JP7419093B2 - Waste plastic sorting equipment, sorting method, and sorting program - Google Patents

Description

The present disclosure relates to a waste plastic sorting device, a sorting method, and a sorting program.

In order to achieve material recycling in the reprocessing of waste plastics, it is required that the products after sorting have little contamination with non-target substances and have high purity. In addition, if the materials include expensive materials, it is required to be able to sort out the expensive materials without leaving anything out. Furthermore, in order to recycle black plastic, which previously had to be thermally recycled because it cannot be separated, there is a need to efficiently identify and sort the material.

Patent Document 1 discloses that an object to be sorted is irradiated with infrared light, reflected light from the object to be sorted is received, and the resin type of the object to be sorted is determined by a pattern matching method using a spectrum based on the reflected light. It is stated that.

JP 2018-100903 Publication

However, in the conventional method described in Patent Document 1 and the like, there is room for improvement in the accuracy of material determination.

The present disclosure aims to provide a sorting device, a sorting method, and a sorting program that can improve the accuracy of determining the material of waste plastic.

ここで、Ｓ _ｏｒｇ （ｎ，ｗ）は、前記反射スペクトル検出部により検出された前記スペクトルであり、Ｗ _ｒｅｆ （ｎ，ｗ）は、前記第１の補正用スペクトルであり、Ｄ _ｒｅｆ （ｎ，ｗ）は、前記第２の補正用スペクトルであり、Ｓ _ｃｏｒ （ｎ，ｗ）は、補正後の前記スペクトルである。
A waste plastic sorting device according to one aspect of an embodiment of the present invention includes an irradiation section that irradiates light onto waste plastic pieces conveyed on a conveyance path, and a device that receives reflected light of the light irradiated by the irradiation section. a reflection spectrum detection unit that detects a spectrum of the reflected light; a first correction spectrum that is obtained by measuring the spectrum detected by the reflection spectrum detection unit under conditions in which the reflected light is bright; and a first correction spectrum that is darker than the bright conditions; a pre-processing unit that performs correction using a second correction spectrum measured under the conditions; and a second unit that determines whether the spectrum corrected by the pre-processing unit belongs to either the waste plastic piece or the conveyance path. a second determination unit that extracts a feature amount from the spectrum determined to be the waste plastic piece by the first determination unit; a third determination unit that determines the material of the waste plastic piece; and a third determination unit that determines the material of the waste plastic piece by controlling the timing of injecting air to the waste plastic piece conveyed on the conveyance path according to the material determination result by the third determination unit. a sorting unit that sorts and collects the waste plastic pieces by material by sorting and dropping them into a plurality of areas ,
The preprocessing unit corrects the spectrum using the following formula,

Here, S _org (n, w) is the spectrum detected by the reflection spectrum detection section, W _ref (n, w) is the first correction spectrum, and D _ref (n, w) is the second correction spectrum, and S _cor (n, w) is the spectrum after correction .

According to the present disclosure, it is possible to provide a sorting device, a sorting method, and a sorting program that can improve the accuracy of determining the material of waste plastic.

A perspective view showing a schematic configuration of a waste plastic material determination device according to an embodiment. Side view of the waste plastic material determination device shown in Figure 1 A plan view of the waste plastic material determination device shown in Figure 1. Functional block diagram of discrimination device Flowchart of waste plastic material determination processing according to the embodiment Diagram showing the method for extracting the spectrum for correction Diagram showing an example of the process of cutting out a characteristic wavelength region from the reflected wave spectrum Diagram showing an example of feature data extraction Diagram showing an example of material discrimination using a decision tree A plan view showing the first material sorting method by the material determining device of this embodiment. A plan view showing the second material sorting method by the material determining device of this embodiment. A plan view showing the third material sorting method by the material determining device of this embodiment A plan view showing the fourth material sorting method by the material determining device of this embodiment. A plan view showing the fifth material sorting method by the material determining device of this embodiment. Diagram showing an example of the operation screen of the material determination device

Embodiments will be described below with reference to the accompanying drawings. In order to facilitate understanding of the description, the same components in each drawing are denoted by the same reference numerals as much as possible, and redundant description will be omitted.

Note that in the following description, the x direction, y direction, and z direction are directions perpendicular to each other. The x and y directions are horizontal, and the z direction is vertical. The x direction is the conveyance direction of the conveyance path 3 of the conveyor 2. The y direction is the width direction of the conveyance path 3 of the conveyor 2. Furthermore, for convenience of explanation, the z-positive direction side may also be expressed as the upper side, and the z-negative direction side may also be expressed as the lower side.

With reference to FIGS. 1 to 3, a schematic configuration of a material determination device 1 as an example of a waste plastic sorting device according to an embodiment will be described. FIG. 1 is a perspective view showing a schematic configuration of a waste plastic material determination device 1 according to an embodiment. FIG. 2 is a side view of the waste plastic material determining apparatus 1 shown in FIG. 1. FIG. 3 is a plan view of the waste plastic material determining apparatus 1 shown in FIG. 1. Here, the waste plastic whose material quality is to be determined is black waste plastic, and a configuration in which two types of materials S1 and S2 (indicated by rectangular and triangular marks in Figures 1 to 3) are mixed is exemplified. explain. In the following, black waste plastic pieces made of two types of materials S1 and S2 may be collectively represented by the symbol S.

