Inspection device and inspection method

Abstract

An inspection device 1 comprises: an optical element 4 that has a property that the transmittance and reflectance thereof monotonically change in a predetermined wavelength region; a light reception unit 5 that receives transmitted light beams Lt transmitted through the optical element 4 and receives reflected light beams Lr reflected by the optical element 4, among light beams Lb from an inspection target object S, and that outputs luminance data of the transmitted light beams Lt and luminance data of the reflected light beams Lr; an image generation unit 6 that generates gravity-center wavelength image data GC of the inspection target object S with respect to the light beams Lb from the inspection target object S, on the basis of the luminance data of the transmitted light beams Lt and the luminance data of the reflected light beams Lr; and an identification unit 7 that identifies the material of a foreign matter W in the inspection target object S, on the basis of the gravity-center wavelength image data GC.

Description

Inspection device and inspection method

[0001] The present disclosure relates to an inspection device and an inspection method.

[0002] Conventionally, there are known inspection devices that inspect an object by irradiating the object with light. For example, a determination device described in Patent Document 1 is an example of such an inspection device. This conventional determination device receives reflected light when infrared light is irradiated onto the object. Then, correlation information is obtained between the spectral intensity of the reflection spectrum or absorption spectrum of the reflected light, excluding the wavelength band of 4.18 μm or more and 4.42 μm or less, and previously acquired spectral data of one or more resin types. The device then determines the resin type of the object based on the highest correlation information that is equal to or greater than a predetermined threshold.

[0003] Japanese Patent Application Laid-Open No. 2018-100903

[0004] The inspection device described above can be applied to, for example, inspecting an object for foreign matter. Foreign matter in an object can be made of various materials, including transparent objects. When inspecting for foreign matter, it is useful to identify the material of the detected foreign matter. However, it may be difficult to accurately identify the material of the foreign matter by simply observing the transmitted light that has passed through the object or the reflected light that has reflected from the object.

[0005] The gist of the present disclosure is as follows.

[0006] [1] An inspection device for identifying the material of foreign matter in an object, comprising: an optical element having a characteristic in which transmittance and reflectance change monotonically in a predetermined wavelength region; a light receiving unit that receives transmitted light that has passed through the optical element and reflected light that has been reflected by the optical element out of light from the object, and outputs luminance data of the transmitted light and luminance data of the reflected light; an image generating unit that generates centroid wavelength image data of the object for the light from the object based on the luminance data of the transmitted light and the luminance data of the reflected light; and an identification unit that identifies the material of foreign matter in the object based on the centroid wavelength image data.

[0007] This inspection device uses an optical element whose transmittance and reflectance change monotonically over a predetermined wavelength range, and generates centroid wavelength image data of the object for light from the object based on the luminance data of the transmitted light that passes through the optical element and the reflected light that is reflected by the optical element. Because the transmittance and reflectance of an optical element monotonically increase or decrease depending on the wavelength, the centroid wavelength image data generated from the luminance data of the transmitted light and the luminance data of the reflected light reflects information about the wavelength shift from the centroid wavelength of the light irradiated onto the object, depending on the material of the foreign object. Therefore, by using the centroid wavelength image data, the material of the foreign object in the object can be accurately identified.

[0008] [2] The inspection device according to [1], wherein the optical element has a characteristic in which the transmittance and reflectance change monotonically in each of a plurality of wavelength ranges different from one another. In this case, it is possible to increase the gradient of the change in the transmittance and reflectance for each wavelength range in the optical element. This improves the wavelength resolution of the centroid wavelength and increases the amount of fluctuation in the centroid wavelength image data in response to a wavelength shift depending on the material of the foreign matter. Therefore, the material of the foreign matter in the target object can be identified with greater accuracy.

[0009] [3] The inspection device according to [1] or [2], wherein the image generation unit generates luminance image data of the object for light from the object based on the luminance data of the transmitted light and the luminance data of the reflected light, and the identification unit identifies the material of a foreign substance in the object based on the centroid wavelength image data and the luminance image data. In this case, by combining the centroid wavelength image data and the luminance image data, the material of a foreign substance in the object can be identified with even greater accuracy.

[0010] [4] The inspection device according to [3], wherein the identification unit identifies the material of a foreign particle in the object based on a correlation between the centroid wavelength image data and the luminance image data. The centroid wavelength of light from the object differs depending on its specific absorption spectrum and for each material. In this case, by using the correlation between the centroid wavelength image data and the luminance image data, the material of the foreign particle in the object can be identified quickly and accurately without using complex techniques such as multivariate analysis.

[0011] [5] The inspection device according to [3] or [4], wherein the image generation unit generates ratio image data based on a ratio between the luminance image data in two different wavelength regions, and the identification unit identifies the material of a foreign particle in the object based on the ratio image data and the centroid wavelength image data in each of the wavelength regions used to generate the ratio image data. In this case, using ratio image data based on the ratio between the luminance image data in two different wavelength regions makes it easier to grasp the trend of the wavelength shift amount in a wide wavelength region. By combining the ratio image data with the centroid wavelength image data in each of the wavelength regions used to generate the ratio image data, the material of a foreign particle in the object can be identified with greater accuracy.

[0012] [6] The inspection device according to any one of [1] to [5], wherein the identification unit identifies whether the material of the foreign matter in the object is transparent based on a comparison between the centroid wavelength image data and a preset threshold. Transparent objects tend to have a bias between the brightness of transmitted light that passes through an optical element and the brightness of reflected light that is reflected by the optical element due to light interference. Therefore, by applying a threshold to the centroid wavelength image data to detect the bias between the brightness of transmitted light and the brightness of reflected light, it is possible to identify with high accuracy whether the material of the foreign matter in the object is transparent.

[0013] [7] The inspection device according to any one of [1] to [6], further comprising a light source unit that outputs inspection light by switching wavelengths over time toward the object, thereby enabling centroid wavelength image data to be easily acquired in a plurality of wavelength regions.

[0014] [8] The inspection device according to any one of [1] to [7], wherein the light receiving unit has a first sensor that receives the transmitted light that has passed through the optical element, and a second sensor that receives the reflected light that has been reflected by the optical element. In this case, the transmitted light and the reflected light can be simply received.

[0015] [9] The inspection device according to any one of [1] to [7], wherein the light receiving unit has a first sensor that receives one of the transmitted light that has passed through the optical element and the reflected light that has reflected from the optical element, and a second sensor that receives light from the object without passing through the optical element. In this case, the transmitted light and the reflected light can be simply received.

[0016]

[10] The inspection device according to any one of [1] to [7], wherein the light receiving unit is configured by a single line sensor including a first light receiving area that receives the transmitted light that has passed through the optical element using a plurality of pixels arranged in one direction, and a second light receiving area that receives light from the object using a plurality of pixels arranged in the same direction without passing through the optical element. With this configuration, transmitted light and reflected light can be received with a smaller system.

