type, and calculate the ratio of the absolute value of the area difference to the standard area of ​​the identical absorption characteristic peak to obtain the area difference rate;

[0028] Calculate the absolute value of the difference between the position information of the absorption characteristic peak of the plastic to be classified and the standard position information of each same absorption characteristic peak in the candidate type, and sum the ratio with the corresponding standard position information of the same absorption characteristic peak to obtain the position difference rate;

[0029] The weighted sum of the difference rate and the position difference rate is calculated to obtain a similarity index; and the type of the plastic to be classified is determined based on the similarity index.

[0030] This method comprehensively considers three characteristic dimensions—peak value, area, and position—to assess the similarity of the plastic being classified against various standard plastic types. This multi-dimensional matching approach more comprehensively and accurately reflects the characteristics of the plastic, avoiding the potential bias associated with single-feature matching, thereby improving classification accuracy and reliability. For example, relying solely on peak value matching can lead to misclassifications due to similar peak values ​​but significant differences in area or position. This method, by integrating multiple features, effectively reduces the likelihood of such misclassifications.

[0031] Preferably, the difference between the peak value of the same absorption characteristic peak value, which is the peak to be identified and the standard peak value, is less than a difference threshold.

[0032] Preferably, the weight coefficients of the area and position information of the characteristic absorption peak are obtained by the following methods, including:

[0033] Extract a number of plastic samples of various types and calculate the mean and variance of the spectral information of each type of plastic sample;

[0034] By formula The importance degree FDR of the plastic spectral information is calculated; where i is the plastic type, k is the area or position information of the absorption characteristic peak, usually expressed as a natural number; uik is the mean value of the spectral information k of the i type plastic, sik is the standard spectral information k of the i type plastic, σi k is the variance of the spectral information k of type i plastic, Σ is the sum of k;

[0035] Based on the importance, the weight coefficients of the area and position information of the characteristic absorption peak are determined.

[0036] The present invention calculates the difference between the mean of spectral information and the standard value and divides it by the sum of squared variances to normalize the characteristic differences into comparable values. This effectively integrates the absorption peak area and position information, and quantitatively assesses the importance of the characteristics through weight assignment. Based on the calculated results, weight coefficients are assigned, which can be dynamically adjusted according to the sensitivity of different plastic types to area and position information. For example, in the classification of polyvinyl chloride (PVC), if position information has a greater impact on classification, its weight coefficient can be increased; if area characteristics are more critical, the weight is increased accordingly. This dynamic adjustment mechanism enables the classification model to adapt to the characteristics of different plastic types, improving classification accuracy and flexibility.

[0037] Preferably, the analyzing the quality of the plastics to be classified includes:

[0038] The image of the plastic to be classified is input into the quality assessment model, and the quality assessment model outputs the quality label of the plastic to be classified; wherein, the quality label includes a good label and a defective label, and the quality assessment model is constructed by an artificial intelligence model.

[0039] Preferably, the quality assessment model is constructed by an artificial intelligence model, including:

[0040] Obtaining a number of plastic images of various types and corresponding quality labels, and marking defects in the plastic images to obtain marked images;

[0041] Integrate the labeled images and quality labels into several groups of training data and test data; use the training data to train the artificial intelligence model; use the test data to test the trained artificial intelligence model, and adjust the artificial intelligence model according to the test results; finally, obtain a quality assessment model with the labeled image as input and the quality label as output; wherein the artificial intelligence model is a BP neural network model or an RBF neural network model.

[0042] Preferably, the second aspect of the present invention provides a mixed plastic sorting system based on infrared spectroscopy, comprising an analysis module and a classification module;

[0043] Analysis module: used to obtain standard infrared spectrum curves of various types of plastics; based on the standard infrared spectrum curves, analyze the standard spectrum information of various types of plastics; wherein the standard spectrum information includes the standard peak value, standard area and standard position information of the characteristic absorption peak;

[0044] Identify the absorption characteristic peaks of the plastic to be classified and obtain spectral information of the plastic to be classified; compare the spectral information of the plastic to be classified with standard spectral information to determine the type of the plastic to be classified; and,

[0045] Collect images of plastics to be classified and analyze the quality of the plastics to be classified;

