Optical discrimination device and optical sorting device

Abstract

This optical discernment device includes: a sensor capable of imaging a subject (granular material 20) in a wavelength region of 200-2,500 nm; a plurality of lights that irradiate the subject and each have different irradiation wavelength regions; a lighting control unit capable of controlling the light emission mode of the lights; and a discernment unit for discerning the quality of the subject on the basis of imaging data acquired by the sensor. The lighting control unit sets a multivariate analysis-based light-quantity balance for the plurality of lights accordance with the discerned quality of the subject.

Description

Optical discrimination device and optical sorting device

[0001] The present disclosure relates to an optical discrimination device and an optical sorting device that are capable of discriminating and sorting the type, state, etc. of objects.

[0002] Various techniques have been disclosed for identifying and characterizing the materials of resins such as plastics. For example, Japanese Patent Laid-Open Publication No. 06-003260 discloses a technique that uses electromagnetic waves ranging from infrared rays to X-rays as a light source for a sorting device.

[0003] Specifically, a configuration has been disclosed in which infrared light in the wavelength range of 2 to 20 μm is used to identify five types of plastic plates (vinyl chloride, polypropylene, urethane, polyethylene, and polystyrene), and a configuration in which a XeCl excimer laser with a wavelength of 308 nm is used to identify the same five types of plastic plates.

[0004] Japanese Patent Application Laid-Open Publication No. 2016-014673 discloses a configuration in which a sorting device includes a plurality of illumination sources that output wavelengths in the range of approximately 400 nm to approximately 2000 nm as a light source. Japanese Patent Application Laid-Open Publication No. 2020-524328 discloses an illumination device that directs electromagnetic energy toward an object or directs light of a specific wavelength toward an object, and discloses that an imaging device can detect reflected light from at least one illumination device.

[0005] Furthermore, Japanese Patent Laid-Open Publication No. 2022-098183 discloses a lighting unit that produces illumination light with a color distribution suited to the type of defect in granular matter, enabling the use of illumination light in an appropriate frequency band for precise determination of granular matter, as a technology for selecting an appropriate wavelength range depending on the material to be sorted. Specifically, the publication discloses that the lighting unit is equipped with a multicolor light-emitting package incorporating multiple LED elements with different emission wavelengths, and that when detecting "immature brown rice," the lighting pattern is controlled to produce strong reddish illumination light, and when detecting "grains damaged by stink bugs," the lighting pattern is controlled to produce illumination light with strong red and green luminosity.

[0006] However, for example, the technology disclosed in JP 2022-098183 A involves selecting illumination light of an appropriate wavelength range using a predetermined lighting pattern of a multi-color light-emitting package based on the opinion of a pre-designated expert using a manual operating tool, and the control is based on an unscientific method that relies on the experience and intuition of the expert.

[0007] In view of the above problems, the present disclosure aims to provide an optical discrimination device and an optical sorting device that can perform appropriate illumination light emission control based on mathematical analysis results, without relying on the experience and intuition of an expert, and that can accurately determine the quality of objects and further sort them.

[0008] According to a first aspect of the present disclosure, there is provided an optical discrimination device including a sensor capable of capturing an image of an object in a wavelength range of 200 to 2500 nm, a plurality of illuminators that irradiate the object with light and each illuminator has a different wavelength range, an illumination control unit that can control the light emission mode of the illuminators, and a discrimination unit that discriminates the quality of the object based on image data acquired by the sensor, wherein the illumination control unit sets a light intensity balance for the plurality of illuminators based on multivariate analysis in accordance with the quality of the object to be discriminated.

[0009] According to a second aspect of the present disclosure, in the first aspect, the optical discrimination device has a light intensity balance generation unit that generates the light intensity balance, and the light intensity balance generation unit is capable of acquiring high-dimensional sample data of object samples of different qualities in advance under multiple types of light intensity balances due to the illumination, and reducing the dimension of the high-dimensional sample data by performing multivariate analysis on the high-dimensional sample data to generate the light intensity balance.

[0010] According to a third aspect of the present disclosure, in the first or second aspect, the plurality of illuminators include a red light source, a green light source, a blue light source, and a near-infrared light source.

[0011] According to a fourth aspect of the present disclosure, in any one of the first to third aspects, the plurality of illuminators include a plurality of near-infrared light sources having different wavelengths.

[0012] According to a fifth aspect of the present disclosure, there is provided an optical sorting device including the optical discrimination device of any one of the first to fourth aspects, and an ejector that selects and removes objects based on a discrimination result of the objects by the discrimination unit.

[0013] According to the present disclosure, it is possible to appropriately control the emission of illumination light based on mathematical analysis techniques, without relying on the experience and intuition of skilled workers as in the past, thereby making it possible to accurately identify and even select objects.