[0017]

[11] The inspection device according to any one of [1] to [7], wherein the light receiving unit is configured with a plurality of line sensors including a first line sensor that receives the transmitted light that has passed through the optical element using a plurality of pixels arranged in one direction, and a second line sensor that receives light from the object using a plurality of pixels arranged in the same direction without passing through the optical element. With this configuration, transmitted light and reflected light can be received with a smaller system.

[0018]

[12] An inspection method for identifying the material of a foreign substance in an object, comprising: a light guiding step of guiding light from the object to an optical element having properties such that transmittance and reflectance change monotonically in a predetermined wavelength range; a light receiving step of receiving, of the light from the object, transmitted light that has passed through the optical element and reflected light that has been reflected by the optical element, and outputting luminance data of the transmitted light and luminance data of the reflected light; a generation step of generating centroid wavelength image data of the object for the light from the object based on the luminance data of the transmitted light and the luminance data of the reflected light; and an identification step of identifying the material of a foreign substance in the object based on the centroid wavelength image data.

[0019] This inspection method uses an optical element whose transmittance and reflectance change monotonically over a predetermined wavelength range, and generates centroid wavelength image data of the object for light from the object based on the luminance data of the transmitted light that passes through the optical element and the reflected light that is reflected by the optical element. Because the transmittance and reflectance of an optical element monotonically increase or decrease depending on the wavelength, the centroid wavelength image data generated from the luminance data of the transmitted light and the luminance data of the reflected light reflects information about the wavelength shift from the centroid wavelength of the light irradiated onto the object, depending on the material of the foreign object. Therefore, by using the centroid wavelength image data, the material of the foreign object in the object can be accurately identified.

[0020]

[13] The inspection method according to

[12] , wherein the light guiding step guides the light from the object to the optical element having a characteristic in which the transmittance and reflectance change monotonically in each of a plurality of different wavelength regions. In this case, it is possible to increase the gradient of the change in transmittance and reflectance for each wavelength region in the optical element. This improves the wavelength resolution of the centroid wavelength and increases the amount of variation in the centroid wavelength image data with respect to a wavelength shift depending on the material of the foreign object. Therefore, the material of the foreign object in the object can be identified with greater accuracy.

[0021]

[14] The inspection method according to

[12] or

[13] , wherein the generating step generates luminance image data of the object for light from the object based on the luminance data of the transmitted light and the luminance data of the reflected light, and the identifying step identifies the material of a foreign object in the object based on the centroid wavelength image data and the luminance image data. In this case, by combining the centroid wavelength image data and the luminance image data, the material of a foreign object in the object can be identified with even greater accuracy.

[0022]

[15] The inspection method according to

[14] , wherein the identifying step identifies the material of a foreign particle in the object based on a correlation between the centroid wavelength image data and the luminance image data. The centroid wavelength of light from the object varies depending on its specific absorption spectrum and for each material. In this case, by using the correlation between the centroid wavelength image data and the luminance image data, the material of the foreign particle in the object can be identified quickly and accurately without using a complex method such as multivariate analysis.

[0023]

[16] The inspection method according to

[14] or

[15] , wherein the generating step generates ratio image data based on a ratio between the luminance image data in two different wavelength regions, and the identifying step identifies the material of the foreign matter in the object based on the ratio image data and the centroid wavelength images in each of the wavelength regions used to generate the ratio image data. In this case, using ratio image data based on the ratio between the luminance image data in two different wavelength regions makes it easier to grasp the trend of the wavelength shift amount in a wide wavelength region. By combining the ratio image data with the centroid wavelength image data in each of the wavelength regions used to generate the ratio image data, the material of the foreign matter in the object can be identified with greater accuracy.

[0024]

[17] The inspection method according to any one of

[12] to

[16] , wherein the identification step identifies whether the material of the foreign matter in the object is transparent based on a comparison between the centroid wavelength image data and a preset threshold. Transparent objects tend to have a bias between the brightness of transmitted light that passes through an optical element and the brightness of reflected light that is reflected by the optical element due to light interference. Therefore, by applying a threshold to the centroid wavelength image data to detect the bias between the brightness of transmitted light and the brightness of reflected light, it is possible to identify with high accuracy whether the material of the foreign matter in the object is transparent.

[0025]

[18] The inspection method according to any one of

[12] to

[16] , further comprising an output step of outputting inspection light by switching wavelengths over time toward the object, thereby enabling centroid wavelength image data to be easily acquired in a plurality of wavelength regions.

[0026]

[19] The inspection method according to any one of

[12] to

[18] , wherein the light receiving step uses a first sensor that receives the transmitted light that has passed through the optical element and a second sensor that receives the reflected light that has been reflected by the optical element. In this case, the transmitted light and the reflected light can be simply received.

[0027]

[20] The inspection method according to any one of

[12] to

[18] , wherein the light receiving step uses a first sensor that receives one of the transmitted light that has passed through the optical element and the reflected light that has reflected from the optical element, and a second sensor that receives light from the object without passing through the optical element. In this case, the transmitted light and the reflected light can be simply received.

[0028]

[21] The inspection method according to any one of

[12] to

[18] , wherein the light receiving step uses a single line sensor including a first light receiving area that receives the transmitted light that has passed through the optical element using a plurality of pixels arranged in one direction, and a second light receiving area that receives light from the object using a plurality of pixels arranged in the one direction without passing through the optical element. With this configuration, transmitted light and reflected light can be received with a smaller system.

[0029]

[22] The inspection method according to any one of

[12] to

[18] , wherein the light receiving step uses a plurality of line sensors including a first line sensor that receives the transmitted light that has passed through the optical element using a plurality of pixels arranged in one direction, and a second line sensor that receives light from the object using a plurality of pixels arranged in the same direction without passing through the optical element. With this configuration, transmitted light and reflected light can be received with a smaller system.

[0030] According to the present disclosure, the material of a foreign substance in an object can be identified with high accuracy.