[0046] Classification module: used to classify the plastics to be classified according to their type and quality.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] The present invention obtains standard infrared spectral curves for various types of plastics and conducts in-depth analysis of their standard spectral information, including the standard peak value, standard area, and standard position of characteristic absorption peaks. This provides an accurate reference for subsequent classification. After identifying the characteristic absorption peaks of the plastic to be classified and obtaining their spectral information, it is carefully compared with the standard spectral information to accurately determine the type of plastic to be classified. Compared with existing sorting machines based on color and shape, the present invention, based on infrared spectroscopy technology, can accurately distinguish plastics of different materials, even those with similar colors and appearances, significantly improving sorting accuracy. In addition to spectral information, the present invention also captures images of the plastics to be classified and analyzes their quality. This multi-dimensional assessment method not only considers the spectral information of the plastics but also takes into account plastic quality analysis, resulting in more comprehensive and reliable classification results, facilitating more targeted processing and utilization of the plastics. The present invention has excellent adaptability to different types of plastics. Whether common polyethylene, polypropylene, or other specialized plastics, accurate classification can be achieved through the establishment and comparison of standard spectral curves, significantly improving sorting efficiency and accuracy, reducing labor costs, and possessing significant economic benefits and application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0050] Figure 1 Schematic diagram of the process of the present invention;

[0051] Figure 2 This is a schematic flow chart of a method for obtaining standard infrared spectrum curves for various plastic types according to the present invention;

[0052] Figure 3 Schematic diagram of the flow chart of the method for distinguishing the types of plastics to be classified according to the present invention. DETAILED DESCRIPTION

[0053] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the embodiments described are only a part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0054] See also Figure 1 The first embodiment of the present invention provides a mixed plastic sorting method based on infrared spectroscopy, comprising:

[0055] Obtain standard infrared spectrum curves for various types of plastics;

[0056] See also Figure 2 Specifically, a number of plastic samples of various types are extracted to obtain infrared spectrum curves of the plastic samples of various types;

[0057] The infrared spectrum curve is divided into several segments according to the variation trend, and several coordinate points are extracted from each segment of the curve; wherein the coordinate points extracted from the same type of infrared spectrum curve have the same horizontal coordinates;

[0058] Extract the ordinate value corresponding to the same abscissa from the same type of infrared spectrum curve, and mark the median of the ordinate value as the standard ordinate value; the abscissa and the standard ordinate value constitute the standard coordinate point; wherein the same type of infrared spectrum curve is the infrared spectrum curve of the same type of plastic;

[0059] Several standard coordinate points are fitted to obtain the standard infrared spectrum curves of various types of plastics.

[0060] For example, suppose a standard infrared spectrum curve of PE plastic is obtained. Infrared spectrum curves of several PE plastics are collected, such as supermarket shopping bags, mineral water bottles, plastic wrap, PE film, and PE packaging materials. Each sample is scanned using a Fourier transform infrared spectrometer to obtain five infrared spectrum curves of PE plastic. The horizontal axis is wave number cm-1, and the vertical axis is absorbance.

[0061] The infrared spectrum curve is divided into several segments according to the change trend. The change trend is traversed from (0, 0). If the infrared spectrum curve of a plastic sample (0, 0)-(x1, y1) is a smooth trend and (x1, y1)-(x2, y2) is an upward trend, then (0, 0)-(x1, y1) is a segment and (x1, y1)-(x2, y2) is a segment. After traversing all coordinate points in sequence, the infrared spectrum curve is divided into several segments.

[0062] Extract a number of coordinate points on each curve segment. For example, if the coordinate points of the (0, 0)-(x1, y1) segment on the infrared spectrum curve of a PE plastic sample are (x10, y10), (x11, y11), ..., (xn, yn), a total of n points, where n is a positive integer; extract coordinate points with horizontal coordinates of x10, x11, ..., xn for the rest of the same type of plastic samples; the vertical coordinate values ​​corresponding to the x10 point of several infrared spectrum curves are several, extract the median of the several vertical coordinate values, and mark it as the standard vertical value; (x10, standard vertical value) is the standard coordinate point; the corresponding standard coordinate points of the rest x11, ..., xn are obtained in this way, and a number of standard coordinate points of this segment of the curve are obtained;

[0063] In this way, several standard coordinate points of the remaining curve segments are obtained, and several standard coordinate points of several curve segments are fitted to obtain the standard infrared spectrum curve of PE type plastic.

