Light splitting type plastic sorting system and method based on near infrared spectrum analysis

Abstract

The invention belongs to the technical field of spectrum detection, and particularly relates to a light splitting type plastic sorting system and method based on near infrared spectrum analysis. In the sorting system, a light source, a light condensing unit, a light splitting unit, a spectrum collecting unit and a discharging device are all arranged above a conveying belt, the light source irradiates an identification area of the conveying belt below, the light condensing unit is located on a path of light reflected by the identification area, and the light splitting unit is located above the light condensing unit; the light splitting unit splits the reflected light into two beams of signal light emitted in the vertical direction, each light outlet of the light splitting unit is provided with an acquisition unit for acquiring light signals, and the two acquisition units transmit the acquired light signals to an identification unit for identification and analysis; the sorting unit is arranged at the tail end of the conveying belt, according to different plastic type combinations, collected characteristic wavelength points can be flexibly changed, and according to spectral analysis results of different points, dichotomy and efficient and accurate sorting are carried out on each type of plastic compared with other types of plastic.

Description

Technical Field

[0001] The present invention belongs to the technical field of spectrum detection, and in particular relates to a spectroscopic sorting system and method based on near-infrared spectrum analysis. Background Art

[0002] In recent years, the widespread use of plastics has intensified the generation of plastic waste. Driven by the concept of circular economy, efficient recycling of plastics has become an urgent need in recent years. Different types of plastics mixed together will cause serious pollution during recycling and reprocessing. Therefore, different plastic resins must be accurately classified before entering the recycling chain. Traditional identification methods include density sorting, temperature difference sorting, wind screening sorting, electrostatic sorting, and flotation separation. The above methods are mainly based on the physical properties of plastics such as density, melting point, and electrical properties. This method is not only subjective but also requires a lot of manpower and material resources.

[0003] At present, the existing plastic sorting systems based on near-infrared spectroscopy technology have the following common spectral acquisition devices: spectrometer, hyperspectral camera, micro-spectral sensor, and a combination of photodiode and narrow-band filter. When the spectral instrument performs spectral detection, plastics can be accurately identified, but the instrument and equipment are complicated to operate, the collection of spectral data is time-consuming, and the manufacturing cost is relatively high. The hyperspectral camera can not only obtain the spectral information of the plastic sample for plastic sorting, but also obtain the characteristic information such as the shape and size of the plastic. However, the data acquisition time of the hyperspectral camera is too long, the recognition efficiency is not high, and the price is expensive, which is difficult to meet the needs of online operations. When using micro-spectral sensors for plastic sorting, the low resolution of the sensor, susceptibility to temperature, and the instability of the light source may have an adverse effect on the acquired spectral data, and the system stability is poor. The combination of multiple groups of photodiodes and narrow-band filters can achieve higher sorting efficiency and miniaturize the equipment. However, each photodiode needs to be matched with a narrow-band filter with a different central wavelength. The use of narrow-band filters will cause the signal intensity received by the photodiode to be greatly reduced, thereby reducing the signal-to-noise ratio of the sorting system and making the recognition error larger. In addition, coated filters are expensive, and increasing the number of filters also increases costs, making them unsuitable for today's sorting needs. Summary of the invention

[0004] The purpose of the present invention is to overcome the deficiencies of the prior art and to provide a spectroscopic plastic sorting system and method based on near-infrared spectroscopy analysis, which can efficiently and flexibly sort different plastic combinations.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] A spectroscopic plastic sorting system based on near-infrared spectroscopy analysis comprises a light source (1), a light focusing unit (2), a light splitting unit (3), a collection unit (4), an identification unit (5), a sorting unit (6), and a conveyor belt (7). The sample to be tested is placed on the conveyor belt (7), the light source (1), the light focusing unit (2), the light splitting unit (3), and the collection unit (4) are all arranged above the conveyor belt (7), the light source (1) emits a light signal to illuminate the conveyor belt (7) above, and the area of ​​the conveyor belt (7) illuminated by the light source (1) is defined as an identification area, the light focusing unit (2) is located on the path of the light reflected by the identification area, the light splitting unit (3) is placed above the light focusing unit (2), and the light splitting unit (3) defines the identification area as a sample to be tested. The reflected light is split into two signal lights emitted in vertical directions; the two light outlets of the splitting unit (3) are each provided with a collecting unit (4) for collecting light signals; the two collecting units (4) are each connected to one identification unit, that is, the two collecting units (4) transmit the collected light signals to one identification unit (5) for identification and analysis; the sorting unit (6) is arranged at the tail end of the conveyor belt (7); the sorting unit (6) is controlled by the identification unit (5) to separate different types of plastics;

[0007] Optionally, the light source (1) is a full-spectrum halogen lamp controlled by a constant current module, which irradiates the identification area of ​​the conveyor belt at a certain angle;

[0008] Optionally, the focusing unit (2) comprises a focusing mirror (21) and a collimating lens (22); the focusing mirror (21) is arranged on a side above the conveyor belt (7) away from the light source (1); the collimating lens (22) is arranged directly above the focusing mirror (21); the light source (1) irradiates the recognition area of ​​the conveyor belt (7), the reflected diffuse light enters the focusing mirror (21) for focusing, the outgoing light of the focusing mirror (21) enters the collimating lens (22), and the outgoing light of the collimating lens (22) enters the light splitting unit;

[0009] Optionally, after the outgoing light of the collimating lens (22) enters the light splitting unit (3), it is split into two or more light splitting beams, and then emitted from two light outlets corresponding to the light splitting unit (3);