[0031] 1 is a schematic diagram showing the configuration of an inspection device according to an embodiment of the present disclosure; FIG. 2 is a schematic diagram showing the relationship between alternating lighting by a light source unit and the conveying direction of an object; FIG. 3 is a diagram showing an example of the relationship between inspection light output from a light source unit and the transmission characteristics of an optical element; FIG. 4 is a schematic diagram showing the configuration of a light receiving unit; FIG. 5 is a diagram showing an example of luminance image data, and FIG. 6 is a diagram showing an example of centroid wavelength image data; FIG. 7 is a diagram showing an example of correlation between centroid wavelength image data and luminance image data for each material of a foreign substance; FIG. 7 is data in which the material is predicted for each pixel from image data of a food acquired by a multispectral camera, and FIG. 8 is a correlation diagram obtained by applying dimensionality compression technology to the image data used for predicting the material in FIG. 8; FIG. 9 is data in which the material is predicted for each pixel from image data of a food acquired by the inspection device of this embodiment, and FIG. 9 is a correlation diagram obtained by applying dimensionality compression technology to the image data used for predicting the material in FIG. 9; 1 is a diagram showing an example of a material whose light absorptivity varies depending on the wavelength region; FIG. 2 is a diagram showing an example of ratio image data, and FIG. 2 is a diagram showing an example of centroid wavelength image data of a wavelength region used to generate non-image data; FIG. 3 is a flowchart showing details of a generation step and an identification step according to a modified example; FIG. 4 is a diagram showing an example of centroid wavelength image data including a non-transparent body and a transparent body, and FIG. 5 is a diagram showing the relationship between the spectrum of a non-transparent body and the transmission characteristics of an optical element; and FIG. 6 is a flowchart showing details of a generation step and an identification step according to a modified example; and FIG. 7 is a schematic diagram showing the configuration of a light receiving unit according to a modified example; and FIG. 8 is a schematic diagram showing another example of a light receiving unit according to a modified example.

[0032] Hereinafter, preferred embodiments of an inspection device and an inspection method according to one aspect of the present disclosure will be described in detail with reference to the drawings.

[0033] FIG. 1 is a schematic diagram illustrating the configuration of an inspection apparatus according to an embodiment of the present disclosure. The inspection apparatus 1 illustrated in FIG. 1 is configured to inspect an object S for foreign matter W by irradiating the object with inspection light La. As described below, the inspection apparatus 1 uses an optical element 4 having a transmission characteristic in which the transmittance changes linearly over a predetermined wavelength range. The inspection apparatus 1 receives, of light Lb from the object S generated by the inspection light La, transmitted light Lt that passes through the optical element 4 and reflected light Lr that is reflected by the optical element 4. Then, centroid wavelength image data GC of the object S is generated based on the luminance data of the transmitted light Lt and the luminance data of the reflected light Lr, and the material of the foreign matter W in the object S is identified based on the generated centroid wavelength image data GC. Note that the term "centroid wavelength" in this embodiment may also be referred to as "center wavelength."

[0034] The object S may be, for example, food. Examples of food include beef, pork, chicken, lamb, and processed foods thereof. The object S is not limited to food, but may be other items such as electronic components. Examples of foreign matter W in the object S include pieces of resin made of various materials. Furthermore, the foreign matter W is not limited to objects that are opaque to visible light, but may also include objects that are transparent to visible light, such as pieces of glass, pieces of film, and pieces of vinyl.

[0035] As shown in FIG. 1, the inspection device 1 includes a transport unit 2, a light source unit 3, an optical element 4 (see FIG. 4), a light receiving unit 5, an image generating unit 6, an identification unit 7, and a display unit 8. The transport unit 2 transports the object S in a predetermined direction. The transport unit 2 is configured, for example, by a belt conveyor. The transport unit 2 transports the object S in the horizontal direction at a constant speed toward the irradiation position of the inspection light La emitted by the light source unit 3. This causes the object S to be scanned with respect to the irradiation position of the inspection light La.

[0036] The light source unit 3 is a part that outputs inspection light La toward the target S. In this embodiment, the light source unit 3 is configured by a light source that can output inspection light La toward the target S by switching the wavelength over time. The inspection light La is, for example, a directional light beam. An example of a light source that outputs such light is a multiband LED bar illumination device. This illumination device incorporates, for example, an LED bar with a center wavelength λ1 of 1200 nm, an LED bar with a center wavelength λ2 of 1300 nm, an LED bar with a center wavelength λ3 of 1450 nm, an LED bar with a center wavelength λ4 of 1550 nm, and an LED bar with a center wavelength λ5 of 1650 nm.

[0037] The light source unit 3 outputs inspection light La of one of the wavelengths while switching between them at high speed toward the object S. The light source unit 3 alternately lights up corresponding to each line of pixels of the light receiving unit 5, so that the object S, which is being transported at a constant speed by the transport unit 2, is repeatedly irradiated in order with inspection light La having center wavelengths in each wavelength range from λ1 to λ5, as shown in Fig. 2. The light source unit 3 may be configured with a variable wavelength light source or a white light source equipped with multiple bandpass filters.

[0038] The optical element 4 is an element having a transmission characteristic in which the transmittance changes linearly in a predetermined wavelength range. Here, the optical element 4 is configured as a light dividing element (beam splitter) called a linear reflectance gradient (LRG) dichroic mirror. The dichroic mirror is configured, for example, of a dielectric multilayer film, and transmits and reflects light Lb from the object S according to its wavelength, thereby dividing the light Lb into transmitted light Lt and reflected light Lr.

[0039] FIG. 3 is a diagram showing an example of the relationship between the inspection light La output from the light source unit 3 and the transmission characteristics of the optical element 4. As shown in FIG. 3, the wavelength range in which the transmittance of the optical element 4 changes linearly is set based on the wavelength range of the inspection light La. In the example of FIG. 3, the transmittance of the optical element 4 is constant around 0% in the wavelength range shorter than the central wavelength λ1, and increases at a constant slope in the wavelength range including the central wavelengths λ1 to λ5. The transmittance of the optical element 4 is constant around 100% in the wavelength range longer than the central wavelength λ5. The transmittance and reflectance of the optical element 4 are negatively correlated, and when one changes in an increasing direction, the other changes in a decreasing direction.

[0040] The light receiving unit 5 receives light Lb from the object S. The light Lb from the object S may be a directional light beam or non-directional diffused or scattered light. The light receiving unit 5 receives transmitted light Lt that passes through the optical element 4 and reflected light Lr that is reflected by the optical element 4, of the light Lb from the object S. In this embodiment, as shown in FIG. 4 , the light receiving unit 5 includes a first sensor 9A that receives transmitted light Lt that passes through the optical element 4 and a second sensor 9B that receives reflected light Lr that is reflected by the optical element 4. The first sensor 9A and the second sensor 9B are each configured, for example, by an area sensor. The first sensor 9A outputs luminance data pieces based on the luminance of the received transmitted light Lt to the image generating unit 6. The second sensor 9B outputs luminance data pieces based on the luminance of the received reflected light Lr to the image generating unit 6.