This black waste plastic material determination device 1 includes a vibrating feeder 8 as an example of a supply unit that sequentially supplies black waste plastic pieces S1 and S2, and a vibrating feeder 8 that conveys the black waste plastic pieces S1 and S2 supplied by the vibrating feeder 8. The main part includes a conveyor 2 as an example of a transport part. The crushed black waste plastic pieces S1 and S2 are supplied to the vibrating feeder 8 via, for example, a charging hopper. The vibrating feeder 8 feeds the black waste plastic pieces S1 and S2 to the conveyor 2 while preventing the black waste plastic pieces S1 and S2 from overlapping each other by vibrating the mounting surface on which the black waste plastic pieces S1 and S2 are placed. The conveyor 2 has a conveyance path 3 on its upper surface, and conveys the black waste plastic pieces S1 and S2 on the conveyance path 3 in a direction away from the vibrating feeder 8.

The material determination device 1 also includes a lighting 10 as an example of an irradiation unit that irradiates infrared rays to the black waste plastic pieces S1 and S2, and an example of a reflection spectrum detection unit that detects the reflection spectra from the black waste plastic pieces S1 and S2. It is equipped with a mid-infrared camera 4 as a main part, and a discrimination device 5 that identifies the material of the black waste plastic pieces S1 and S2 based on the reflection spectrum detected by the mid-infrared camera 4. The illumination 10 includes a lamp 10A (see FIG. 6) which is an infrared light source such as a halogen tungsten lamp, and irradiates infrared rays from the lamp 10A toward the black waste plastic pieces S1 and S2. Further, the lighting 10 is installed so that the reflected light from the black waste plastic pieces S1 and S2 enters the mid-infrared camera 4, and the lighting 10 is installed so that the reflected light from the black waste plastic pieces S1 and S2 enters the mid-infrared camera 4. ).

For example, as shown in FIG. 1, one mid-infrared camera 4 can measure the entire width of the conveyor 2, and can measure the entire width of the conveyor 2 by dividing it into multiple (for example, 318) areas along the width. Near-infrared reflected light from the waste plastic pieces S1 and S2 can be received, and the spectrum of the reflected light can be measured for each area. The mid-infrared camera 4 is, for example, a camera equipped with a spectrometer in the mid-infrared wavelength region of 3 μm or more. The mid-infrared camera 4 performs measurement at a scan frequency of 230 Hz, for example, and transmits 318 pieces of spectrum data to the discrimination device 5 for each scan. Based on the 318 spectrum data received from the mid-infrared camera 4, the discrimination device 5 outputs material determination results for each of the 318 regions to the injection control section 6, which will be described later.

Furthermore, the material determination device 1 is provided with an injection nozzle 7 that injects air laterally or diagonally in a direction intersecting the conveyance direction on the downstream side of the conveyor 2 in the conveyance direction. A plurality of injection nozzles 7 (for example, 318) are arranged in parallel in the width direction of the conveyor 2, and the operation of each nozzle is controlled by the injection control section 6. The injection control unit 6 injects or does not inject air from the injection nozzle 7 according to the material determination result received from the discriminator 5, so that air can be sprayed into a plurality of areas (for example, a collection hopper) divided by a partition plate 9, for example. etc.), the black waste plastic pieces S1 and S2 are sorted and dropped, and waste plastic of a desired material is collected. That is, in this embodiment, the injection control unit 6, the injection nozzle 7, and the partition plate 9 select a desired material from the waste plastic pieces flowing on the conveyance path 3 of the conveyor 2 based on the material determination result by the discrimination device 5. It functions as a collection device 12 (sorting section) that collects things.

The operation of the material determination device 1 will be explained. For example, when the crushed black waste plastic pieces S1 and S2 are supplied to the vibrating feeder 8 via a charging hopper, the vibrating feeder 8 applies vibration to the supplied black waste plastic pieces S1 and S2 and then weighs them. It is conveyed downstream and supplied to the conveyor 2 so that it does not become a problem.

The black waste plastic pieces S1 and S2 supplied to the conveyance path 3 on the upper surface of the conveyor 2 are conveyed in the conveyance direction on the positive Light is irradiated. The mid-infrared camera 4 receives the infrared light emitted from the illumination 10 and reflected by the black waste plastic pieces S1 and S2, and outputs the light reception result (data of the light reception spectrum) to the discrimination device 5.

The discrimination device 5 identifies the material of the black waste plastic pieces S1 and S2 based on the light reception result input from the mid-infrared camera 4. Note that details of the material determination method by the determination device 5 will be described later with reference to FIGS. 4 to 9. The discrimination device 5 outputs the material identification result to the injection control section 6.

The injection control unit 6 selects the injection nozzle 7 according to the material from among the plurality of injection nozzles 7 arranged, and transmits a control signal at a timing. The injection nozzle 7 that has received the control signal opens its nozzle opening and injects air. By injecting air from the injection nozzle 7 at an appropriate timing based on the discrimination result of the discrimination device 5, materials to be sorted and those not to be sorted can be separated and collected.