[0064] Based on the standard infrared spectrum curve, the standard spectrum information of each type of plastic is analyzed; wherein the standard spectrum information includes the standard peak value, standard area and standard position information of the characteristic absorption peak;

[0065] Specifically, based on the standard infrared spectrum curve, the standard peak value and standard position information of the characteristic absorption peak are obtained;

[0066] The standard position information of the characteristic absorption peak includes the peak position, starting position, ending position and half-height width; the starting position is the position where the characteristic absorption peak starts, and the ending position is the position where the characteristic absorption peak ends;

[0067] The area of ​​the characteristic absorption peak is calculated as follows:

[0068] A number of coordinate points are extracted from the characteristic absorption peak, and the curve equation of the characteristic absorption peak is constructed based on multinomial equation fitting; based on the starting position and ending position of the absorption characteristic peak, the curve equation is integrated to obtain the standard area of ​​the characteristic absorption peak.

[0069] For example: Taking a certain PE plastic as an example, extract the peak point of the standard infrared curve at 2920cm-1 and determine the starting / ending position. Starting position: scan from 2920cm-1 to the left and find the wave number (assuming it is 2950cm-1) where the absorbance drops to the baseline (such as 0.05); ending position: scan from 2920cm-1 to the right and find the wave number (assuming it is 2850cm-1) where the absorbance drops to the baseline.

[0070] Calculate the half-height width: half-height = peak absorbance × 0.5 = 0.85 × 0.5 = 0.425; find the wave number corresponding to the absorbance = 0.425 on both sides of the peak (such as 2940cm-1 and 2900cm-1), half-height width = 2940–2900 = 40cm-1; in the range of 2850–2950cm-1, evenly select 10 wave number points and their absorbance (from the standard infrared spectrum curve): According to the polynomial equation fitting, the curve equation of the characteristic absorption peak of the PE plastic at 2920cm-1 is obtained, marked as A(x); through The area of ​​the characteristic absorption peak at 2920 cm-1 is calculated; wherein B is the initial position of the characteristic absorption peak, and C is the end position of the characteristic absorption peak.

[0071] Identify the absorption characteristic peaks of the plastics to be classified and obtain the spectral information of the plastics to be classified;

[0072] Based on the standard spectrum information, the peak values ​​of the characteristic absorption peaks of each type of plastic are extracted to obtain a peak sequence; wherein the peak values ​​in the peak sequence are different;

[0073] Extract the peak to be identified from the infrared spectrum curve of the plastic to be classified, calculate the difference between the peak value of the peak to be identified and each peak value in the peak sequence, and determine whether the difference is less than the difference threshold; if so, mark the peak to be identified as a characteristic absorption peak; wherein the peak to be identified is the maximum point in the infrared spectrum curve of the plastic to be classified.

[0074] For example, based on standard infrared spectrum information, it is assumed that the characteristic absorption peaks of PE (polyethylene), PP, and PVC (polyvinyl chloride) plastics are extracted to form a peak sequence; among them, only one identical peak is retained;

[0075] Taking PVC (polyvinyl chloride) as an example, Table 1 shows the matching relationship between the plastic to be classified and the standard peak value of the PVC type;

[0076] Table 1 shows the matching relationship between the plastics to be classified and the standard peak values ​​of PVC types

[0077]

[0078]

[0079] Compare the spectral information of the plastic to be classified with the standard spectral information to determine the type of the plastic to be classified;

[0080] See also Figure 3 Specifically, the number of identical peaks between the peak of the plastic to be classified and the standard peak of each type of plastic is counted, and the plastic type whose number of identical peaks exceeds the quantity threshold is the candidate type of the plastic to be classified;

[0081] Calculate the absolute value of the area difference between the area of ​​the characteristic absorption peak of the plastic to be classified and the standard area of ​​each identical absorption characteristic peak in the candidate type, and calculate the ratio of the absolute value of the area difference to the standard area of ​​the identical absorption characteristic peak to obtain the area difference rate;