[0010] Optionally, a light focusing unit (2) and a light splitting unit (3) are provided on the path of the reflected light in the identification area of ​​each transport channel, and two corresponding collection units (4) of the same light splitting unit (3) are electrically connected to the same identification unit (5);

[0011] Optionally, the acquisition unit (4) comprises two micro-spectral sensors (41) and a signal acquisition module (42), one micro-spectral sensor (41) is electrically connected to one signal acquisition module (42), the micro-spectral sensor (41) is arranged at a light outlet corresponding to the light splitting unit (3), and the split light beam is irradiated on a light receiving area of ​​the corresponding micro-spectral sensor (41);

[0012] Optionally, the signal acquisition module (42) includes a voltage controller (421), a digital-to-analog converter (424), an IV converter (422), a first analog-to-digital converter (425), a voltage follower (423), a second analog-to-digital converter (426), and a data transmission module (427); the micro-spectral sensor (41) is electrically connected to the voltage controller (421), the IV converter (422), and the voltage follower (423), respectively; the voltage controller (421) is connected to the data transmission module (427) through the digital-to-analog converter (424); the IV converter (422) is connected to the first analog-to-digital converter (425) and the data transmission module (427) in sequence; similarly, the voltage follower (423) is connected to the second analog-to-digital converter (426) and the data transmission module (427) in sequence. ;

[0013] Optionally, the conveyor belt (7) includes more than one conveying channel arranged side by side, and the signal light emitted by the light source (1) is irradiated to the identification area on each conveying channel; a corresponding sorting unit (6) is arranged at the tail end of each conveying channel, and each identification unit (5) controls the sorting unit (6) at the tail end of the corresponding conveying channel to sort the tested objects conveyed to the tail end of the current conveying channel;

[0014] Optionally, the sorting unit (6) includes an air jet valve (61) and a plastic receiving box (62), wherein the air jet valve (61) is arranged directly below the tail end of each conveying path, and the plastic collecting box (62) is arranged below the tail end of the conveyor belt (7).

[0015] The present invention also provides a spectroscopic plastic sorting method based on near-infrared spectral analysis, which is applied to a spectroscopic detection device based on spectrum as described above, and comprises the following steps:

[0016] S1. Collect spectral data of five kinds of plastics, namely PET, PC, PE, PS and PP, and use PCA principal component analysis to reduce the dimensionality of the spectral data of the five kinds of plastics to obtain several optimal characteristic wavelength points;

[0017] S2. Determine whether the signal light emitted by the light source is irradiated on the identification area of ​​the conveyor belt, determine whether the conveyor belt is at a uniform speed and the conveying speed v, determine the distance L between the conveyor belt identification area and the end of the conveyor belt, and pre-calculate the time t for the sample to be tested to travel from the identification area to the end of the conveyor belt 0 ;

[0018] S3, the unloading device accurately places the tested samples on each conveying lane of the conveyor belt at equal time intervals;

[0019] S4, determine two or more characteristic wavelengths corresponding to the plastics identified on each conveyor belt, for example, conveyor belt conveyor lane 1 identifies PET and non-PET, and sets the wavelengths λ collected by the two sensors respectively 1 , 2 PET and non-PET are divided into two groups by the ratio of the collected wavelengths. Similarly, the spectrum collection unit corresponding to each transport channel is set with a corresponding specific collection wavelength;

[0020] S5. The sample to be tested is evenly placed on each conveying lane of the conveyor belt through the unloading device. When it is transported to the identification area, the light emitted by the light source irradiates the object to be tested. The reflected light of the object to be tested in the identification area is split by the spectroscopic unit into two signal lights emitted in the vertical direction, namely, light beam a and light beam b. Light beam a and light beam b are emitted from two light outlets of the spectroscopic unit respectively.

[0021] S6. The collection units at the two light outlets of the same spectroscopic unit collect the light intensity of the light beam a and the light beam b respectively, and each of the two collection units transmitted to the corresponding spectrum collection unit collects the light intensity of the sample to be tested at different characteristic wavelength points according to the needs, and then enters the same identification unit for identification and analysis;

[0022] S7. When the tested object reaches the tail end of the conveyor belt, the sorting unit corresponding to each conveying channel is controlled by the identification unit to classify the tested plastic into two categories.

[0023] Optionally, S1 includes the following specific steps:

[0024] S11. Analyze the diffuse reflectance spectrum curves of five plastics, PET, PC, PE, PS, and PP, in the near-infrared band. Through analysis, it is found that there is an obvious absorption peak in the 1550nm-1850nm band. Subsequently, the corresponding spectrum sensor is selected according to the selected specific band. The technicians determine the set wavelength range and each wavelength point within the set wavelength range according to the detection purpose;

[0025] S12. The plastic spectrum data is analyzed using principal component analysis. According to the principal component analysis, the contribution rates of the first two principal components are greater than 1%, and the cumulative contribution rate reaches 99.685%. The load coefficients of the two principal components with the largest variance contribution rate take a maximum value and a minimum value, and the wavelengths corresponding to the two extreme values ​​are recorded as the first wavelength λ 1 , the second wavelength λ 2 .