In the examples shown in FIGS. 2 and 3, the black waste plastic pieces S1 on the conveyor 2 receive air from the air injection nozzle 7 that has received the control signal, and are blown off and dropped into the collection devices 12 provided for each material. It will be collected. Further, the black waste plastic pieces S2 on the conveyor 2 do not receive air from the injection nozzle 7, so they are collected by a different collecting device 12 from the black waste plastic pieces S1. In this way, by spraying and stopping the spray nozzle 7, black waste plastic pieces of a plurality of materials can be sorted and collected by material.

With reference to FIGS. 4 to 9, a method for determining the material of waste plastic using the determining device 5 will be described. FIG. 4 is a functional block diagram of the discrimination device 5.

As shown in FIG. 4, the discrimination device 5 includes a preprocessing section 51, a first determination section 52, a second determination section 53, and a third determination section 54.

The preprocessing unit 51 performs preprocessing such as correcting and processing the reflection spectra of the black waste plastic pieces S1 and S2 detected by the mid-infrared camera 4. The preprocessing unit 51 corrects the detected reflection spectrum using, for example, a spectrum measured under bright reflected light conditions and a spectrum measured under dark conditions. "Dark conditions" means conditions that are relatively darker than the above-mentioned "bright conditions." Further, the preprocessing unit 51 performs processing to cut out a predetermined frequency range from the corrected reflection spectrum.

The first determination unit 52 determines whether the spectrum detected by the mid-infrared camera 4 belongs to the waste plastic pieces S1, S2 or the conveyance path 3 of the conveyor 2. The first determination unit 52 performs determination using a trained One Class SVM (Support Vector Machine).

One Class SVM is a type of SVM that is a type of machine learning classification algorithm. In SVM, the identification boundary is set so that the Euclidean distance between the support vectors of each class (the one closest to other classes in the learning data) is maximized. Furthermore, if the features are nonlinear, a kernel is used to map the data to the feature space. By appropriately selecting the kernel, it is possible to draw identification boundaries even in complex data arrangements.

In One Class SVM, a method called a kernel trick is used for one type of learning data to map the data to a feature space in a high-dimensional space. At this time, since the learning data is mapped to be located far from the origin, data that is not similar to the original learning data gathers near the origin. This property is used to distinguish between normal data (conveyor 2) and abnormal data (objects (waste plastic pieces S1, S2)).

By using a One Class SVM with excellent pattern recognition ability in the first determination unit 52, it is possible to determine with high accuracy whether the reflection spectrum is reflected by the waste plastic pieces S1 and S2 or reflected by the conveyance path 3 of the conveyor 2. can be identified. Note that a classification method of supervised machine learning other than One Class SVM may be applied to the first determination unit 52.

The second determining unit 53 extracts feature data Score1 and Score2 (feature amounts) from the spectrum determined by the first determining unit 52 to be a piece of waste plastic. The second determination unit 53 performs determination using a learned PLS (Partial Least Squares) method.

PLS is a type of regression algorithm for supervised machine learning, and performs regression analysis between only a small number of principal components calculated from explanatory variables and a target variable. In PLS, principal components are calculated so that their covariance with the target variable is large. In this embodiment, two types of feature data Score1 and Score2 are calculated using PLS based on the explanatory variables of the reflection spectra determined to have been reflected by the waste plastic pieces S1 and S2.

By using PLS in the second determination unit 53, it is possible to reduce the multidimensional explanatory variables of the reflection spectrum to a small number of feature quantities, so it is possible to extract appropriate feature data Score1 and Score2 that are easier to distinguish. . Note that a multivariate analysis method of machine learning other than PLS may be applied to the second determination unit 53.

The third determination unit 54 determines the material of the waste plastic pieces S1 and S2 corresponding to the spectra based on the two feature values Score1 and Score2 of the reflection spectrum extracted by the second determination unit 53. The third determination unit 54 makes a determination using a learned decision tree. A decision tree is a type of supervised learning classification algorithm. A decision tree is a tree structure that represents rules for classifying objective variables, and is frequently used in classification problems.

By using a decision tree in the third determination unit 54, the materials of the waste plastic pieces S1 and S2 can be determined with high accuracy from the two feature quantities Score1 and Score2 of the reflection spectrum. Note that a classification method of supervised learning of machine learning other than the decision tree may be applied to the third determination unit 54.

The discrimination device 5 is physically configured as a computer system including a CPU (Central Processing Unit), a main storage device such as a RAM (Random Access Memory) and a ROM (Read Only Memory), a communication module, an auxiliary storage device, and the like. can do. Each function of the discrimination device 5 shown in FIG. 4 operates various hardware under the control of the CPU by loading predetermined computer software ( selection program) into the CPU, RAM, etc. This is achieved by reading and writing. That is, by executing the sorting program according to the present embodiment on a computer, the discriminating device 5 functions as the preprocessing section 51, the first determining section 52, the second determining section 53, and the third determining section 54 in FIG. do.

The selection program of this embodiment is stored, for example, in a storage device included in a computer. Note that a part or all of the screening program may be transmitted via a transmission medium such as a communication line, and may be received and recorded (including installation) by a communication module included in a computer. In addition, the selection program may be configured such that part or all of it is stored in a portable storage medium such as a CD-ROM, DVD-ROM, or flash memory, and then recorded (including installation) in a computer. Good too.