[0082] Calculate the absolute value of the difference between the position information of the absorption characteristic peak of the plastic to be classified and the standard position information of each identical absorption characteristic peak in the candidate type, and sum the ratio with the corresponding standard position information of the identical absorption characteristic peak to obtain the position difference rate; wherein, the identical absorption characteristic peak refers to the difference between the peak of the peak to be identified and the standard peak being less than the difference threshold;

[0083] Calculate the weighted sum of the difference rate and the position difference rate to obtain a similarity index; determine whether the similarity index is greater than a preset similarity threshold; if yes, the absorption characteristic peak of the plastic to be classified matches the absorption characteristic peak of the candidate type; if no, the absorption characteristic peak of the plastic to be classified does not match the absorption characteristic peak of the candidate type;

[0084] Count the number of matching absorption characteristic peaks between the plastic to be classified and each candidate type; determine whether the number of matching absorption characteristic peaks is greater than the matching number threshold; if yes, the candidate type is the type of the plastic to be classified; if not, the candidate type is not the type of the plastic to be classified.

[0085] The weight coefficients of the area and position information of the characteristic absorption peak are obtained by the following methods:

[0086] Extract a number of plastic samples of various types and calculate the mean and variance of the spectral information of each type of plastic sample;

[0087] By formula The importance degree FDR of the plastic spectral information is calculated; where i is the plastic type, k is the area or position information of the absorption characteristic peak, uik is the mean value of the spectral information k of the i type plastic, sik is the standard spectral information k of the i type plastic, σi k is the variance of the spectral information k of type i plastic;

[0088] Based on the importance, the weight coefficients of the area and position information of the characteristic absorption peak are determined.

[0089] For example: Assume that the peak value of a plastic to be classified is the same as the standard peak value of PE, PP, and PVC type plastics (cm -1 ) The number is shown in Table 2: If the difference threshold is: ±10cm -1 , quantity threshold: ≥3 matching peaks;

[0090] Table 2 Number of peaks that are identical between the peak of a plastic to be classified and the standard peak of PE, PP, and PVC plastics

[0091]

[0092] Result: PE and PVC are candidate types of plastics to be classified.

[0093] Calculate the area difference ratio between each matching characteristic peak: for example, the area difference ratio of the matching peaks between the plastic to be classified and PE = |area of ​​the peak to be identified at position 2918 - area of ​​the absorption characteristic peak at position 2920 of PE | / area of ​​the absorption characteristic peak at position 2920 of PE; the rest are calculated in the same way;

[0094] It should be noted that the matching peaks between the plastic to be classified and PE are absorption characteristic peaks with the same peak value.

[0095] Calculate the position difference rate: the position information includes the peak position and half-height width. Calculate the ratio of the sum of the absolute values ​​of the differences to the standard value: for example, the position difference rate of the matching peaks between the plastic to be classified and PE = |peak position of the peak to be identified at 2918 - peak position of the absorption characteristic peak at 2920 of PE | / peak position of the absorption characteristic peak at 2920 of PE + |half-height width of the peak to be identified at 2918 - half-height width of the absorption characteristic peak at 2920 of PE | / half-height width of the absorption characteristic peak at 2920 of PE + |initial position of the peak to be identified at 2918 - initial position of the absorption characteristic peak at 2920 of PE | / initial position of the absorption characteristic peak at 2920 of PE + |termination position of the peak to be identified at 2918 - termination position of the absorption characteristic peak at 2920 of PE | / termination position of the absorption characteristic peak at 2920 of PE;

[0096] Calculate the weight coefficients of the area and position information of the absorption characteristic peak; if the FDR of the area is greater than the FDR of the position information, the weight coefficient of the position is greater than the weight coefficient of the position information; otherwise, the weight coefficient of the position is less than or equal to the weight coefficient of the position information;

[0097] Similarity index = w1×area difference rate + w2×position difference rate; where w1 and w2 are the weight coefficients of area difference rate and position difference rate, respectively.

[0098] Assuming that the similarity index between the plastic to be classified and PE is less than the similarity index threshold, the type of the plastic to be classified is PE material.