[0026] S13, analyzing the spectrum data of five kinds of plastics, the first wavelength λ corresponding to the first principal component 1 =

[0027] 1670nm, second wavelength λ 2 =1754nm, find the ratio of the reflectivity of these five plastics at the first wavelength and the second wavelength. The ratios of PC, PS and PE are significantly different from those of other plastics, so the above three plastics can be sorted first;

[0028] S14, the first wavelength λ corresponding to the second principal component 1 =1670nm, second wavelength λ 2 =1694nm, the ratio of the wavelength corresponding to each plastic has a large degree of differentiation, and PET and PP plastics can be sorted.

[0029] Optionally, S5 includes the following specific steps:

[0030] S51, the reflected light of the object to be tested in the identification area is firstly focused by entering the condenser, and the light emitted by the condenser is collimated by the collimating lens and then enters the spectroscopic unit. The spectroscopic unit splits the light emitted by the condenser into two signal lights emitted in vertical directions, namely, light beam a and light beam b. The light outlet of the spectroscopic unit projects the split light beams to the corresponding spectrum acquisition unit.

[0031] Optionally, S6 includes the following specific steps:

[0032] S61, the collection units at the two light outlets of the same light splitting unit collect the light intensity of light beam a and light beam b respectively, convert them into current signals, and then send them to the same identification unit. The current signal corresponding to light beam a is recorded as I λ1 , the current signal corresponding to the beam b is recorded as I λ2 ;

[0033] S62, the recognition unit calculates the reflectivity ratio R of the light beam a and the light beam b: R=I λ1 / I λ2 ;

[0034] S63, if R 0 <R<R 1, the recognition unit determines that the sample under test in the current conveyor identification area is recycled plastic, otherwise it is non-recycled plastic, where R 0 , R 1 is the first sorting parameter, R of each conveying lane 0 , R 1 Are different.

[0035] Optionally, S7 includes the following specific steps:

[0036] S71, the recognition unit records the time consumed in each round of recognition and analysis as t1. When the recognition unit determines that the sample under test in the recognition area of ​​the current conveying channel is the recycled plastic, the recognition unit controls the jet valve of the corresponding conveying channel to spray after the time interval t0-t1 from the time the judgment result is obtained, so that the sample under test is pushed by the gas ejected from the corresponding jet valve when falling from the tail end of the conveyor belt and falls into the corresponding plastic recycling receiving box of the current conveying channel. Otherwise, the jet valve of the corresponding conveying channel will not operate.

[0037] The beneficial effects of the present invention are:

[0038] (1) Compared with the spectrometer measurement in the prior art, the spectroscopic sorting system based on near-infrared spectral analysis proposed in the present invention has a simple structure, is easy to operate, and has a low design cost while ensuring the sorting accuracy.

[0039] (2) According to the present invention, for different plastics, each conveying lane will be set with a corresponding first wavelength and a second wavelength for collection. The two micro-spectral sensors corresponding to each conveying lane can flexibly change the characteristic wavelength collected, and distinguish different plastics by the reflectivity ratio of the two wavelengths.

[0040] (3) The present invention uses coaxial light splitting to simultaneously collect two different characteristic wavelengths for ratio, thereby avoiding the influence of temperature and light source instability, overcoming the problem of poor sampling stability of the micro-spectral sensor, and improving system stability;

[0041] (4) When the spectrometer in the prior art collects spectral data of a sample, it usually continuously collects spectral information at a certain wavelength interval, and extracts useful characteristic wavelengths through subsequent processing. Each collection time is too long. The present invention uses two miniature spectral sensors to simultaneously collect the optimal detection wavelength representing the significant information of the sample under test, and each collection of a wavelength point only takes 1 ms, which greatly shortens the collection time.

[0042] (5) Compared with the spectral instruments in the prior art, which can only continuously scan and collect spectral information under a certain wavelength, it is impossible to collect spectral information of a single characteristic wavelength according to demand. The present invention only needs to set and collect specific wavelengths through the full-spectrum light source and the spectral collection unit, and collect the spectral information of multiple characteristic wavelengths set in a specific band at a single point, thereby realizing dynamic sampling and online detection. While improving the flexibility of the entire sorting system, the sorting efficiency is also improved.

[0043] (6) In the present invention, a plurality of conveyor lanes are arranged on the conveyor belt, and each conveyor lane has a fixed area as an identification area. That is, the present invention realizes the simultaneous sorting of multiple conveyor lanes with a single light source, thereby improving the sorting efficiency and further saving costs.

[0044] In response to the need to sort multiple groups of different types of plastic combinations, the spectroscopic sorting system of the present invention selects corresponding characteristic wavelengths for different plastic combinations, and then sets them as the corresponding plastics to be identified in each conveyor lane and the wavelengths to be collected by the spectral collection unit at a single point. It can efficiently sort a variety of plastic combinations including but not limited to PET, PVC, PE, PS, PP, etc., and can flexibly realize the identification and sorting of different types of plastic combinations without changing the sorting system. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic diagram of the two-dimensional structure of the optical separation sorting system of the present invention;

[0046] Figure 2 It is a schematic diagram of a three-dimensional structure of part of the optical separation sorting system of the present invention;

[0047] Figure 3 This is the structure diagram of the spectrum acquisition module;

[0048] Figure 4 It is a flow chart of the detection method of the present invention;

[0049] The actual correspondence between the reference numerals and component names of the present invention is as follows:

[0050] 1. Light source; 2. Focusing unit; 21. Condenser; 22. Collimating lens; 3. Spectroscopic unit; 4. Collection unit; 41. Micro-spectral sensor; 42. Signal acquisition module; 421. Voltage controller; 422. IV converter; 423. Voltage follower; 424. Digital-to-analog converter; 425. First analog-to-digital converter; 426. Second analog-to-digital converter; 427. Data transmission module; 5. Identification unit; 6. Sorting unit; 61. Jet valve; 62. Material collection box; 7. Conveyor belt; 8. Unloading unit; 9. Sample to be tested. DETAILED DESCRIPTION

[0051] In order to make the technical solution of the present invention clearer and more specific, the present invention is clearly and completely described below in conjunction with the accompanying drawings. Any equivalent replacement of the technical features of the technical solution of the present invention and any solution derived by conventional reasoning by ordinary technicians in the field without making any creative work shall fall within the protection scope of the present invention.