The discrimination device 5 may be a circuit configured with an analog circuit, a digital circuit, or an analog/digital mixed circuit. Further, a control circuit for controlling each function of the discrimination device 5 may be provided. Each circuit may be implemented using an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or the like.

Similarly, the injection control unit 6 can also be physically configured as a computer system including a CPU, RAM and ROM, a communication module, an auxiliary storage device, etc., and predetermined computer software is loaded into the CPU, RAM, etc. The function is realized by

FIG. 5 is a flowchart of waste plastic material determination processing according to the embodiment. Each process of the flowchart shown in FIG. 5 is executed by the discrimination device 5.

In step S01, the preprocessing unit 51 acquires the spectrum S _org (n, w) by the mid-infrared camera 4. Here, n is the number of sensors (the number of spectrum detection areas divided in the width direction of the conveyor 2 by the mid-infrared camera 4), and when the number of sensors is 318, there are 0 to 317 corresponding to each detection area. An integer is used. w is the wavelength of the spectrum, and in this embodiment, a total of 131 wavelengths are set between 2700 (nm) and 5300 (nm) in 20 (nm) increments, and an integer from 0 to 130 corresponding to each wavelength is set. is used. That is, S _org (n, w) represents the numerical value of the intensity of the spectrum of wavelength w in the n-th spectrum detection area along the width direction of the conveyor 2.

In step S02, the preprocessing unit 51 corrects the spectrum S _org (n, w) acquired in step S01 to calculate a corrected spectrum S _cor (n, w). With this correction, the spectrum due to the influence of changes in the concentration of water vapor and carbon dioxide in the measurement space, the temperature of the black waste plastic pieces S1 and S2 to be measured, aging of the lighting 10 and mid-infrared camera 4, the position on the conveyor 2, etc. Can absorb differences in strength characteristics. The corrected spectrum S _cor (n, w) can be calculated using the following equation (1), for example.

ここで、Ｗ_ｒｅｆ（ｎ，ｗ）は、反射光が明るい条件で計測した第１の補正用スペクトルである。Ｄ_ｒｅｆ（ｎ，ｗ）は、反射光が上記の明るい条件よりも暗い条件で計測した第２の補正用スペクトルである。これらの補正用スペクトルＷ_ｒｅｆ（ｎ，ｗ）、Ｄ_ｒｅｆ（ｎ，ｗ）は、例えば、材質判別処理を実行する前に中赤外線カメラ４の校正を行うときに抽出できる。

Here, W _ref (n, w) is the first correction spectrum measured under conditions where the reflected light is bright. D _ref (n, w) is a second correction spectrum measured under a condition in which the reflected light is darker than the bright condition described above. These correction spectra W _ref (n, w) and D _ref (n, w) can be extracted, for example, when calibrating the mid-infrared camera 4 before executing the material discrimination process.

FIG. 6 is a diagram showing a method for extracting the correction spectra W _ref (n, w) and D _ref (n, w). As shown in FIG. 6, a calibration plate 11 for acquiring a correction spectrum is installed in the imaging area of the mid-infrared camera 4 on the conveyance path 3 of the conveyor 2, and the spectrum of reflected light by the mid-infrared camera 4 is By performing the detection, the correction spectra W _ref (n, w) and D _ref (n, w) can be obtained.

In the case of the first correction spectrum W _ref (n, w) measured under conditions where the reflected light is bright, a calibration plate 11 (aluminum, stainless steel, etc.) that reflects all wavelengths in the mid-infrared region was placed and the illumination 10 was turned on. In the state, data of all wavelengths (w = 0 (2700), 2 (2720), ..., 130 (5300)) are collected for all sensors (n = 0, 1, 2, ..., 317). get.

In the case of the second correction spectrum D _ref (n, w) measured under conditions where the reflected light is dark, a calibration plate 11 (made of aluminum, stainless steel, etc.) that reflects all wavelengths in the mid-infrared region was placed and the lighting 10 was turned off. state (or with the camera shutter closed), all wavelengths (w = 0 (2700), 2 (2720), ...) for all sensors (n = 0, 1, 2, ..., 317). , 130 (5300)).

For _example , as shown by the dotted _arrow in FIG. It is preferable that the device be installed so as to be movable between the position of the imaging area of the mid-infrared camera 4 and a standby position outside the imaging area of the mid-infrared camera 4 and the irradiation range of the illumination 10 on the top of the camera. In other words, it is preferable that the calibration plate 11 is fixable at a predetermined position within the field of view of the mid-infrared camera 4 and at a predetermined position outside the field of view, and is movable between the two predetermined positions. It is preferable that the calibration plate 11 is processed so that its main surface, which receives light from the illumination 10, has a large surface roughness. Thereby, it is possible to suppress the occurrence of halation due to reflected light.

Further, the conveyor 2 may be stopped when acquiring the correction spectrum. In this case, if the calibration plate 11 is not placed correctly in the imaging area of the mid-infrared camera 4 due to some malfunction in the operation of the calibration plate 11, the infrared rays on the conveyor path 3 of the conveyor 2 will be irradiated by the infrared rays of the illumination 10. The temperature of the parts will rise and there is a risk of burnout or ignition. Therefore, if the calibration plate 11 is not fixed within the field of view of the mid-infrared camera 4, it is preferable to provide an interlock to prevent the illumination 10 from emitting infrared rays.

