X-ray inspection equipment and item inspection system

Abstract

The present invention provides an X-ray inspection device and an article inspection system equipped therewith that can perform highly accurate relative mass inspection without requiring the hassle of changing product settings, even when inspecting multiple types of items with different components on the same line. [Solution] An X-ray inspection apparatus that performs relative mass inspection of an object to be inspected PF based on an X-ray inspection image Dpx obtained by X-ray imaging of the object to be inspected PF at a predetermined transport position, comprising: a morphological feature imaging unit 51 that images the appearance of the object to be inspected PF and outputs appearance image data Dea that includes classifiable morphological features of the object to be inspected PF; a variety feature image recognition unit 53 that detects variety-specific image feature quantities Fc that can identify the type of content component of the object to be inspected PF from the appearance image data; and an inspection control unit 30 that variably sets specific inspection parameters γ and Wr for relative mass inspection according to the variety-specific image feature quantities Fc recognized therein.

Description

[Technical Field]

[0001] The present invention relates to an X-ray inspection apparatus and an article inspection system, and more particularly to an X-ray inspection apparatus capable of measuring at least the relative mass of an object to be inspected based on the dose (transmission amount) of X-rays transmitted through the object and the distribution of the transmission amount, and an article inspection system equipped therewith. [Background technology]

[0002] Conventionally, X-ray inspection devices and product inspection systems equipped with such devices are known to have a function (hereinafter simply referred to as the relative mass measurement function) that measures the relative mass or volume of an object under inspection based on the measured amount of X-rays transmitted through each unit transmission area of ​​the object under inspection.

[0003] Known X-ray inspection devices and product inspection systems that have this type of relative mass measurement function include, for example, those that irradiate an object to be measured between an X-ray generator and an X-ray detection unit with X-rays from the X-ray generator, detect the amount of X-rays that have passed through the object to be measured with the X-ray detection unit, calculate the mass thickness for each unit transmission region of the object to be measured (mass thickness of the material constituting the unit transmission region = density ρ × thickness x) based on the distribution of the detected amount of X-rays transmitted or the corresponding absorbed dose, and calculate the total mass of the object to be inspected, which corresponds to the integral value of the mass thickness of the entire transmission region by integrating the mass thickness for each unit transmission region (see, for example, Patent Document 1).

[0004] The relative mass measurement method in this material inspection system is based on the following idea: When an incident X-ray of intensity I0 is attenuated by absorption during transmission through a material and detected by the X-ray detection unit after transmission, the detected X-ray intensity I is equal to the material's intrinsic mass absorption coefficient μ relative to the incident intensity I0. m Since it decreases exponentially depending on the linear absorption coefficient μ / density ρ of the material and the mass thickness of the material (=density ρ × thickness x), it can be expressed by the following equation

[11] , where the mass thickness ρ·x of the material is the logarithmic decay rate corresponding to the X-ray transmittance (ratio of X-ray intensity before and after transmission) I / I0 and the mass absorption coefficient μ of the material m Based on this, it can be expressed by the following equation

[12] . I = I0·e -μm·ρ·x ······

[11] ρ·x = -(1 / μ m )·ln(I / I0) ·······

[12]

[0005] Other conventional X-ray inspection devices with relative mass measurement capabilities are also known, for example, that perform a conversion process from density data of X-ray images in each transmission region to density data of equivalent thickness images corresponding to the thickness of the object under inspection in each transmission region, measure the volume of the object under inspection in multiple transmission regions based on the density data of equivalent thickness images corresponding to each of the multiple transmission regions, and then convert the measured volume values ​​for each transmission region into converted values ​​in mass units using a preset conversion ratio to measure the mass of the object under inspection (see, for example, Patent Document 2). [Prior art documents] [Patent Documents]

[0006] [Patent Document 1] Japanese Patent Publication No. 2002-296022 [Patent Document 2] Japanese Patent Publication No. 2006-300887 [Overview of the project] [Problems that the invention aims to solve]

[0007] However, with conventional X-ray inspection devices and product inspection systems that have the relative mass inspection function described above, when multiple types of products that are nearly identical in appearance but have different internal components are manufactured and inspected on the same production line, in order to perform highly accurate relative mass inspection that can handle even subtle differences in internal components, it was necessary to read the identification information such as label markings for each type of product with different internal components, understand the differences in internal components from other types, and then determine whether inspection parameters for the inspected item are necessary and make any necessary setting changes. Therefore, when product types are frequently switched or when multiple types are mixed in the production line, there was a problem in that frequent changes to the inspection parameter settings were time-consuming.

[0008] On the other hand, when multiple types of articles with similar external appearances but different internal components are manufactured and inspected on the same production line, failing to set the inspection parameters for each type as described above, which enables highly accurate relative mass inspection, could lead to large errors in relative mass measurements due to inappropriate settings, potentially degrading the accuracy of the mass inspection.