[0052] refer to Figure 1-Figure 4 , a spectroscopic plastic sorting method based on near infrared spectroscopy analysis, comprising the following steps:

[0053] S1. Collect spectral data of five kinds of plastics, namely PET, PC, PE, PS and PP, and use PCA principal component analysis to reduce the dimensionality of the spectral data of the five kinds of plastics to obtain several optimal characteristic wavelength points;

[0054] S2, determine that the signal light emitted by the light source (1) is irradiated on the identification area of ​​the conveyor belt, determine whether the conveyor belt (7) is moving at a uniform speed and the conveying speed v, determine the distance L between the identification area of ​​the conveyor belt and the end of the conveyor belt, and pre-calculate the time t for the measured sample (9) to travel from the identification area to the end of the conveyor belt 0 ;

[0055] S3, the unloading device (8) accurately places the tested samples (9) on each conveying lane of the conveyor belt at equal time intervals;

[0056] S4, determining two or more characteristic wavelengths corresponding to the plastics identified on each conveyor belt, for example, conveyor belt (7) conveyor lane 1 identifies PET and non-PET, and setting the wavelengths λ collected by two spectral sensors (41) respectively 1 , 2 PET and non-PET are divided into two groups by the ratio of the collected wavelengths. Similarly, the spectrum collection unit corresponding to each transport channel is set with a corresponding specific collection wavelength;

[0057] S5. The sample to be tested is evenly placed on each conveying lane of the conveyor belt through the unloading device (8). When it is transported to the identification area, the light emitted by the light source (1) irradiates the object to be tested (9). The reflected light of the object to be tested (9) in the identification area is split by the light splitting unit (3) into two signal light beams emitted in a vertical direction, namely, light beam a and light beam b. Light beam a and light beam b are emitted from two light outlets of the light splitting unit respectively.

[0058] S6, the collection units at the two light outlets of the same spectroscopic unit (3) respectively collect the light intensity of the light beam a and the light beam b, and each is transmitted to the two collection units (4) of the corresponding spectrum collection unit to collect the light intensity of the sample to be tested at different characteristic wavelength points according to the needs, and then enter the same identification unit (5) for identification and analysis;

[0059] S7. When the tested object (9) reaches the tail end of the conveyor belt (9), the sorting unit (6) corresponding to each conveying channel is controlled by the identification unit (5) to classify the tested plastic (9) into two categories.

[0060] In this implementation, S1 includes the following steps:

[0061] S11. Analyze the diffuse reflectance spectrum curves of five kinds of plastics, PET, PC, PE, PS and PP, in the near-infrared band. If there is an obvious absorption peak in the 1550nm-1850nm band, select the corresponding spectrum sensor according to the specific band. The technicians determine the set wavelength range and each wavelength point within the set wavelength range according to the detection purpose.

[0062] S12. Analyze the plastic spectrum data using principal component analysis. According to the principal component analysis, the contribution rates of the first two principal components are both greater than 1%, and the cumulative contribution rate reaches 99.685%. The load coefficients of the two principal components with the largest variance contribution rate take a maximum value and a minimum value, and the wavelengths corresponding to the two extreme values ​​are recorded as the first wavelength λ 1 , the second wavelength λ 2 ;

[0063] S13, analyzing the spectrum data of five kinds of plastics, the first wavelength λ corresponding to the first principal component 1 =

[0064] 1670nm, second wavelength λ 2 =1754nm, find the ratio of the reflectivity of these five plastics at the first wavelength and the second wavelength. The ratios of PC, PS and PE are different from those of other plastics. So first sort the above three plastics.

[0065] S14, the first wavelength λ corresponding to the second principal component 1 =1670nm, second wavelength λ 2 =1694nm, the ratio of the wavelength corresponding to each plastic has a large degree of differentiation, and PET and PP plastics can be sorted.

[0066] In this implementation, S6 includes the following steps:

[0067] S61, the collection units (4) at the two light outlets of the same light splitting unit (3) respectively collect the light intensity of the light beam a and the light beam b, convert them into current signals, and then send them to the same identification unit (5), and the current signal corresponding to the light beam a is recorded as I λ1 , the current signal corresponding to the beam b is recorded as I λ2 ;

[0068] S62, the recognition unit calculates the reflectivity ratio R of the light beam a and the light beam b: R=I λ1 / Iλ2 ;

[0069] S63, if R 0 <R<R 1 , then the identification unit (5) determines that the sample (9) under test in the identification area of ​​the current conveying path is recycled plastic, otherwise it is non-recycled plastic, wherein R 0 , R 1 is the first sorting parameter.