As shown in FIG. 6, the illumination 10 includes a lamp 10A (such as a sheath heater, a carbon lamp, or a kanthal lamp) that is an infrared light source, and a reflection plate 10B that collects the heat of the lamp 10A. The lamp 10A is formed to extend along the width direction (y direction) of the conveyor 2, and is arranged so as to radiate infrared rays in all directions around the axis along the y axis. The reflector plate 10B is arranged on the opposite side of the conveyor path 3 of the conveyor 2 with respect to the lamp 10A, and is curved along the circumferential direction around the axis of the lamp 10A. can collect infrared rays emitted to the opposite side, reflect them, and send them to the conveyor 2 side. The reflecting plate 10B is made of, for example, aluminum, stainless steel, or an aluminum plated member.

Returning to FIG. 5, in step S03, the preprocessing unit 51 cuts out a characteristic wavelength region from the corrected spectrum S _cor (n, w). FIG. 7 is a diagram illustrating an example of processing for cutting out a characteristic wavelength region from a reflected wave spectrum. The horizontal axis in FIG. 7 indicates the wavelength (nm) of the spectrum, and the vertical axis indicates the intensity of the spectrum at each wavelength. FIG. 7 shows an example of spectra of each material: ABS, HIPS, PP, and PE. In the example of FIG. 7, spectra in the wavelength ranges of 3250 to 3750 (nm) and 4400 to 4600 (nm) are cut out.

Returning to FIG. 5, in step S04, the first determination unit 52 uses the spectrum corrected in step S02 and the characteristic wavelength region cut out in step S03 to determine whether each spectrum is on the belt of the conveyor 2. (conveyance path 3) or an object (waste plastic) on the conveyance path 3 is determined. In this embodiment, the first determination unit 52 performs determination using a trained One Class SVM.

In step S05, the second determination unit 53 extracts two types of feature data Score1 and Score2 from the spectrum determined to be an object (waste plastic) in step S4 using the learned PLS. FIG. 8 is a diagram showing an example of feature data extraction. The horizontal axis in FIG. 8 indicates the first feature data (Score1), and the vertical axis indicates the second feature data (Score2). FIG. 8 shows an example of extraction of the four types of materials, ABS, HIPS, PP, and PE illustrated in FIG. 7. As shown in FIG. 8, it can be seen that in the two-dimensional space based on the two feature data Score1 and Score2, the area plotted for each material can be divided. Note that the number of feature data may be other than two types.

Returning to FIG. 5, in step S06, the third determination unit 54 determines the material using the learned decision tree based on the two types of feature data Score1 and Score2 extracted in step S5. FIG. 9 is a diagram showing an example of material discrimination using a decision tree. In this embodiment, in order to ultimately identify four types of materials (PE, PP, ABS, HIPS), the decision tree has two levels of conditional branches as shown in FIG. In the first layer, a set of feature data Score1 and Score2 is divided into two groups G1 and G2 using a conditional branching function f1 (Score1, Score2). One group G1 is further divided into two groups G11 and G12 using a conditional branching function f2 (Score1, Score2) in the second layer. The other group G2 is further divided into two groups G21 and G22 using a conditional branching function f3 (Score1, Score2) in the second layer. As a result, the set of feature data Score1 and Score2 is classified into four groups G11, G12, G21, and G22, and the materials of each group are determined to be PE, PP, ABS, and HIPS, respectively.

As described above, the discrimination device 5 of the waste plastic material discrimination device 1 according to the present embodiment uses the spectrum of the reflected light of the light irradiated onto the conveyance path 3 of the conveyor 2 by the illumination 10, which is detected by the mid-infrared camera 4. The first determining unit 52 determines whether the plastic piece S belongs to the waste plastic piece S or the conveyance path 3, and two types of characteristic data Score1 and Score2 are generated from the spectrum determined to be the waste plastic piece S by the first determining unit 52. and a third determining unit 54 that determines the materials S1 and S2 of the waste plastic piece S based on the feature data Score1 and Score2 extracted by the second determining unit 53.

With this configuration, the input information of the reflection spectrum can be narrowed down to the spectrum of the black waste plastic piece S by object discrimination by the first determination unit 52, and dimensionality from spectrum information to feature data can be determined by feature extraction by the second determination unit 53. Through three stages of determination processing including compression and classification processing by the third determination unit 54 and data narrowing down, output information on the material of the waste plastic can be obtained. Therefore, the waste plastic material determination device 1 of the present embodiment can determine the material quality of waste plastic by considering various conditions, and can perform the determination in detail over multiple stages. The accuracy of plastic material determination can be improved.

Further, the discriminating device 5 of the waste plastic material discriminating device 1 according to the present embodiment converts the reflection spectrum S _org (n, w) detected by the mid-infrared camera 4 into the first one measured under conditions where the reflected light is bright. A preprocessing unit 51 is provided that performs correction using a correction spectrum W _ref (n, w) and a second correction spectrum D _ref (n, w) measured under relatively darker conditions than this bright condition. .