[0014] FIG. 1 is a schematic diagram of an inspection unit of an optical discrimination device and an optical sorting device according to an embodiment of the present disclosure. FIG. 2 is a schematic block diagram illustrating a control configuration of an optical discrimination device and an optical sorting device according to an embodiment of the present disclosure. FIG. 3 is a schematic diagram illustrating a case where a light source having five wavelength regions is used for illumination according to an embodiment of the present disclosure. FIG. 4 is a diagram illustrating an aspect of multivariate analysis (principal component analysis) according to an embodiment of the present disclosure. FIG. 5 is a diagram illustrating a method for generating a balance of illumination light intensity by multivariate analysis (principal component analysis) according to an embodiment of the present disclosure. FIG. 6 is a diagram illustrating a method for reducing the dimension of data obtained under high-dimensional illumination conditions by multivariate analysis (principal component analysis) according to an embodiment of the present disclosure.

[0015] An example of an embodiment of the present disclosure will now be described with reference to the drawings. As shown in the schematic diagram of Fig. 1, an optical discrimination device 1 and an optical sorting device 2 in this embodiment are intended to treat, for example, granular objects (granular objects 20) as targets, and have an inspection unit that optically inspects the quality of the granular objects 20.

[0016] As shown in the figure, the inspection unit has at least a sensor 10 capable of capturing an image of a granular object 20 in the wavelength range of 200 to 2500 nm, and a plurality of lights 11 that irradiate the granular object 20 with light and each irradiate with a different wavelength range.

[0017] 2, the system includes at least an illumination control unit 31 capable of controlling the light emission mode of the illumination 11, and a discrimination unit 30 that discriminates the quality of the granular matter 20 based on the imaging data acquired by the sensor 10. The discrimination unit 30 discriminates the quality of the granular matter 20, for example, by comparing the gradation value of the imaging data acquired by the sensor 10 with a threshold value. The illumination control unit 31 can set the light intensity balance of the plurality of illumination units 11 based on multivariate analysis in accordance with the quality of the granular matter 20 to be discriminated.

[0018] More specifically, the sensor 10 of this embodiment can use a wideband image sensor, such as a wideband image sensor (wideband monosensor) that has spectral sensitivity from visible light to near-infrared light. By using such a wideband monosensor, it is no longer necessary to install multiple sensors corresponding to multiple illuminations 11 with different wavelength ranges, thereby reducing costs.

[0019] In addition, the sensor 10 is not necessarily limited to a wideband mono sensor, and the sensor 10 can be configured to be equipped with multiple image sensors of various types capable of capturing images in a specified wavelength range, and to be capable of capturing images of the granular matter 20 in the wavelength range of 200 to 2500 nm described above.

[0020] As described above, this embodiment has a plurality of illuminators 11 that emit light in different wavelength ranges. Specifically, the illuminators 11 include at least a red light source (R), a green light source (G), a blue light source (B), and a near-infrared light source (NIR). The light intensity balance of each light source is set by the illumination control unit 31. In other words, the light intensity of each light source is set individually. Note that the types and numbers of light sources described above are merely examples, and two or more of any light sources can be arranged.

[0021] The light quantity balance described above is generated by the light quantity balance generation unit 32. Specifically, the light quantity balance generation unit 32 acquires high-dimensional sample data of granular material samples of different qualities in advance under a plurality of types of light quantity balances by the illumination 11, and performs multivariate analysis on the high-dimensional sample data to reduce the dimension thereof and generate the light quantity balance.

[0022] Regarding the multivariate analysis in the light quantity balance generation unit 32, the following will explain the case where the light source (lights 11A to 11E) having the five wavelength ranges shown in FIG. 3 is used, taking principal component analysis as an example.

[0023] First, as a pre-processing step, granular samples of different qualities are prepared in advance, and high-dimensional sample data for discriminating the quality of the granular objects are obtained under high-dimensional illumination conditions using the five types of light sources. １ ~X ５ Under certain high-dimensional illumination conditions, images of particulate matter samples with different qualities are captured to obtain high-dimensional sample data.

[0024] Next, principal component analysis is performed on the high-dimensional sample data. Figure 5 shows an example of the contribution rate of each principal component and the basic formula for principal component analysis. As shown in the figure, the cumulative contribution rate up to the third principal component is approximately 90%, so it is possible to reduce the dimension (dimensionality) from five dimensions to three or less principal components.

[0025] Then, based on the principal component coefficients obtained by the principal component analysis, the light intensity balance of each of the lighting devices 11A to 11E (variable X １ ~X ５ ) and at the same time obtain image information of the granular object 20 to be discriminated, the image information can be obtained by reducing the dimensions to a one-dimensional space represented by the first principal component, or a two-dimensional space represented by the first and second principal components, or a three-dimensional space represented by the first, second and third principal components, as shown in FIG.

[0026] By using the above-described processing to specifically control the balance of the light intensity of each of the illuminators 11A-11E (for example, by controlling only illumination pattern 1, or illumination patterns 1 and 2, or illumination patterns 1, 2, and 3 in FIG. 6), it is possible to obtain quality discrimination performance equivalent to that under high-dimensional illumination conditions, such as the reflected and transmitted light, reflected light, and transmitted light of a red light source (R), a green light source (G), and a blue light source (B), and the reflected and transmitted light, reflected light, and transmitted light of multiple near-infrared light sources (NIR). In other words, by balancing the light intensity of each of the illuminators 11A-11E using multivariate analysis, it is possible to accurately discriminate the quality of the granular material 20 (for example, non-defective, defective, foreign matter, etc.) under low-dimensional illumination conditions.