[0041] Returning to FIG. 1 , the image generation unit 6, the identification unit 7, and the display unit 8 are physically configured by a computer 10 including a storage device such as a RAM or a ROM, a processor (arithmetic circuit) such as a CPU, a communication interface, etc. The computer 10 may be, for example, a personal computer, a cloud server, or a smart device (smartphone, tablet terminal, etc.). The computer 10 may be configured by an FPGA (Field Programmable Gate Array) or an ASIC (Application Specific Integrated Circuit), etc. The computer 10 is connected to the light receiving unit 5 via a wired or wireless connection so as to be able to communicate information.

[0042] The image generation unit 6 generates centroid wavelength image data GC to be used for inspecting the object S based on the light reception results from the light receiving unit 5. In this embodiment, the image generation unit 6 generates luminance image data GL and centroid wavelength image data GC of the object S for the light Lb from the object S based on the luminance data of the transmitted light Lt and the luminance data of the reflected light Lr. The image generation unit 6 classifies each of the multiple luminance data pieces acquired at different center wavelengths for each line by wavelength range and combines the luminance data pieces for each wavelength range. The image generation unit 6 then performs the following arithmetic processing on the combined luminance data to generate luminance image data GL and centroid wavelength image data GC for each wavelength range. The image generation unit 6 outputs the generated luminance image data GL and centroid wavelength image data GC to the identification unit 7.

[0043] The luminance image data GL is data indicating the total amount of light, which is the sum of the luminance of the transmitted light Lt and the luminance of the reflected light Lr. The pixel value of each pixel included in the luminance image data GL is expressed as T+R, where T is the count value of the luminance of the transmitted light Lt received by each pixel of the first sensor 9A and R is the count value of the luminance of the reflected light Lr received by each pixel of the second sensor 9B. The centroid wavelength image data GC is data indicating the relative values ​​of the luminance of the transmitted light Lt and the luminance of the reflected light Lr. The pixel value of each pixel included in the centroid wavelength image data GC is expressed as (T-R) / (2×(T+R)), where T is the count value of the luminance of the transmitted light Lt received by each pixel of the first sensor 9A and R is the count value of the luminance of the reflected light Lr received by each pixel of the second sensor 9B.

[0044] The value of the centroid wavelength itself is the wavelength at which the transmittance of the optical element 4 is 0%. Ｒ , and the wavelength at which the transmittance of the optical element 4 becomes 50% is λ Ｍ , and the wavelength at which the transmittance of the optical element 4 becomes 100% is λ Ｔ In this case, λ Ｍ + (λ Ｔ -λ Ｒ ) (TR) / (2×(T+R)). In identifying the material of the foreign matter W, it is acceptable if the quantitativeness of the value of the centroid wavelength itself is lost. For this reason, in this embodiment, λ Ｍ = 0 and λ Ｔ -λ Ｒ = 1, the value normalized as the centroid wavelength image data GC is defined as the value of the centroid wavelength image data GC. The value of the centroid wavelength image data GC obtained in this way may be different depending on the material in order to identify the material of the foreign matter W. Therefore, the transmittance and reflectance of the optical element 4 do not necessarily need to change linearly with wavelength. They may change monotonically, such as monotonically increasing or decreasing.

[0045] The identification unit 7 is a part that identifies the material of the foreign matter W in the object S based on the centroid wavelength image data GC. After identifying the material of the foreign matter W, the identification unit 7 outputs information indicating the identification result to the display unit 8. In this embodiment, the identification unit 7 identifies the material of the foreign matter W in the object S based on the centroid wavelength image data GC and luminance image data GL received from the image generation unit 6. Fig. 5(a) is a diagram showing an example of the luminance image data GL. Fig. 5(b) is a diagram showing an example of the centroid wavelength image data GC.

[0046] 5A shows luminance image data GL1 in the wavelength region of center wavelength λ1 (=1200 nm), luminance image data GL2 in the wavelength region of center wavelength λ2 (=1300 nm), luminance image data GL3 in the wavelength region of center wavelength λ3 (=1450 nm), luminance image data GL4 in the wavelength region of center wavelength λ4 (=1550 nm), and luminance image data GL5 in the wavelength region of center wavelength λ5 (=1650 nm). Each of the luminance image data GL1 to GL5 shows the detection results for resin pieces of ABS (acrylonitrile butadiene styrene), PE (polyethylene), PC (polycarbonate), and PP (polypropylene).

[0047] 5B shows centroid wavelength image data GC1 in the wavelength region of center wavelength λ1 (=1200 nm), centroid wavelength image data GC2 in the wavelength region of center wavelength λ2 (=1300 nm), centroid wavelength image data GC3 in the wavelength region of center wavelength λ3 (=1450 nm), centroid wavelength image data GC4 in the wavelength region of center wavelength λ4 (=1550 nm), and centroid wavelength image data GC5 in the wavelength region of center wavelength λ5 (=1650 nm). Like the luminance image data GL1 to GL5, each of the centroid wavelength image data GC1 to GC5 shows the detection results for resin pieces of ABS (acrylonitrile butadiene styrene), PE (polyethylene), PC (polycarbonate), and PP (polypropylene).

[0048] The luminance of the light Lb from the object S varies for each wavelength region and material depending on its unique absorption spectrum. Therefore, the material of the foreign matter W in the object S can be identified by using the luminance image data GL generated from the luminance data of the transmitted light Lt and the luminance data of the reflected light Lr. Furthermore, the transmittance of the optical element 4 varies linearly with wavelength (see FIG. 3 ). Therefore, if a wavelength shift occurs in the light Lb from the object S depending on the material of the foreign matter W, the ratio of transmission to reflection in the optical element 4 changes. Therefore, the centroid wavelength image data GC generated from the luminance data of the transmitted light Lt and the luminance data of the reflected light Lr reflects information regarding the wavelength shift from the centroid wavelength of the inspection light La irradiated on the object S. Therefore, the use of the centroid wavelength image data GC allows the material of the foreign matter W in the object S to be identified with high accuracy.

[0049] By using both the luminance image data GL and the centroid wavelength image data GC in each wavelength region, differences appear in the luminance value of each pixel of the luminance image data GL and the luminance value of each pixel of the centroid wavelength image data GC for each material of the foreign matter W, making it possible to identify with sufficient accuracy the material of the foreign matter W in the target S. The identification of the material of the foreign matter W by the identification unit 7 may be performed, for example, by image processing of the luminance image data GL and the centroid wavelength image data GC, but in this embodiment, the material of the foreign matter W in the target S is identified based on the correlation between the centroid wavelength image data GC and the luminance image data GL.

[0050] 6 is a diagram showing an example of the correlation between the centroid wavelength image data GC and the luminance image data GL for each material of foreign matter, in which the horizontal axis represents the count value of the luminance image data GL and the vertical axis represents the count value of the centroid wavelength image data GC, and the correlation values ​​for PP (polypropylene), PE (polyethylene), and PS (polystyrene) are plotted.