In this way, by correcting the reflection spectrum S _org (n, w) using the correction spectra W _ref (n, w) and D _ref (n, w) using, for example, equation (1), Differences in spectral intensity characteristics caused by the temperature of the black waste plastic pieces S1 and S2 to be measured, age-related deterioration of the mid-infrared camera 4, position on the conveyor 2, etc. can be suppressed. Therefore, by using the corrected spectrum S _cor (n, w) to perform learning and judgment in the first judgment section 52, second judgment section 53, and third judgment section 54, the material quality of the waste plastic can be determined. Judgment accuracy can be further improved.

The preprocessing unit 51 further performs processing to cut out a predetermined frequency range from the corrected spectrum S _cor (n, w), and outputs the processed spectrum to the first determination unit 52 .

With this configuration, parts that are strongly related to the material of waste plastic can be extracted from the spectrum and used for learning and determination by the first determining section 52, second determining section 53, and third determining section 54. It is possible to further improve the accuracy of determining the material of waste plastic by reducing the amount of noise that may impede the determination.

Note that in this embodiment, the preprocessing unit 51 performs two processes: a process of correcting the reflection spectrum S _org (n,w) and a process of cutting out a predetermined frequency range, but only one of the two processes is performed. It may also be configured to do so.

Further, in the present embodiment, the case where the waste plastic to be determined as a material is black waste plastic S has been described as an example, but waste plastic of other colors such as red or blue may be used, for example. Further, waste plastics of different colors may be used in combination.

With reference to FIGS. 10 to 14, a method of sorting waste plastic of a desired material using the material determination device 1 of this embodiment will be described. FIG. 10 is a plan view showing the first material selection method by the material determination device 1 of this embodiment. FIG. 10 shows a simplified diagram corresponding to the plan view of the material determination device 1 shown in FIG. 3. From FIG. 10 onward, as an example of the sorting method, an example will be described in which a plastic mix in which five types of materials (1), (2), (3), (4), and (5) are mixed is to be sorted.

In the example of FIG. 10, a configuration is illustrated in which the conveyor 2 and the collection device 12 form a single conveyance path that is not divided in the width direction (y direction) of the conveyor 2. In the collection device 12, the waste plastics are separated using the partition plate 9 shown in FIG. 2 as a boundary by spraying and stopping the spray nozzle 7, so that the objects to be sorted are roughly classified into two types. Therefore, in order to separate a plastic mix containing five types of materials to be sorted into individual materials, it is necessary to inject air from one of the collecting devices 12-1 (for example, from the injection nozzle 7). It is necessary to repeat the process of separating each type of waste plastic into a device that collects the waste plastic. That is, as shown in FIG. 10, first, only material (1) is classified from the plastic mix and collected by the collection device 12-1. At this time, the remaining plastic mix collected in the other collecting device 12-2 contains other four types of materials (2) to (5). Next, the remaining plastic mix containing the four types is fed into the material determining device 1 again and classified into any one of (2) to (5). By repeating this procedure four times, it is possible to separate each of the five types of materials (1) to (5).

FIG. 11 is a plan view showing a second material selection method by the material determination device 1A of this embodiment. From FIG. 11 onward, the conveyance path 3 of the conveyor 2 is divided into two systems, a first system and a second system, in the width direction. More specifically, each of the input port (vibrating feeder 8), conveyor 2, and collection device 12 is divided into two in the width direction. Note that the input port 8 and the conveyor 2 are not made up of two components, but a single component is provided with a partition or the like to prevent mixing between systems. For example, the conveyance path 3 of the conveyor 2 can be divided into two systems by providing a partition wall along the conveyance direction at a substantially central position in the width direction.

In the following description, the first system is represented by the subscript A, and the second system is represented by the subscript B. Also, elements corresponding to the collection device 12-1 in FIG. 10 are expressed as "collection device A1" and "collection device B1", and elements corresponding to the collection device 12-2 in FIG. 10 are expressed as "collection device A2" and " Collection device B2".

In the example of FIG. 11, waste plastic pieces made of the same material are collected in the first system and the second system. For example, as shown in Figure 11, a plastic mix in which five types of materials (1) to (5) are mixed is supplied to the input ports A and B of the first and second systems, and the material determination is made in each system. The waste plastic pieces of the same material (1) are collected by the collection devices A1 and B1. Further, the collection devices A2 and B2 collect the remaining waste plastics in which materials (2) to (5) are mixed.

FIG. 12 is a plan view showing a third material selection method by the material determination device 1B of this embodiment. In the example of FIG. 12, the first system collects the waste plastic pieces of the first material and supplies the remaining waste plastic pieces to the second system, and the second system collects the remaining waste plastic pieces. Collect waste plastic pieces of the second material from among them. In the example of FIG. 12, a plurality of types of mixed materials can be separated into three types: a first material, a second material, and other materials.

In the example of FIG. 12, a plastic mix containing a mixture of five types of materials (1) to (5) is supplied to the input port A of the first system, and the material determination is performed on the conveyor A of the first system. , waste plastic pieces of material (1) are collected by the collection device A1. In addition, the collection device A2 collects waste plastic in which the remaining materials (2) to (5) are mixed.

Next, the waste plastics collected by the collection device A2 and containing the remaining materials (2) to (5) are transported to the input port B of the second system by the transport device 13, and are supplied to the input port B. Ru. The material quality is determined by the conveyor B of the second system, and waste plastic pieces of material (2) are collected by the collection device B1. The collection device B2 collects the waste plastic in which the remaining materials (3) to (5) are mixed.