[0009] Furthermore, when consumer products consist of multiple varieties with differing ingredients, consumers can often easily and quickly identify the product they want to buy based on partial differences in packaging or visible product form, without having to meticulously check identification information on labels, etc. However, such differences in characteristic points have not been effectively utilized in changing the settings of the aforementioned inspection parameters.

[0010] The present invention aims to solve the aforementioned unresolved problems of the conventional approach, and to provide an X-ray inspection device and an article inspection system equipped therewith that can perform highly accurate relative mass inspection without requiring the change of product type settings, even when inspecting multiple types of inspected items with different components on the same line. [Means for solving the problem]

[0011] (1) An X-ray inspection apparatus according to the present invention is an X-ray inspection apparatus that, in order to achieve the above objective, performs a relative mass inspection of an object to be inspected based on an X-ray inspection image taken by X-ray transmission at a predetermined transport position of an object to be inspected being transported in a predetermined direction, and is characterized by comprising: a morphological feature imaging unit (e.g., a visible light camera) that images the object to be inspected and outputs classification image data including classifiable morphological features of the object to be inspected; a variety feature image recognition unit that detects variety-specific image feature quantities that can identify the content component types of the object to be inspected from the classification image data; and an inspection control unit that variably sets specific inspection parameters for the relative mass inspection according to the variety-specific image feature quantities recognized by the variety feature image recognition unit.

[0012] In this configuration, the X-ray inspection device captures the object to be inspected by X-ray transmission at a predetermined transport position to generate an X-ray inspection image. Based on this X-ray inspection image, a relative mass inspection of the object to be inspected is performed. Simultaneously, the object to be inspected is captured by a morphological feature imaging unit such as a visible light camera, and classification image data is output. The variety feature image recognition unit then detects variety-specific image features from the classification image data that can identify the type of content component of the object to be inspected. The inspection control unit then variably sets specific inspection parameters for the relative mass inspection in the X-ray inspection device according to these variety-specific image features. Therefore, for example, when the shapes are substantially the same and the package color, product image, product type display, etc., indicating the package type can be visually identified separately from identification information such as product identification labels, product identification becomes possible from the same perspective as the consumer, and inspection parameters can be set or changed according to the identification result. As a result, even when multiple similar types of inspected objects have the same or similar shapes, and when the ingredient labels on the inspected objects in the manufacturing line are difficult to see from above or the side, effective variety identification and rapid, accurate setting of inspection parameters based on the identification results can be easily performed without cumbersome setup work. Furthermore, since the image features of each variety are detected from the classification image data, there is no need to read identification codes or textual information.

[0013] (2) In a preferred embodiment of the present invention, when a plurality of objects to be inspected are transported in the predetermined direction, the morphological feature imaging unit images each of the plurality of objects to be inspected and outputs classification image data for each object to be inspected, the variety feature image recognition unit detects the image feature quantities for each variety of each object to be inspected from the classification image data of each object to be inspected, and the inspection control unit dynamically sets specific inspection parameters for each variety of each object to be inspected in the X-ray inspection apparatus according to the image feature quantities for each variety of each object to be inspected recognized by the variety feature image recognition unit.

[0014] In this case, when multiple identical items are transported one after the other, the specific inspection parameters for that variety are maintained. However, when multiple different types of items are transported one after the other, before the subsequent item reaches a predetermined transport position, an X-ray inspection image is created, and relative mass inspection is performed by the X-ray inspection device based on that X-ray inspection image, the specific inspection parameters for relative mass inspection of the preceding item (hereinafter simply referred to as the preceding item) can be changed from those previously identified by the variety feature image recognition unit based on the image features of the preceding item (hereinafter simply referred to as the preceding item) to specific inspection parameters for relative mass inspection of the subsequent item (hereinafter simply referred to as the succeeding item) based on the image features of the succeeding item (hereinafter simply referred to as the succeeding item). Therefore, even when switching varieties or when multiple varieties pass through the inspection line together, the setting of specific inspection parameters for each variety can be performed quickly, accurately, and easily without any hassle.

[0015] (3) The variety feature image recognition unit detects variety-specific image feature quantities from the classification image data using an object detection method that can identify the content component types of the object to be inspected, and the inspection control unit recognizes the variety of the object to be inspected based on the variety-specific image feature quantities, and can be configured to variably set specific inspection parameters for the relative mass inspection according to the content component types of the object to be inspected identified by the variety.

[0016] In this case, since the object detection method can detect image features specific to each variety from the classification image data, the position and number of image elements (objects) that contribute to variety identification can also be effectively utilized for variety identification. The object detection method referred to here may be, for example, one that uses a neural network or one that uses a rule-based object detection algorithm.

[0017] (4) The variety feature image recognition unit detects an image of a two-dimensional code from the classification image data, and the inspection control unit executes a predetermined code reading process for reading the two-dimensional code to recognize the variety of the inspected object, and according to the content component type of the inspected object specified by the variety, it is also possible to adopt a configuration in which specific inspection parameters for the relative mass inspection are variably set.