[0070] In this implementation, S7 includes the following steps:

[0071] S71, the identification unit (9) records the time spent in each round of identification analysis as t1. When the identification unit (5) determines that the sample under test in the identification area of ​​the current conveying channel is the recycled plastic, the identification unit (5) controls the jet valve (61) of the corresponding conveying channel to spray after the time interval t0-t1 from the time the determination result is obtained, so that when the sample under test falls from the tail end of the conveyor belt (7), it is pushed by the gas ejected by the corresponding jet valve (61) and falls into the corresponding plastic recycling receiving box (62) of the current conveying channel. Otherwise, the jet valve (61) of the corresponding conveying channel does not operate.

[0072] Example:

[0073] like Figure 1 The figure is a schematic structural diagram of a spectroscopic plastic sorting system based on near-infrared spectroscopy analysis of the present invention.

[0074] A spectroscopic plastic sorting system based on near-infrared spectral analysis comprises a full-spectrum light source (1), a light focusing unit (2), a spectroscopic unit (3), a collection unit (4), an identification unit (5), a sorting unit (6), a conveyor belt (7), and a material unloading unit (8).

[0075] In the sorting system, a light source (1), a focusing unit (2), a spectroscopic unit (3), a spectrum collection unit (4), and a feeding unit (8) are all placed above a conveyor belt (7); the feeding unit (8) is arranged above the head end of the conveyor belt (7), and the sorting unit (6) is arranged below the tail end of the conveyor belt (7); the plastic (9) to be tested falls from the discharge port at the bottom of the feeding device (8) to the head end of the conveyor belt (7) below, and reaches the tail end of the conveyor belt (7) as the conveyor belt (7) is conveyed; the light source (1) irradiates the identification area of ​​the conveyor belt (7) below, the focusing unit (2) is located on the path of the light reflected from the identification area, the spectroscopic unit (3) is located above the focusing unit (2), and the spectrum collection unit (4) is arranged on the path of the outgoing light of the spectroscopic unit (3); when the sample (9) to be tested is conveyed to the identification area of ​​the conveyor belt (7), the full spectrum The light emitted by the light source (1) is irradiated onto the sample to be tested (9), and the focusing unit (2) focuses and collimates the light diffusely reflected by the sample to be tested (9) and then emits it to the spectroscopic unit (3). The spectroscopic unit (3) splits the incident light into two mutually orthogonal signal lights. Two corresponding collection units (4) are arranged at the spectroscopic port of one spectroscopic unit (3) to collect the light signal under the corresponding incident light; the two collection units (4) corresponding to the same spectroscopic unit (3) are electrically connected to a recognition unit (5), and one recognition unit (5) is electrically connected to a sorting unit (6); the two collection units (4) transmit the collected light signals to the corresponding recognition unit (5) for recognition and analysis. When the plastic reaches the tail end of the conveyor belt (7), the sorting unit (6) is controlled by the recognition unit (5) to sort the plastic into the corresponding material box (62).

[0076] Specific:

[0077] The light source (1) adopts a full-spectrum halogen lamp controlled by a constant current module, which can generate stable light in the near-infrared band.

[0078] Five different types of plastics, PET, PC, PE, PS, and PP, were recorded as the tested items.

[0079] Optionally, one or more discharge ports are provided on the unloading device (8), and one or more conveying paths arranged side by side are correspondingly provided on the conveyor belt (7), and the number and positions of the discharge ports correspond to the number and positions of the conveying paths, see Figure 2 Each conveyor path has an identification area illuminated by the light emitted by the light source (1), such as Figure 2 Each conveyor lane is very narrow, and the samples to be tested on each conveyor lane can only be arranged in sequence along the conveying direction, so it is almost impossible for two or more tested objects to appear side by side in the identification area on the same conveyor lane.

[0080] Figure 2The sample to be tested (9), the collection unit (3), and the sorting unit (6) corresponding to each conveying channel are shown. The specific structure is shown in Figure 1 .

[0081] A focusing unit (2) and a splitting unit (3) are arranged on the reflected light path of the identification area on each conveying path. The focusing unit (2) includes a condenser (21) and a collimating lens (22). The collimating lens (22) is arranged directly above the condenser (21). The sample to be tested (9) reflects diffuse reflected light with sample spectrum information. After the diffuse reflected light is converged by the condenser (21), it is emitted to the collimating lens (22) above. The collimating lens (22) collimates the incident light and then emits it to the splitting unit (3). The splitting unit (3) is a splitting plate. The splitting ratio of the splitting plate in the 1200-1850nm band is 50 / 50±5%. The splitting unit (3) splits the incident light into two signal lights with approximately the same energy and orthogonal to each other and emits them from the light outlet of the splitting unit (3). The collection units (4) provided at each light outlet collect light signals of a wavelength set in advance, and convert the light intensity into a voltage signal and transmit it to an identification unit (5). The identification unit (5) identifies the object to be tested based on the calculated light intensity of the two different wavelengths collected to determine whether it is the identified plastic. At the same time, the identification unit (5) calculates the time t when the sample to be tested reaches the end of the conveyor belt according to the conveyor belt speed. 0 , and controls the separation unit (6) to perform corresponding separation.

[0082] The spectrum acquisition unit (4) includes two micro-spectral sensors (41) and a signal acquisition module (42). The micro-spectral sensors (41) collect spectrum data of the sample to be tested in the wavelength range of 1550nm-1850nm for qualitative analysis and classification. The micro-spectral sensors (41) are respectively arranged at the light outlets of the light-splitting unit (3). After the split light beams are emitted from the light outlets, they are irradiated on the light receiving area of ​​the micro-spectral sensor (41). The micro-spectral sensor (41) and the signal acquisition module (42) are electrically connected. The micro-spectral sensor (41) adjusts the light transmission wavelength to the set wavelength according to the signal acquisition module (42). Only the light beams that meet the light transmission wavelength and are irradiated on the light receiving area will be collected by the micro-spectral sensor (41) and sent to the signal acquisition module (42). The signal acquisition module (42) sends the collected light intensity and other light beam information to the recognition unit (5) for analysis.