[0051] As shown in Fig. 6, PP, PE, and PS exist in different ranges on the vertical axis of the correlation diagram. In the example of Fig. 6, PP is located in the range where the count value of the centroid wavelength image data GC is 0 to -0.045, PE is located in the range where the count value of the centroid wavelength image data GC is -0.045 to -0.062, and PS is located in the range where the count value of the centroid wavelength image data GC is less than -0.062. Therefore, by setting -0.045 as threshold A and -0.062 as threshold B for the count value of the centroid wavelength image data GC, it is possible to identify the materials of these resins by comparing the centroid wavelength image data GC with thresholds A and B.

[0052] Furthermore, as shown in Fig. 6, even if the resin is made of the same material, the range of existence on the horizontal axis of the correlation diagram differs depending on the color. In the example of Fig. 6, white PP is located in the range where the count value of the luminance image data GL is around 0.45, red PP is located in the range where the count value of the luminance image data GL is 0.2 to 0.3, and dark PP is located in the range where the count value of the luminance image data GL is less than 0.2. Therefore, by setting a threshold value for the count value of the centroid wavelength image data GC in the same way as for the centroid wavelength image data GC, it is possible to identify the color of the resin.

[0053] Identifying the material of a foreign object W based on the correlation between centroid wavelength image data GC and luminance image data GL is applicable to materials other than resin. Figure 7(a) shows data predicting the material of each pixel from image data obtained by capturing images of foreign objects in food using a multispectral camera. In the example shown in Figure 7(a), a frozen pilaf containing ingredients such as rice and shrimp is contaminated with foreign objects such as cockroaches and plastic pieces. Figure 7(b) shows a correlation diagram obtained by applying a dimensionality reduction technique (UMAP: Uniform Manifold Approximation and Projection) to the image data used to predict the material in Figure 7(a), resulting in a two-dimensional representation of the image data. The vertical and horizontal axes represent feature values ​​in arbitrary units. As shown in Figure 7(b), when image data captured using a multispectral camera is two-dimensionalized using UMAP, the distribution ranges of the feature values ​​for each material are somewhat uniform, but the distribution ranges of the feature values ​​for each material are close to each other, necessitating the challenge of improving the accuracy of material identification using a threshold.

[0054] On the other hand, Figure 8(a) shows data in which the material of a foreign object in food is predicted for each pixel from image data obtained by capturing an image of the foreign object contaminated in food using the inspection device 1 according to this embodiment. The example of Figure 8(a) shows data in which the material of the same region of the same food material as Figure 7(a) is predicted for each pixel from centroid wavelength image data and luminance image data. Figure 8(b) is a correlation diagram in which the centroid wavelength image data and luminance image data used to predict the material in Figure 8(a) are two-dimensionalized using UMAP, with the vertical and horizontal axes representing feature quantities in arbitrary units. As shown in Figure 8(b), when the correlation between centroid wavelength image data and luminance image data is used, the distribution ranges of the feature quantities for each material are clearly separated compared to Figure 7(b). Therefore, the accuracy of material identification using a threshold value is improved.

[0055] The display unit 8 is a part that displays the inspection results of the target object S. The display unit 8 is configured, for example, by a general display, a touch panel display, or the like. The display unit 8 displays the identification results of the material of the foreign substance W based on the information indicating the identification results output from the identification unit 7. There are no particular limitations on the display method, and in addition to displaying the identification results of the material of the foreign substance W, the luminance image data GL, the centroid wavelength image data GC, the correlation value between the centroid wavelength image data GC and the luminance image data GL, and the like may be displayed together with the identification results.

[0056] 9 is a flowchart showing an inspection method according to an embodiment of the present disclosure. As shown in FIG. 9, the inspection method according to this embodiment includes an output step S01, a light guiding step S02, a light receiving step S03, a generating step S04, an identifying step S05, and a display step S06. The inspection method according to this embodiment is carried out using, for example, the inspection device 1.

[0057] The output step S01 is a step of outputting inspection light La with wavelengths switched over time toward the object S. In the output step S01, the LED bars of the light source unit 3, which is configured by, for example, a multiband LED bar illumination device, are alternately lit, and inspection light La with wavelengths switched over time is output toward the object S.

[0058] The light guiding step S02 is a step of guiding the light Lb from the object S to the optical element 4 having a transmission characteristic in which the transmittance changes linearly in a predetermined wavelength region. In the light guiding step S02, in accordance with the transmission characteristic of the optical element 4, a part of the light Lb from the object S becomes transmitted light Lt, and the remaining part of the light Lb from the object S becomes reflected light Lr.

[0059] The light receiving step S03 is a step of receiving, of the light Lb from the object S, transmitted light Lt that has passed through the optical element 4 and reflected light Lr that has been reflected by the optical element 4, and outputting luminance data of the transmitted light Lt and luminance data of the reflected light Lr. In the light receiving step S03, of the light Lb from the object S, the transmitted light Lt that has passed through the optical element 4 is received by the first sensor 9A, and the luminance data of the transmitted light Lt is output to the computer 10. Also, of the light Lb from the object S, the reflected light Lr that has been reflected by the optical element 4 is received by the second sensor 9B, and the luminance data of the reflected light Lr is output to the image generating unit 6 of the computer 10.

[0060] The generation step S04 is a step of generating centroid wavelength image data GC of the object S for the light Lb from the object S based on the luminance data of the transmitted light Lt and the luminance data of the reflected light Lr. In the generation step S04, the centroid wavelength image data GC is generated by the image generation unit 6 of the computer 10, and the generated centroid wavelength image data GC is output to the identification unit 7.

[0061] The identification step S05 is a step of identifying the material of the foreign matter W in the target S based on the centroid wavelength image data GC. In the identification step S05, the identification unit 7 of the computer 10 analyzes the centroid wavelength image data GC for each wavelength region and identifies the material of the foreign matter W in the target S. Information indicating the identification result is then output to the display unit 8.

[0062] The display step S06 is a step of displaying the identification result of the material of the foreign substance W in the target object S. In the display step S06, the identification result output from the identification unit 7 is output to the display unit 8. In the display step S06, in addition to displaying the identification result of the material of the foreign substance W, the luminance image data GL, the centroid wavelength image data GC, the correlation value between the centroid wavelength image data GC and the luminance image data GL, etc. may be displayed together with the identification result.

[0063] FIG. 10 is a flowchart showing an example of the details of the generating step S04 and the identifying step S05. As shown in FIG. 10, first, multiple pieces of luminance data are acquired for each wavelength range by alternatingly lighting the light source unit 3 (step S11). Next, the acquired pieces of luminance data are combined for each wavelength range to obtain luminance data for each wavelength range (step S12). Furthermore, the combined luminance data is subjected to arithmetic processing to generate luminance image data GL and centroid wavelength image data GC for each wavelength range (step S13). Then, the material of the foreign matter W in the target object S is identified based on the correlation value between the centroid wavelength image data GC and the luminance image data GL for each wavelength range (step S14).