FIG. 13 is a plan view showing a fourth material selection method by the material determination device 1C of this embodiment. In the example of FIG. 13, the first system collects waste plastic pieces of the first material and a small amount of other materials, and supplies the collected waste plastic pieces to the second system. , the waste plastic pieces made of the first material are collected from among the waste plastic pieces made of the first material and a small amount of other materials. In the example shown in FIG. 13, plastic pieces made of a predetermined type of material can be sorted to a high purity.

In the example shown in FIG. 13, a plastic mix in which five types of materials (1) to (5) are mixed is supplied to the input port A of the first system, and the material determination is performed on the conveyor A of the first system. , the material (1) and trace amounts of waste plastic pieces (2) to (5) are collected by the collection device A1. In addition, the collection device A2 collects waste plastic in which the remaining materials (2) to (5) and a trace amount of (1) are mixed.

Next, the waste plastic collected by the collection device A1 and containing a mixture of material (1) and trace amounts of (2) to (5) is transported to the input port B of the second system by the conveyance device 13, and is transported to the input port B of the second system. supplied to B. The material quality is determined by the conveyor B of the second system, material (1) is sorted again by the collection device B1, and waste plastic pieces of material (1) are collected. The material (1) collected by this collection device B1 has a higher purity than the material collected by the collection device A1. The collection device B2 collects the remaining waste plastics containing materials (1) to (5).

FIG. 14 is a plan view showing the fifth material selection method by the material determination device 1D of this embodiment. In the example of FIG. 14, the first system excludes waste plastic pieces of the first material and a small amount of other materials, and supplies the remaining excluded waste plastic pieces to the second system. In the system, waste plastic pieces containing the first material and a small amount of other materials are further excluded from other waste plastic pieces, and waste plastic pieces that do not contain the first material are collected. In the example shown in FIG. 14, plastic pieces made of one type of predetermined material can be more reliably sorted out from mixed materials.

In the example shown in FIG. 14, a plastic mix containing five types of materials (1) to (5) is supplied to the input port A of the first system, and the material is determined by the conveyor A of the first system. , the material (1) and trace amounts of waste plastic pieces (2) to (5) are collected by the collection device A1. Further, the collection device A2 collects waste plastic in which the remaining materials (2) to (5) and a small amount of (1) are mixed.

Next, the waste plastics collected by the collection device A2, in which materials (2) to (5) and a small amount of (1) are mixed, are transported by the transport device 13 to the input port B of the second system, and are transported to the input port B of the second system. B is supplied. The material quality is determined by the conveyor B of the second system, and the material (1) is sorted again by the collection device B1, and waste plastic pieces containing a mixture of material (1) and trace amounts of materials (2) to (5) are collected. collected. The collection device B2 collects waste plastic in which the remaining materials (2) to (5) and a trace amount of (1) are mixed.

FIG. 15 is a diagram showing an example of the operation screen of the material determination device 1. The operation screen shown in FIG. 15 is displayed on a display device installed in the main body of the material determination device 1, for example. As shown in FIG. 15, the operation screen lists the names of the plastic materials to be sorted, and selects the first system ("primary" in FIG. 15) and the second system ("secondary" in FIG. 15). ) It is possible to individually select the material to be sprayed and sorted for each type.The display device on which the operation screen is displayed is, for example, a touch panel, and the material to be sprayed and sorted can be selected individually by pressing the "OFF" display in the "Spraying selection" field. By switching to the "ON" display, the injection nozzle 7 can be set to inject air and separate the collection device in the case of the material (ABS in FIG. 15). ” column to display the ratio of each material mixed in the material according to the determination result of the material determination process.

The present embodiment has been described above with reference to specific examples. However, the present disclosure is not limited to these specific examples. Design changes made by those skilled in the art as appropriate to these specific examples are also included within the scope of the present disclosure as long as they have the characteristics of the present disclosure. The elements included in each of the specific examples described above, their arrangement, conditions, shapes, etc. are not limited to those illustrated, and can be changed as appropriate. The elements included in each of the specific examples described above can be appropriately combined as long as no technical contradiction occurs.

1, 1A, 1B, 1C, 1D Waste plastic material determination device (waste plastic sorting device)
2 Conveyor 3 Conveyance path 4 Mid-infrared camera (reflection spectrum detection unit)
5 Discrimination device 51 Pre-processing section 52 First judgment section 53 Second judgment section 54 Third judgment section 12, 12-1, 12-2, A1, A2, B1, B2 Collection device (sorting section)
S1, S2 Black waste plastic pieces

Claims (13)

A waste plastic sorting device,
an irradiation unit that irradiates light onto waste plastic pieces being conveyed on the conveyance path;
a reflection spectrum detection unit that receives reflected light of the light irradiated by the irradiation unit and detects the spectrum of the reflected light;
The spectrum detected by the reflection spectrum detection unit is determined by using a first correction spectrum measured under conditions in which the reflected light is bright and a second correction spectrum measured under conditions darker than the bright conditions. a preprocessing unit for correction;
a first determination unit that determines whether the spectrum corrected by the preprocessing unit belongs to either the waste plastic piece or the conveyance path;
a second determination unit that extracts a feature amount from the spectrum determined to be the waste plastic piece by the first determination unit;
a third determination unit that determines the material of the waste plastic piece based on the feature amount extracted by the second determination unit;
By controlling the timing of injecting air to the waste plastic pieces transported on the conveyance path and sorting the waste plastic pieces into a plurality of areas and letting them fall, according to the material determination result by the third determination unit, a sorting unit that sorts and collects the waste plastic pieces according to the material;
Equipped with
The preprocessing unit corrects the spectrum using the following formula,