[0083] Figure 3The structure schematic diagram of the spectrum acquisition unit 4 is shown, a micro-spectral sensor (41) is connected to a signal acquisition module (42), and the signal acquisition module (42) includes a voltage controller (421), an IV converter (422), a voltage follower (423), a digital-to-analog converter (424), a first analog-to-digital converter (425), a second analog-to-digital converter (426), and a data transmission module (427). Specifically, the micro-spectral sensor (41) is electrically connected to the voltage monitor (421), the IV converter (422), the voltage follower (423), the digital-to-analog converter (424), the first analog-to-digital converter (425), the second analog-to-digital converter (426), and the data transmission module (427). The thermistor inside the sensor collects temperature in real time and converts it into voltage, which is converted into a digital signal by the second analog-to-digital converter 426 and transmitted to the identification unit (5) through the data transmission module (427). According to the collected temperature, the control voltage applied to the micro-spectral sensor (41) by the voltage controller (421) is adjusted through the digital-to-analog converter (424). According to different control voltages loaded into the micro-spectral sensor (41), spectral information of corresponding wavelengths is collected and converted into voltage signals through an IV converter (422). The first analog-to-digital converter (425) converts the electrical signal into a digital signal, which is then transmitted to the identification unit (5) through a data transmission module (427). The identification unit (5) identifies the corresponding sample under test according to the information in the microprocessor.

[0084] In this embodiment, the micro-spectral sensor (41) has a built-in MEMS-FPI tunable filter, an InGaAs photodiode and a thermistor. When the split light beam is irradiated on the light receiving area, the temperature of the light receiving area will change. The temperature change of the light receiving area will affect the light transmission wavelength of the micro-spectral sensor (41). Therefore, it is necessary to collect the voltage on the thermistor in the light receiving area to obtain the temperature in the light receiving area; the voltage controller (421) is adjusted according to the real-time temperature in the light receiving area to adjust the voltage at both ends of the MEMS-FPI tunable filter in the micro-spectral sensor (41), and control the light of a specific wavelength to pass through the light receiving area of ​​the micro-spectral sensor (41), so that the micro-spectral sensor (41) collects spectral information of the corresponding wavelength; and it is ensured that the wavelength collected by the micro-spectral sensor (41) is the corresponding wavelength set according to the demand.

[0085] A sorting unit (6) is provided at the tail end of each conveying path, and the identification unit (5) is electrically connected to the sorting unit (6) of the corresponding conveying path. The sorting unit (6) comprises an air jet valve (61) and a plastic collection box (62). The air jet valve (61) is provided directly below the tail end of each conveying path, and the plastic collection box (62) is arranged below the tail end of the conveyor belt (7); each air jet valve (61) is controlled by an execution signal of the identification unit (5) to spray air, and sorts the sample to be tested into the plastic collection box (62) of the corresponding plastic type.

[0086] Before the present invention is officially put into use, it is necessary to first collect the spectral information of the plastic to be sorted in the 1550nm-1850nm band, use PCA to process the spectral data, and select the two optimal wavelengths λ for each plastic to distinguish other plastics. 1 , 2 In this embodiment, five kinds of plastics, including but not limited to PET, PC, PE, PS, and PP, are selected as sorting samples, and three optimal characteristic wavelengths are selected to sort the five kinds of plastics.

[0087] After processing the spectral data, two optimal wavelengths are selected for each of the above-mentioned plastics that have obvious characteristics compared to other plastics. The technicians set in advance the two characteristic wavelength points collected by the collection unit (4) corresponding to each conveying lane.

[0088] The present invention can flexibly change the optimal characteristic wavelength collected by using a micro-spectral sensor (41) for different plastics, and identify different plastics through the reflectivity ratio of two wavelengths, with simple operation, high identification efficiency and high accuracy.

[0089] Compared with the spectrometer measurement in the prior art, the spectroscopic sorting system based on near-infrared spectral analysis proposed in the present invention has a simple structure, is easy to operate, and has a low design cost while ensuring the sorting accuracy. Compared with the spectral instruments in the prior art, which can only continuously scan and collect spectral information under a wavelength, it is impossible to collect spectral information of a single-point characteristic wavelength according to demand. The present invention only needs to set the collection of a specific wavelength through a full-spectrum light source and a spectral acquisition unit, and collects the spectral information of multiple characteristic wavelengths set at a single point in a specific band, thereby realizing dynamic sampling and online detection. While improving the flexibility of the entire sorting system, the sorting efficiency is also improved. The present invention uses coaxial spectroscopy to simultaneously collect two different characteristic wavelengths for ratio, thereby avoiding the influence of temperature and light source instability, overcoming the problem of poor sampling stability of micro-spectral sensors, and improving system stability;

[0090] In the present invention, a plurality of conveyor paths are arranged on the conveyor belt (7), and a fixed area is provided on each conveyor path as an identification area, that is, the present invention realizes the simultaneous sorting of multiple conveyor paths with a single light source, improves the sorting efficiency, and further saves costs. For the sorting requirements of multiple groups of different types of plastic combinations, the spectroscopic sorting system of the present invention selects corresponding characteristic wavelengths for different plastic combinations, and then sets the wavelengths corresponding to the plastics to be identified on each conveyor path and the single-point collection of the spectrum collection unit, so as to efficiently sort multiple plastic combinations including but not limited to PET, PVC, PE, PS, PP, etc., and can flexibly realize the identification and sorting of different types of plastic combinations without changing the sorting system.