[0099] Collect images of plastics to be classified and analyze the quality of the plastics to be classified;

[0100] The image of the plastic to be classified is input into the quality assessment model, and the quality assessment model outputs the quality label of the plastic to be classified; the quality assessment model is constructed by an artificial intelligence model, and the quality labels include good labels and defective labels, which are usually represented by "0" and "1" respectively.

[0101] Among them, the quality assessment model is constructed by an artificial intelligence model, including:

[0102] Obtaining a number of plastic images of various types and corresponding quality labels, and marking defects in the plastic images to obtain marked images;

[0103] Integrate the labeled images and quality labels into several groups of training data and test data; use the training data to train the artificial intelligence model; use the test data to test the trained artificial intelligence model, and adjust the artificial intelligence model according to the test results; finally, obtain a quality assessment model with the labeled image as input and the quality label as output; wherein the artificial intelligence model is a BP neural network model or an RBF neural network model.

[0104] Sorting out plastics according to their type and quality;

[0105] Specifically, plastics of the same type are classified, and within the same type, plastics are further divided into plastics of good quality and plastics with defects.

[0106] A second aspect of the present invention provides a mixed plastic sorting system based on infrared spectroscopy, comprising an analysis module and a classification module;

[0107] The analysis module obtains standard infrared spectrum curves of various types of plastics; based on the standard infrared spectrum curves, analyzes standard spectrum information of various types of plastics; wherein the standard spectrum information includes standard peak value, standard area and standard position information of characteristic absorption peaks;

[0108] Identify the absorption characteristic peaks of the plastic to be classified and obtain spectral information of the plastic to be classified; compare the spectral information of the plastic to be classified with standard spectral information to determine the type of the plastic to be classified; and,

[0109] Collect images of plastics to be classified and analyze the quality of the plastics to be classified;

[0110] The classification module classifies the plastics to be classified according to their type and quality.

[0111] Some of the data in the above formula are calculated by removing the dimensions and taking their numerical values. The formula is a formula that is closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0112] The above embodiments are only used to illustrate the technical method of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A mixed plastic separation method based on infrared spectroscopy, characterized in that: include: Obtaining standard infrared spectrum curves for various types of plastics; analyzing standard spectrum information for various types of plastics based on the standard infrared spectrum curves; wherein the standard spectrum information includes standard peak value, standard area, and standard position information of characteristic absorption peaks; Identify the absorption characteristic peaks of the plastics to be classified and obtain the spectral information of the plastics to be classified; The method of identifying the absorption characteristic peaks of the plastics to be classified comprises: Based on the standard spectrum information, the peak values ​​of the characteristic absorption peaks of each type of plastic are extracted to obtain a peak sequence; wherein the peak values ​​in the peak sequence are different; Extracting a peak to be identified from the infrared spectrum curve of the plastic to be classified, calculating the difference between the peak value of the peak to be identified and each peak value in the peak sequence, and determining whether the difference is less than a difference threshold; if so, marking the peak to be identified as a characteristic absorption peak; if not, not marking it; wherein the peak to be identified is a maximum point in the infrared spectrum curve of the plastic to be classified; Compare the spectral information of the plastic to be classified with the standard spectral information to determine the type of the plastic to be classified; Count the number of peaks that are identical between the peak of the plastic to be classified and the standard peak of each type of plastic. The plastic type with the number of identical peaks exceeding the quantity threshold is the candidate type of the plastic to be classified; Calculate the absolute value of the area difference between the area of ​​the characteristic absorption peak of the plastic to be classified and the standard area of ​​each identical absorption characteristic peak in the candidate type, and calculate the ratio of the absolute value of the area difference to the standard area of ​​the identical absorption characteristic peak to obtain the area difference rate; Calculate the absolute value of the difference between the position information of the absorption characteristic peak of the plastic to be classified and the standard position information of each same absorption characteristic peak in the candidate type, and sum the ratio with the corresponding standard position information of the same absorption characteristic peak to obtain the position difference rate; Calculating a weighted sum of the difference rate and the position difference rate to obtain a similarity index; determining the type of the plastic to be classified based on the similarity index; and, collecting images of the plastics to be classified and analyzing the quality of the plastics to be classified; The plastics to be sorted are sorted according to their type and quality.