[0064] As described above, the inspection device 1 uses the optical element 4 having transmission characteristics in which the transmittance changes linearly in a predetermined wavelength range, and generates centroid wavelength image data GC of the object S with respect to the light Lb from the object S based on the luminance data of the transmitted light Lt that passes through the optical element 4 and the reflected light Lr that is reflected by the optical element 4. Because the transmittance of the optical element 4 changes linearly with wavelength, the centroid wavelength image data GC generated from the luminance data of the transmitted light Lt and the luminance data of the reflected light Lr reflects information regarding the amount of wavelength shift from the centroid wavelength of the inspection light La irradiated on the object S, depending on the material of the foreign matter W. Therefore, by using the centroid wavelength image data GC, the material of the foreign matter W in the object S can be accurately identified.

[0065] In this embodiment, the image generation unit 6 generates luminance image data GL of the object S for the light Lb from the object S based on the luminance data of the transmitted light Lt and the luminance data of the reflected light Lr. The identification unit 7 then identifies the material of the foreign matter W in the object S based on the centroid wavelength image data GC and the luminance image data GL. In this way, by combining the centroid wavelength image data GC and the luminance image data GL, the material of the foreign matter W in the object S can be identified with even greater accuracy.

[0066] In this embodiment, the identification unit 7 identifies the material of the foreign matter W in the object S based on the correlation between the centroid wavelength image data GC and the luminance image data GL. The centroid wavelength of the light Lb from the object S differs depending on the specific absorption spectrum and for each material. Therefore, by using the correlation between the centroid wavelength image data GC and the luminance image data GL, the material of the foreign matter W in the object S can be identified quickly and accurately without using a complex method such as multivariate analysis.

[0067] In this embodiment, the inspection device 1 includes a light source unit 3 that outputs inspection light La by switching wavelengths over time toward the target object S. By including such a light source unit 3, centroid wavelength image data GC can be easily acquired in a plurality of wavelength regions.

[0068] In this embodiment, the light receiving unit 5 has a first sensor 9A that receives transmitted light Lt that has passed through the optical element 4, and a second sensor 9B that receives reflected light Lr that has been reflected by the optical element 4. With this configuration, the transmitted light Lt and reflected light Lr can be easily received.

[0069] The present disclosure is not limited to the above-described embodiment. For example, the optical element 4 may have a transmission characteristic in which the transmittance changes linearly in each of a plurality of different wavelength regions, as shown in Fig. 11. Such a transmission characteristic may be realized by designing a dielectric multilayer film of a single tilted dichroic mirror, or may be realized by stacking a plurality of tilted dichroic mirrors or by switching between them using a filter changer.

[0070] In the example of FIG. 11 , the transmittance of the optical element 4 varies linearly and independently with respect to the wavelength ranges including the central wavelength λ1, the wavelength range including the central wavelength λ3, and the wavelength range including the central wavelength λ5. In this case, the transmittance of the optical element 4 varies from approximately 0% to approximately 100% in each wavelength range, resulting in a steeper gradient of the change in the transmittance of the optical element 4 for each wavelength range compared to the example of FIG. 3 . The steeper gradient of the change in the transmittance of the optical element 4 results in a larger change in the relative values ​​of the transmittance and reflectance relative to the amount of wavelength shift when a wavelength shift occurs. This improves the wavelength resolution of the centroid wavelength, and increases the amount of fluctuation in the centroid wavelength image data GC with respect to a wavelength shift that corresponds to the material of the foreign matter W. Therefore, the material of the foreign matter W in the target S can be identified with greater accuracy.

[0071] As shown in Fig. 12, the optical element 4 may have transmission characteristics in which wavelength regions in which the transmittance increases linearly and wavelength regions in which the transmittance decreases linearly alternately. In the example of Fig. 12, the transmittance of the optical element 4 increases linearly in a wavelength region including the center wavelength λ1, decreases linearly in a wavelength region including the center wavelength λ3' (=1400 nm), and increases linearly in a wavelength region including the center wavelength λ5. The dielectric multilayer film of such an optical element 4 can be designed more easily than that of the optical element 4 having the transmission characteristics shown in Fig. 11.

[0072] Furthermore, the image generation method by the image generation unit 6 and the method of identifying the material of the foreign matter W by the identification unit 7 are not limited to the above embodiment, and various methods can be applied. The image generation unit 6 may generate ratio image data GR based on the ratio between the luminance image data GL, GL in two different wavelength regions, and the identification unit 7 may identify the material of the foreign matter W in the target S based on the ratio image data GR and the centroid wavelength image data GC in each of the wavelength regions used to generate the ratio image data GR.

[0073] Depending on the material of the foreign matter W, the light absorptivity may differ depending on the wavelength region. For example, for materials A and B shown in FIG. 13 , material A has a higher absorptivity than material B around a wavelength of 1300 nm, while material B has a higher absorptivity than material A around a wavelength of 1450 nm. For materials A and B, the image generation unit 6 divides the pixel values ​​of each pixel contained in luminance image data GL1 acquired in a wavelength region including the 1300 nm wavelength by the pixel values ​​of each pixel contained in luminance image data GL2 acquired in a wavelength region including the 1450 nm wavelength to generate ratio image data GR (see FIG. 14( a)). The identification unit 7 also generates centroid wavelength image data GC2 in a wavelength region including the 1300 nm wavelength and centroid wavelength image data GC3 in a wavelength region including the 1450 nm wavelength (see FIG. 14( b)).

[0074] FIG. 15 is a flowchart showing an example of the details of the generating step S04 and the identifying step S05 in the above-described modified example. Steps S21 to S23 shown in FIG. 15 are the same as steps S11 to S13 shown in FIG. 10 , and therefore will not be described here. After generating the luminance image data GL and centroid wavelength image data GC for each wavelength region, ratio image data GR based on the ratio between the luminance image data GL, GL for the two different wavelength regions and centroid wavelength image data GC for the wavelength region used to generate the ratio image data GR are generated (step S24). Then, the material of the foreign matter W in the target object S is identified based on the generated ratio image data GR and the centroid wavelength image data GC for the wavelength region used to generate the ratio image data GR (step S25).

[0075] According to this method, by using the ratio image data GR based on the ratio of the luminance image data GL in two different wavelength regions, it becomes easier to grasp the trend of the wavelength shift amount in a wide wavelength region. By combining the ratio image data GR with the centroid wavelength image data GC in each wavelength region used to generate the ratio image data GR, it is possible to more accurately identify the material of the foreign matter W in the target S.