[0091] At the same time, compared with the spectrometer measurement in the prior art, the present invention has low cost, and while ensuring the sorting accuracy, the present invention is simple to operate and has high sorting efficiency.

[0092] The sorting system of the present invention has been verified for five types of plastics, namely PET, PC, PE, PS and PP, and the sorting accuracy is almost the same as that of the spectrometer measurement in the prior art.

[0093] The sorting system and sorting method of the present invention are not limited to the above five types of plastics: PET, PC, PE, PS, and PP, but can also identify and sort other types of plastics. It only requires re-analyzing the spectral information of the object to be tested and selecting the optimal characteristic wavelength for ratio discrimination or linear discrimination.

[0094] The techniques, shapes, and structural parts not described in detail in the present invention are all well-known techniques.

[0095] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A spectroscopic plastic sorting system based on near infrared spectroscopy analysis, characterized in that: include: A light source (1), a light focusing unit (2), a light splitting unit (3), a collection unit (4), an identification unit (5), a sorting unit (6), and a conveyor belt (7); The sample to be tested is placed on the conveyor belt, and the light source (1), the light focusing unit (2), the light splitting unit (3), and the collection unit (4) are all arranged above the conveyor belt (7); The light source (1) emits a light signal to illuminate the conveyor belt (7) above, and the area of ​​the conveyor belt (7) illuminated by the light source (1) is defined as an identification area; The light focusing unit (2) is located on the path of the light reflected from the identification area, the light splitting unit (3) is placed above the light focusing unit (2), and the light splitting unit (3) splits the light reflected from the identification area into two signal lights emitted in vertical directions; The two light outlets of the light splitting unit (3) are both provided with the collecting unit (4) for collecting light signals, and the two collecting units (4) are both connected to one identification unit (5), that is, the two collecting units (4) transmit the collected light signals to one identification unit (5) for identification and analysis; The sorting unit (6) is arranged at the tail end of the conveyor belt (7), and the sorting unit (6) is controlled by the identification unit (5) to separate different types of plastics.

2. A spectroscopic plastic sorting system based on near infrared spectroscopy analysis, characterized in that: The focusing unit (2) comprises a focusing mirror (21) and a collimating lens (22); the focusing mirror (21) is arranged on a side above the conveyor belt (7) away from the light source (1); the collimating lens (22) is arranged directly above the focusing mirror (21); the diffusely reflected light irradiated by the light source (1) and reflected in the identification area of ​​the conveyor belt (7) enters the focusing mirror (21); the outgoing light of the focusing mirror (21) enters the collimating lens (22); and the outgoing light of the collimating lens (22) enters the light splitting unit (3).

3. A spectroscopic plastic sorting system based on near infrared spectroscopy analysis according to claim 2, characterized in that: After the outgoing light of the collimating lens (22) enters the light splitting unit (3), it is split into two or more light splitting beams and then emitted from the light outlet corresponding to the light splitting unit (3). A focusing unit and a light splitting unit are arranged on the reflected light path of the identification area of ​​each conveying channel. Two collecting units corresponding to the same light splitting unit are electrically connected to the same identification unit. A corresponding sorting unit (6) is arranged at the tail end of each conveying channel. Each identification unit (5) controls the sorting unit (6) at the tail end of the corresponding conveying channel to sort the tested objects (9) that are transported to the tail end of the current conveying channel.

4. The spectroscopic plastic sorting system based on near infrared spectroscopy analysis according to claim 3 is characterized in that: The acquisition unit (4) comprises two micro-spectral sensors (41) and a signal acquisition module (42); one micro-spectral sensor (41) is electrically connected to one signal acquisition module (42); the micro-spectral sensor (41) is arranged at a light outlet corresponding to the light splitting unit (3); and the split light beam is irradiated on a light receiving area of ​​the corresponding micro-spectral sensor (41).

5. The spectroscopic plastic sorting system based on near infrared spectroscopy analysis according to claim 4 is characterized in that: The signal acquisition module (42) comprises a voltage controller (421), a digital-to-analog converter (424), an IV converter (422), a first analog-to-digital converter (425), a voltage follower (423), a second analog-to-digital converter (426), and a data transmission module (427). The micro-spectral sensor (41) is electrically connected to the voltage controller (421), the IV converter (422), and the voltage follower (423), respectively. The voltage controller (421) is connected to the data transmission module (427) via the digital-to-analog converter (424). The IV converter (422) is connected to the first analog-to-digital converter (425) and the data transmission module (427) in sequence. Similarly, the voltage follower (423) is connected to the second analog-to-digital converter (426) and the data transmission module (427) in sequence.

6. The spectroscopic plastic sorting system based on near infrared spectroscopy analysis according to claim 5, characterized in that: The conveyor belt (7) includes more than one conveying channel arranged in parallel, and each conveying channel has a signal light emitted by the light source (1) irradiating the identification area; a corresponding sorting unit (6) is arranged at the tail end of each conveying channel, and each identification unit (5) controls the sorting unit (6) at the tail end of the corresponding conveying channel to sort the tested objects conveyed to the tail end of the current conveying channel; the sorting unit (6) includes an air jet valve (61) and a plastic receiving box (62), the air jet valve (61) is arranged directly below the tail end of each conveying channel, and the plastic collecting box (62) is arranged below the tail end of the conveyor belt.