2. The mixed plastic separation method based on infrared spectroscopy according to claim 1, characterized in that: The method of obtaining the standard infrared spectrum curves of various types of plastics includes: Extract a number of plastic samples of various types and obtain infrared spectrum curves of the plastic samples of various types; The infrared spectrum curve is divided into several segments according to the change trend, and several coordinate points are extracted from each segment of the curve; wherein the coordinate points extracted from the same type of infrared spectrum curve have the same horizontal coordinates; Extract the ordinate value corresponding to the same abscissa from the same type of infrared spectrum curve, and mark the median of the ordinate value as the standard ordinate value; the abscissa and the standard ordinate value constitute the standard coordinate point; wherein the same type of infrared spectrum curve is the infrared spectrum curve of the same type of plastic; Several standard coordinate points are fitted to obtain the standard infrared spectrum curves of various types of plastics.

3. The mixed plastic separation method based on infrared spectroscopy according to claim 1, characterized in that: The standard spectral information for analyzing various types of plastics includes: Based on the standard infrared spectrum curve, the peak value, area and position information of the characteristic absorption peak are obtained; the position information of the characteristic absorption peak includes the peak position, starting position, ending position and half-height width; the starting position is the position where the characteristic absorption peak starts, and the ending position is the position where the characteristic absorption peak ends; The area of ​​the characteristic absorption peak is calculated as follows: A number of coordinate points are extracted from the characteristic absorption peak, and the curve equation of the characteristic absorption peak is constructed based on the curve equation fitting; based on the starting position and the ending position of the absorption characteristic peak, the curve equation is integrated to obtain the standard area of ​​the characteristic absorption peak.

4. The mixed plastic separation method based on infrared spectroscopy according to claim 1, characterized in that: The same absorption characteristic peak value is a peak to be identified, and the difference between the peak value and the standard peak value is less than a difference threshold.

5. The mixed plastic separation method based on infrared spectroscopy according to claim 1, characterized in that: The weight coefficients of the area and position information of the characteristic absorption peak are obtained by: Extract a number of plastic samples of various types and calculate the mean and variance of the spectral information of each type of plastic sample; By the formula FDR=Σ k (uik-sik) / Σ k The importance degree FDR of the plastic spectral information is calculated; where i is the plastic type, k is the area or position information of the absorption characteristic peak, uik is the mean value of the spectral information k of the i type plastic, sik is the standard spectral information k of the i type plastic, is the variance of the spectral information k of type i plastic; Based on the importance, the weight coefficients of the area and position information of the characteristic absorption peak are determined.

6. The mixed plastic separation method based on infrared spectroscopy according to claim 1, characterized in that: The analysis of the quality of the plastics to be sorted includes: The image of the plastic to be classified is input into the quality assessment model, and the quality assessment model outputs the quality label of the plastic to be classified; wherein, the quality label includes a good label and a defective label, and the quality assessment model is constructed by an artificial intelligence model.

7. The mixed plastic separation method based on infrared spectroscopy according to claim 6, characterized in that: The quality assessment model is constructed by an artificial intelligence model, including: Obtaining a number of plastic images of various types and corresponding quality labels, and marking defects in the plastic images to obtain marked images; Integrate the labeled images and quality labels into several groups of training data and test data; use the training data to train the artificial intelligence model; use the test data to test the trained artificial intelligence model, and adjust the artificial intelligence model according to the test results; finally, obtain a quality assessment model with the labeled image as input and the quality label as output; wherein the artificial intelligence model is a BP neural network model or an RBF neural network model.

8. A mixed plastics sorting system based on infrared spectroscopy, operating based on the mixed plastics sorting method based on infrared spectroscopy according to any one of claims 1 to 7, characterized in that: Includes analysis module and classification module; Analysis module: used to obtain standard infrared spectrum curves of various types of plastics; based on the standard infrared spectrum curves, analyze the standard spectrum information of various types of plastics; wherein the standard spectrum information includes the standard peak value, standard area and standard position information of the characteristic absorption peak; Identify the absorption characteristic peaks of the plastic to be classified and obtain spectral information of the plastic to be classified; compare the spectral information of the plastic to be classified with standard spectral information to determine the type of the plastic to be classified; and, Collect images of plastics to be classified and analyze the quality of the plastics to be classified; Classification module: used to classify the plastics to be classified according to their type and quality.

Images