[0076] Furthermore, the identification unit 7 may identify whether the material of the foreign matter W in the target object S is transparent or not, based on a comparison between the centroid wavelength image data GC and a preset threshold value. If the foreign matter W is transparent, the incidence of the inspection light La on the foreign matter W may produce specularly reflected light that is specularly reflected from the surface of the foreign matter W, specularly reflected light that passes through the foreign matter W and is specularly reflected from the back surface, and multiple reflected light that is light that has passed through the foreign matter W and is then multiple-reflected within the foreign matter W. These reflected lights strengthen or weaken each other depending on their phase differences determined by the wavelength and the optical path length, and therefore, when the detected light Lb is dispersed, a spectrum with periodic peaks and valleys on the wavelength axis is obtained.

[0077] For example, as shown in Fig. 16(a), when a non-transparent body Pn is included in the centroid wavelength image data GC, the light transmittance for wavelengths within a certain range (for example, a region where the transmittance of the optical element 4 changes linearly) tends to be approximately constant as shown in Fig. 16(b). With a non-transparent body Pn, there is little deviation between the luminance of the transmitted light Lt that has passed through the optical element 4 and the luminance of the reflected light Lr that has been reflected by the optical element 4, and the count value of the centroid wavelength image data GC tends to be close to 0.

[0078] On the other hand, when a transparent body Pt is included in the centroid wavelength image data GC, periodicity tends to appear in the light transmittance for wavelengths within a certain range (e.g., a region where the transmittance of the optical element 4 changes linearly), as shown in Figures 16(c) and 16(d). As described above, the centroid wavelength image data GC is data that indicates the relative values ​​of the luminance of the transmitted light Lt and the luminance of the reflected light Lr. For this reason, with a transparent body Pt, as shown in Figure 16(c), the luminance of the transmitted light Lt that has passed through the optical element 4 may be significantly greater than the luminance of the reflected light Lr that has been reflected by the optical element 4, or as shown in Figure 16(d), the luminance of the transmitted light Lt that has passed through the optical element 4 may be significantly smaller than the luminance of the reflected light Lr that has been reflected by the optical element 4.

[0079] When taking such a tendency into consideration, the discrimination unit 7 sets, for example, an upper threshold X and a lower threshold Y for the count value of the centroid wavelength image data GC. If the count value of the centroid wavelength image data GC is equal to or greater than Y and equal to or less than X, the discrimination unit 7 discriminates the foreign matter W as a non-transparent body Pn, and if the count value of the centroid wavelength image data GC is less than Y or greater than X, the discrimination unit 7 discriminates the foreign matter W as a transparent body Pt. In this way, by applying a threshold to the centroid wavelength image data GC and detecting the deviation between the luminance of the transmitted light Lt and the luminance of the reflected light Lr, it is possible to discriminate with high accuracy whether the material of the foreign matter W in the target S is a transparent body Pt.

[0080] FIG. 17 is a flowchart showing an example of the details of the generating step S04 and the identifying step S05 in the above-described modified example. Steps S31 to S33 shown in FIG. 17 are the same as steps S11 to S13 shown in FIG. 10 , and therefore will not be described here. After generating the luminance image data GL and centroid wavelength image data GC for each wavelength region, the count value of the generated centroid wavelength image data GC is compared with preset thresholds (here, upper and lower thresholds) (step S34), and it is determined whether the count value of the centroid wavelength image data GC is within the threshold range (step S35). If the count value of the centroid wavelength image data GC is within the threshold range in step S35, it is determined that the foreign matter W is a non-transparent body Pn (step S36). If the count value of the centroid wavelength image data GC is not within the threshold range in step S35, it is determined that the foreign matter W is a transparent body Pt (step S37).

[0081] When identifying whether a foreign object W is a transparent body Pt, the foreign object W may first be identified as a transparent body Pt, and then the material of the foreign object W determined to be a non-transparent body Pn may be identified using the correlation described above or the comparative image data GR. This allows the material of the foreign object W to be identified efficiently.

[0082] Various modifications can also be applied to the configuration of the light receiving unit 5. As shown in Fig. 18, the light receiving unit 5 may have a first sensor 19A that receives one of the transmitted light Lt that has passed through the optical element 4 and the reflected light Lr that has reflected from the optical element 4, and a second sensor 19B that receives the light Lb from the object S without passing through the optical element 4. Even with such a configuration, the transmitted light Lt and the reflected light Lr can be simply received.

[0083] 18, a light dividing element (beam splitter) such as a half mirror 15 is arranged in front of the optical element 4, and reflected light Lr that passes through the half mirror 15 and is reflected by the optical element 4 is received by a first sensor 19A, and a portion of light Lb that is reflected by the half mirror 15 is received by a second sensor 9B without passing through the optical element 4. In this case, the luminance of the reflected light Lr received by the first sensor 9A can be subtracted from the luminance of the light received by the second sensor 9B, thereby indirectly determining the luminance of the transmitted light Lt that passes through the half mirror 15 and then the optical element 4.

[0084] 19A, the light receiving unit 5 may be configured with a line sensor instead of an area sensor, from the viewpoint of receiving the transmitted light Lt and the reflected light Lr in a smaller system. In this case, as shown in Fig. 19A, the light receiving unit 5 may be configured with a single line sensor 21 including a first light receiving area 22A that receives the transmitted light Lt that has passed through the optical element 4 using a plurality of pixels arranged in one direction, and a second light receiving area 22B that receives the light Lb from the target S without passing through the optical element 4 using a plurality of pixels arranged in the same direction as the first light receiving area 22A.

[0085] An inclined dichroic mirror constituting the optical element 4 may be formed directly on the light-receiving surface of the first light-receiving region 22A. In the configuration of Fig. 19(a), similar to the example of Fig. 18, the luminance of the reflected light Lr reflected by the optical element 4 can be indirectly determined by subtracting the luminance of the light Lb received by the first light-receiving region 22A from the luminance of the transmitted light Lt received by the second light-receiving region 22B.

[0086] In addition, the light receiving unit 5 may be composed of multiple line sensors including a first line sensor 21A that receives transmitted light Lt that has passed through the optical element 4 using multiple pixels arranged in one direction, and a second line sensor 21B that receives light Lb from the object S without passing through the optical element 4 using multiple pixels arranged in the same direction as the first line sensor 21A.

[0087] In the example of FIG. 19( b), the first line sensors 21A are arranged in two stages. In this case, optical elements 4 with different wavelength ranges in which transmittance changes linearly can be attached to each of the two stages of the first line sensors 21A, allowing for the generation of centroid wavelength image data GC, luminance image data GL, etc. over a wider wavelength range. When line sensors are arranged in multiple stages as in the example of FIG. 19( b), it is expected that the areas from which luminance data fragments are acquired by each line sensor will be shifted relative to the transport direction of the object S. In this case, delay processing may be performed on the circuit of the line sensor whose area from which luminance data fragments are acquired is ahead in the transport direction of the object S, so that each line sensor outputs luminance data fragments for the same area of ​​the object S.