[0027] The above-described processing allows the imaging information of the granules 20 to be represented in a space of three or less dimensions, making it possible to create a three-dimensional optical correlation diagram of each granule 20 on a display and plot and display the imaging information. This allows the operator to easily input and edit the discrimination threshold for the granules 20.

[0028] (Optical sorting device) The above describes an example of the discrimination processing configuration of the optical discrimination device 1 in this embodiment, but the optical discrimination device 1 can be provided in a known sorting device for granular materials, etc., to form an optical sorting device 2.

[0029] The optical sorting device 2 can be equipped with, for example, an inclined chute within the device body that serves as a means for transporting granular material 20, an inspection unit that performs optical inspection on the granular material 20 falling from the chute and is equipped with at least the aforementioned sensor 10 and lighting 11, and an ejector that is provided below the inspection unit and can, for example, blow away defective granular material 20 with a blast of air.

[0030] 1, the controller 3 of the optical sorting device 2 is provided with an ejector control unit 33, which controls the operation of the ejector 41 via an ejector drive circuit 40. That is, based on the discrimination result of the granular material 20 by the discrimination unit 30 described above, the granular material 20 can be sorted and removed by the ejector 41. The controller 3 may include a processor and a memory, and realize the functions of the discrimination unit 30, the illumination control unit 31, the light quantity balance generation unit 32, and the ejector control unit 33 by executing a program stored in the memory.

[0031] Other Embodiments The above describes an embodiment of the optical discrimination device and optical sorting device of the present disclosure. However, the present disclosure is not necessarily limited to the above-described embodiment and includes, for example, the following modifications.

[0032] For example, rice can be used as the object to be discriminated. That is, from the rice to be discriminated, it is possible to distinguish and sort out good quality white rice from defective colored rice, lightly burnt rice, and shirataka rice.

[0033] Furthermore, the objects to be discriminated are not limited to the rice mentioned above, but may include grains such as wheat, beans, nuts, etc., as well as small articles such as pellets, beads, and other resin pieces, pharmaceuticals, ores, whitebait, etc. The optical discrimination device and optical sorting device disclosed herein can be effectively applied to cases where raw materials to be discriminated are sorted into good and bad products, or where foreign matter mixed in the raw materials is removed.

[0034] In the above-described embodiment, an embodiment using principal component analysis has been described as an example of multivariate analysis, but the present invention is not necessarily limited to such an analysis method, and other methods such as linear discriminant analysis (LDA) that can be used as a classification model can be used as long as they enable dimensionality reduction.

[0035] Although the embodiments and some modifications of the present disclosure have been described above, these descriptions are intended to facilitate understanding of the present teachings and are not intended to limit the present invention. The present invention may be modified or improved without departing from the spirit thereof, and the present invention includes equivalents thereof. Furthermore, the components described in the claims and specification may be combined or omitted to the extent that at least some of the above-mentioned problems can be solved or at least some of the effects can be achieved.

[0036] REFERENCE SIGNS LIST 1 Optical discrimination device 2 Optical sorting device 3 Controller 10 Sensor 11 (11A, 11B, 11C, 11D, 11E) Illumination 20 Granular matter 30 Discrimination unit 31 Illumination control unit 32 Light quantity balance generation unit 33 Ejector control unit 40 Ejector drive circuit 41 Ejector

Claims

1. An optical discrimination device comprising: a sensor capable of capturing an image of an object in the wavelength range of 200 to 2500 nm; a plurality of lights that irradiate the object with light and each light has a different wavelength range; a lighting control unit capable of controlling the light emission mode of the lights; and a discrimination unit that discriminates the quality of the object based on the imaging data acquired by the sensor, wherein the lighting control unit sets a light intensity balance for the plurality of lights based on multivariate analysis in accordance with the quality of the object to be discriminated.

2. An optical discrimination device as described in claim 1, further comprising a light quantity balance generation unit that generates the light quantity balance, the light quantity balance generation unit being capable of acquiring high-dimensional sample data of object samples of different qualities in advance under a plurality of types of light quantity balance due to the illumination, and performing multivariate analysis on the high-dimensional sample data to reduce its dimension and generate the light quantity balance.

3. The optical discrimination device according to claim 1 or 2, wherein the plurality of illuminations include a red light source, a green light source, a blue light source, and a near-infrared light source.

4. An optical discrimination device according to any one of claims 1 to 3, wherein the plurality of illuminations include a plurality of near-infrared light sources having different wavelengths.

5. An optical sorting device comprising the optical discrimination device according to any one of claims 1 to 4, and further comprising an ejector for selectively removing objects based on the discrimination results of the objects by the discrimination unit.

Images