7. A spectroscopic plastic sorting method based on near infrared spectroscopy analysis, characterized in that: The following steps are involved: S1. Collect spectral data of five kinds of plastics, namely PET, PC, PE, PS and PP, and use PCA principal component analysis to reduce the dimensionality of the spectral data of the five kinds of plastics to obtain several optimal characteristic wavelength points; S2, determine that the signal light emitted by the light source (1) is irradiated on the identification area of ​​the conveyor belt, determine whether the conveyor belt (7) is moving at a uniform speed and the conveying speed v, determine the distance L between the identification area of ​​the conveyor belt and the end of the conveyor belt, and pre-calculate the time t0 for the sample (9) to travel from the identification area to the end of the conveyor belt; S3, the unloading device (8) accurately places the tested samples (9) on each conveying lane of the conveyor belt at equal time intervals; S4, determining two or more characteristic wavelengths corresponding to the plastics identified on each conveyor belt, for example, conveyor belt (7) conveyor lane 1 identifies PET and non-PET, setting wavelengths λ1 and λ2 collected by two spectral sensors (41) respectively, and dividing PET and non-PET into two by the ratio of the collected wavelengths. Similarly, the spectral collection unit corresponding to each conveyor lane is set with a corresponding collection specific wavelength; S5. The sample to be tested is evenly placed on each conveying lane of the conveyor belt through the unloading device (8). When it is transported to the identification area, the light emitted by the light source (1) irradiates the object to be tested (9). The reflected light of the object to be tested (9) in the identification area is split by the light splitting unit (3) into two signal light beams emitted in a vertical direction, namely, light beam a and light beam b. Light beam a and light beam b are emitted from two light outlets of the light splitting unit respectively. S6. The acquisition units at the two light output ports of the same beam splitting unit (3) respectively acquire the light intensities of beam a and beam b. Each of the two acquisition units (4) that deliver to the corresponding spectral acquisition unit acquires the light intensity of the sample to be measured at different characteristic wavelength points as required, and then enters the same identification unit (5) for identification and analysis; S7. When the item to be measured (9) reaches the end of the conveyor belt (9), the sorting unit (6) corresponding to each conveying channel is controlled by the identification unit (5) to perform binary classification on the plastic to be measured (9).

8. The method for separating plastics by spectroscopic analysis based on near infrared spectroscopy according to claim 7, characterized in that: The specific steps of S1 are as follows: S11. Analyze the diffuse reflection spectral curves of five plastics, namely PET, PC, PE, PS, and PP, in the near-infrared band. If there are obvious absorption peaks in the 1550 nm - 1850 nm band, then select the corresponding spectral sensor according to the specific band. The technical personnel determine the set wavelength range and each wavelength point within the set wavelength range according to the detection purpose; S12. Use the principal component analysis method to analyze the plastic spectral data. According to the principal component analysis, the contribution rates of the first two principal components of the first two principal components are both greater than 1%, and the cumulative contribution rate reaches 99.685%. Take a maximum value and a minimum value for the load coefficients of the two principal components with the largest variance contribution rate. The wavelengths corresponding to the two extreme values are recorded as the first wavelength λ1 and the second wavelength λ2; S13. Analyze the spectral data of the five plastics. The first wavelength λ1 corresponding to the first principal component is 1670 nm, and the second wavelength λ2 is 1754 nm. Calculate the ratio of the reflectivities of the five plastics at the first wavelength and the second wavelength. The ratios of PC, PS, and PE plastics are relatively large compared to other plastics. First, sort the above three plastics; S14. The first wavelength λ1 corresponding to the second principal component is 1670 nm, and the second wavelength λ2 is 1694 nm. The ratio of each plastic corresponding to the wavelength has a large discrimination degree. Sort PET and PP plastics.

9. The method for separating plastics by spectroscopic analysis based on near infrared spectroscopy according to claim 7, characterized in that: The specific steps of S6 are as follows: S61, the collection units (4) at the two light outlets of the same light splitting unit (3) respectively collect the light intensity of the light beam a and the light beam b, convert them into current signals, and then send them to the same identification unit (5), and the current signal corresponding to the light beam a is recorded as I λ1 , the current signal corresponding to the beam b is recorded as I λ2 ; S62, the recognition unit calculates the reflectivity ratio R of the light beam a and the light beam b: R=I λ1 / I λ2 ; S63. If R0 < R < R1, the identification unit (5) determines that the sample to be measured (9) within the identification area of the current conveying channel is the recycled plastic, otherwise it is non-recycled plastic, where R0 and R1 are the first sorting parameters.

10. The method for separating plastics by spectroscopic analysis based on near infrared spectroscopy according to claim 7, characterized in that: The specific steps of S7 are as follows: S71. The identification unit (9) records the time consumed for each round of identification and analysis as t1. When the identification unit (5) determines that the sample to be measured within the identification area of the current conveying channel is the recycled plastic, then starting from the time when the discrimination result is obtained, the identification unit (5) controls the air jet valve (61) of the corresponding conveying channel to jet air after a time interval of t0 - t1, so that when the sample to be measured falls from the end of the conveyor belt (7), it is pushed by the gas ejected from the corresponding air jet valve (61) and falls into the plastic recycling receiving box (62) corresponding to the current conveying channel. Otherwise, the air jet valve (61) of the corresponding conveying channel does not act.

Images