[0088] 1...inspection device, 3...light source unit, 4...optical element, 5...light receiving unit, 6...image generation unit, 7...identification unit, 9A...first sensor, 9B...second sensor, 19A...first sensor, 19B...second sensor, 21, 21A, 21B...line sensor, 22A...first light receiving area, 22B...second light receiving area, GC...centroid wavelength image data, GL...luminance image data, GR...ratio image data, Lb...light from object, Lt...transmitted light, Lr...reflected light, S...object, W...foreign matter, Pn...non-transparent body, Pt...transparent body.

Claims

1. An inspection device for identifying the material of foreign matter in an object, comprising: an optical element having characteristics in which transmittance and reflectance change monotonically in a predetermined wavelength range; a light receiving unit that receives, of light from the object, transmitted light that has passed through the optical element and reflected light that has been reflected by the optical element, and outputs luminance data of the transmitted light and luminance data of the reflected light; an image generating unit that generates centroid wavelength image data of the object for the light from the object based on the luminance data of the transmitted light and the luminance data of the reflected light; and an identifying unit that identifies the material of foreign matter in the object based on the centroid wavelength image data.

2. The inspection device according to claim 1, wherein said optical element has a characteristic that transmittance and reflectance change monotonically in each of a plurality of wavelength ranges different from one another.

3. An inspection device as claimed in claim 1 or 2, wherein the image generation unit generates luminance image data of the object in response to light from the object based on the luminance data of the transmitted light and the luminance data of the reflected light, and the identification unit identifies the material of a foreign object in the object based on the centroid wavelength image data and the luminance image data.

4. The inspection device according to claim 3, wherein the identification unit identifies the material of the foreign matter in the object based on the correlation between the centroid wavelength image data and the luminance image data.

5. An inspection device according to claim 3 or 4, wherein the image generation unit generates ratio image data based on the ratio between the luminance image data in two mutually different wavelength regions, and the identification unit identifies the material of a foreign substance in the object based on the ratio image data and the centroid wavelength image data in each of the wavelength regions used to generate the ratio image data.

6. An inspection device according to any one of claims 1 to 5, wherein the identification unit identifies whether the material of the foreign matter in the object is transparent or not based on a comparison between the centroid wavelength image data and a preset threshold value.

7. An inspection device according to any one of claims 1 to 6, comprising a light source unit that outputs inspection light by switching wavelengths over time toward the object.

8. An inspection device according to any one of claims 1 to 7, wherein the light receiving unit has a first sensor that receives the transmitted light that has passed through the optical element, and a second sensor that receives the reflected light that has been reflected by the optical element.

9. An inspection device according to any one of claims 1 to 7, wherein the light receiving unit has a first sensor that receives either the transmitted light that has passed through the optical element or the reflected light that has reflected from the optical element, and a second sensor that receives light from the object without passing through the optical element.

10. An inspection device according to any one of claims 1 to 7, wherein the light receiving unit is configured by a single line sensor including a first light receiving area that receives the transmitted light that has passed through the optical element using a plurality of pixels arranged in one direction, and a second light receiving area that receives light from the object using a plurality of pixels arranged in the same direction without passing through the optical element.

11. An inspection device according to any one of claims 1 to 7, wherein the light receiving unit is composed of a plurality of line sensors including a first line sensor that receives the transmitted light that has passed through the optical element using a plurality of pixels arranged in one direction, and a second line sensor that receives light from the object using a plurality of pixels arranged in the same direction without passing through the optical element.

12. An inspection method for identifying the material of a foreign substance in an object, comprising: a light guiding step of guiding light from the object to an optical element having properties such that transmittance and reflectance change monotonically in a predetermined wavelength range; a light receiving step of receiving, of the light from the object, transmitted light that has passed through the optical element and reflected light that has been reflected by the optical element, and outputting luminance data of the transmitted light and luminance data of the reflected light; a generation step of generating centroid wavelength image data of the object for the light from the object based on the luminance data of the transmitted light and the luminance data of the reflected light; and an identification step of identifying the material of a foreign substance in the object based on the centroid wavelength image data.

13. An inspection method according to claim 12, wherein in the light guiding step, light from the object is guided to the optical element having a characteristic in which the transmittance and reflectance change monotonically in each of a plurality of mutually different wavelength regions.

14. An inspection method as claimed in claim 12 or 13, wherein the generating step generates luminance image data of the object in response to light from the object based on the luminance data of the transmitted light and the luminance data of the reflected light, and the identifying step identifies the material of a foreign object in the object based on the centroid wavelength image data and the luminance image data.

15. An inspection method according to claim 14, wherein said identifying step identifies the material of a foreign substance in said object based on the correlation between said centroid wavelength image data and said luminance image data.

16. An inspection method as claimed in claim 14 or 15, wherein in the generating step, ratio image data is generated based on the ratio between the luminance image data in two mutually different wavelength regions, and in the identifying step, the material of the foreign matter in the object is identified based on the ratio image data and the centroid wavelength image data in each of the wavelength regions used to generate the ratio image data.

17. An inspection method according to any one of claims 12 to 16, wherein the identification step identifies whether the material of the foreign matter in the object is transparent or not based on a comparison between the centroid wavelength image data and a preset threshold value.

18. An inspection method according to any one of claims 12 to 16, further comprising an output step of outputting inspection light, the wavelength of which is switched over time, toward the object.

19. An inspection method according to any one of claims 12 to 18, wherein the light receiving step uses a first sensor that receives the transmitted light that has passed through the optical element, and a second sensor that receives the reflected light that has been reflected by the optical element.

20. An inspection method according to any one of claims 12 to 18, wherein the light receiving step uses a first sensor that receives either the transmitted light that has passed through the optical element or the reflected light that has reflected from the optical element, and a second sensor that receives light from the object without passing through the optical element.

21. An inspection method according to any one of claims 12 to 18, wherein the light receiving step uses a single line sensor including a first light receiving area that receives the transmitted light that has passed through the optical element using a plurality of pixels arranged in one direction, and a second light receiving area that receives light from the object using a plurality of pixels arranged in the same direction without passing through the optical element.

22. An inspection method according to any one of claims 12 to 18, wherein the light receiving step uses a plurality of line sensors including a first line sensor that receives the transmitted light that has passed through the optical element using a plurality of pixels arranged in one direction, and a second line sensor that receives light from the object using a plurality of pixels arranged in the same direction without passing through the optical element.

Images