Here, S _org (n, w) is the spectrum detected by the reflection spectrum detection section, W _ref (n, w) is the first correction spectrum, and D _ref (n, w) is the second correction spectrum, S _cor (n, w) is the spectrum after correction,
Waste plastic sorting equipment. The preprocessing unit performs processing to cut out a predetermined frequency range from the corrected spectrum,
The first determination unit performs the determination using the spectrum processed by the preprocessing unit.
The waste plastic sorting device according to claim 1 . The first determination unit performs the determination using a trained One Class SVM.
The waste plastic sorting device according to claim 1 or 2 . The second determination unit makes the determination using the learned PLS,
The waste plastic sorting device according to any one of claims 1 to 3 . The third determination unit makes a determination using a learned decision tree.
The waste plastic sorting device according to any one of claims 1 to 4 . The sorting unit includes a collection device that collects pieces of one material from the waste plastic pieces flowing along the conveyance path based on the determination result of the third determination unit.
The waste plastic sorting device according to any one of claims 1 to 5 . The conveyance path is divided into two systems, a first system and a second system, in the width direction.
The waste plastic sorting device according to claim 6 . collecting waste plastic pieces of the same material in the first system and the second system;
The waste plastic sorting device according to claim 7 . In the first system, waste plastic pieces of a first material are collected, and the remaining waste plastic pieces are supplied to the second system,
In the second system, waste plastic pieces of a second material are collected from the remaining waste plastic pieces.
The waste plastic sorting device according to claim 7 . In the first system, waste plastic pieces of the first material and a small amount of other materials are collected, and the collected waste plastic pieces are supplied to the second system,
In the second system, waste plastic pieces made of a first material are collected from waste plastic pieces made of the first material and a small amount of other materials.
The waste plastic sorting device according to claim 7 . In the first system, waste plastic pieces of the first material and a small amount of other materials are excluded, and the remaining waste plastic pieces that are excluded are supplied to the second system,
In the second system, waste plastic pieces containing the first material and a small amount of other materials are further excluded from the other waste plastic pieces, and waste plastic pieces that do not contain the first material are collected. do,
The waste plastic sorting device according to claim 7 . A method for sorting waste plastic,
an irradiation step of irradiating light onto the waste plastic pieces being conveyed on the conveyance path;
a reflection spectrum detection step of receiving reflected light of the light irradiated in the irradiation step and detecting a spectrum of the reflected light;
The spectrum detected in the reflection spectrum detection step is obtained by using a first correction spectrum measured under conditions in which the reflected light is bright and a second correction spectrum measured under conditions darker than the bright conditions. a preprocessing step for correcting the
a first determination step of determining whether the spectrum corrected in the pre-processing step belongs to either the waste plastic piece or the conveyance path;
a second determination step of extracting a feature amount from the spectrum determined to be the waste plastic piece in the first determination step;
a third determination step of determining the material of the waste plastic piece based on the feature amount extracted in the second determination step;
By controlling the timing of injecting air to the waste plastic pieces conveyed on the conveyance path according to the material determination result of the third determination step, and sorting the waste plastic pieces into a plurality of areas and dropping them, a sorting step of sorting and collecting the waste plastic pieces according to the material;
including;
The preprocessing unit corrects the spectrum using the following formula,

Here, S _org (n, w) is the spectrum detected in the reflection spectrum detection step, W _ref (n, w) is the first correction spectrum, and D _ref (n , w) is the second correction spectrum, S _cor (n, w) is the spectrum after correction,
How to sort waste plastic. A waste plastic sorting program,
An irradiation function that irradiates light onto waste plastic pieces transported on the transport path,
a reflection spectrum detection function that receives reflected light of the light irradiated by the irradiation function and detects the spectrum of the reflected light;
The spectrum detected by the reflection spectrum detection function is obtained by using a first correction spectrum measured under conditions where the reflected light is bright and a second correction spectrum measured under conditions darker than the bright conditions. Pre-processing function to correct
a first determination function that determines whether the spectrum corrected by the preprocessing function is from the waste plastic piece or the conveyance path;
a second determination function that extracts a feature amount from the spectrum determined to be the waste plastic piece by the first determination function;
a third determination function that determines the material of the waste plastic piece based on the feature amount extracted by the second determination function;
By controlling the timing of injecting air to the waste plastic pieces conveyed on the conveyance path according to the material determination result by the third determination function, and sorting the waste plastic pieces into a plurality of areas and dropping them, a sorting function that sorts and collects the waste plastic pieces according to the material;
to be realized by a computer ,
The preprocessing function corrects the spectrum using the following formula,

Here, S _org (n, w) is the spectrum detected by the reflection spectrum detection function, W _ref (n, w) is the first correction spectrum, and D _ref (n, w) is the second correction spectrum, S _cor (n, w) is the spectrum after correction,
Waste plastic sorting program.

Images