[0018] In this case, when a two-dimensional code can be read as an image feature amount for each variety from the classification image data, the variety can be identified from the two-dimensional code, and specific inspection parameters are variably set using the identification result.

[0019] (5) The article inspection system according to the present invention is an article inspection system provided with an X-ray inspection device having the configuration of (1) or (2) above, wherein the morphological feature imaging unit is mounted on an appearance inspection device arranged in front of the X-ray inspection device, and classification image data including classifiable morphological features of the inspected object transmitted and output from the appearance inspection device is received and input to the inspection control unit of the X-ray inspection device.

[0020] With this configuration, in the article inspection system of the present invention, when classification image data of the inspected object transmitted and output from the appearance inspection device is received and input to the inspection control unit of the X-ray inspection device, the variety feature image recognition unit detects image feature amounts for each variety of the inspected object from the classification image data, and specific inspection parameters for the relative mass inspection in the X-ray inspection device are variably set by the inspection control unit according to the image feature amounts for each variety. Therefore, it is not necessary to set a dedicated appearance imaging unit in the X-ray inspection device, and the system cost can be reduced.

Advantages of the Invention

[0021] According to the present invention, it is possible to provide an X-ray inspection device capable of performing high-precision relative mass inspection without causing the trouble of changing the variety setting even when inspecting a plurality of types of inspected objects with different content components on the same line, and an article inspection system equipped with the same. [Brief explanation of the drawing]

[0022] [Figure 1] This is a schematic diagram of an X-ray inspection apparatus according to one embodiment of the present invention. [Figure 2] Figure 1 is a block diagram of the relative mass measurement function unit in the X-ray inspection apparatus shown. [Figure 3] Figure 3(a) shows a first parameter setting table as a typical example of variable parameter settings in the inspection control unit of an X-ray inspection apparatus according to one embodiment of the present invention, in which brightness P0 indicating background image density, brightness P1 indicating foreground image density, and unsaturated maximum brightness Qmax indicating background image density are set. Figure 3(b) shows a second parameter setting table in which the correction index value γ in equation

[11] , which indicates the exponential decrease in transmitted X-ray intensity with respect to irradiation intensity for each type of product, and the weight of the measurement standard product for that type are set. [Figure 4] Figure 4(a) shows the relative mass inspection results for multiple inspected objects when inspection parameters are set according to the product type in an X-ray inspection apparatus according to one embodiment of the present invention. The vertical axis shows the difference between the actual weight and the measured value, and the horizontal axis shows the actual weight of each inspected object. Figure 4(b) shows the relative mass inspection results for a comparative example, which corresponds to the case when inspection parameters are not set according to the product type in an X-ray inspection apparatus according to one embodiment of the present invention. The vertical axis shows the difference between the actual weight and the measured value, and the horizontal axis shows the actual weight of each inspected object. [Figure 5] This is a schematic diagram of an article inspection system including an X-ray inspection apparatus according to another embodiment of the present invention. [Modes for carrying out the invention]

[0023] Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings. (One embodiment) Figures 1 to 4 show an X-ray inspection apparatus according to one embodiment of the present invention.

[0024] First, let me explain its structure.

[0025] The X-ray inspection apparatus 1 shown in Figure 1 includes an article transport unit 10, an X-ray imaging unit 20, an inspection control unit 30, an operation display unit 40, and a PLC 50. The article transport unit 10 is used as an inspection conveyor, with a front conveyor 14 positioned in front of it and a rear conveyor 15, which is equipped with sorting equipment (not shown), positioned behind it.

[0026] The item transport unit 10, acting as an inspection conveyor, transports the item to be inspected PF (item) in the direction of the rightward arrow (transport direction) in Figure 1. For example, by motor-driving either roller 12 or 13 across a loop-shaped inspection belt 11, the item to be inspected PF is transported at a constant speed on the transport path 11a, which is the upper section of the belt, and passes through the inspection area of ​​the X-ray imaging unit 20. In this example, the item transport unit 10 is a belt conveyor with a motor-driven roller 12 or 13 and a flat transport path 11a. However, it may also be a system in which the item to be inspected PF is pumped through a tubular transport path, or a system in which the item to be inspected PF passes through the inspection area by its own weight.

[0027] The inspected object PF is not particularly limited, but examples include retort foods in which food is sealed in a bag-like shape by vapor deposition or heat sealing of packaging film, such as multiple types of curry with different levels of spiciness or ingredients, or small processed foods in which multiple types of jelly-like foods are sealed in a long, flat tube shape by vapor deposition or heat sealing of packaging film, such as pet food that is small enough to be held with a fingertip.

[0028] The X-ray imaging unit 20 has an X-ray generator 21 and an X-ray detector 23 positioned with the transport path 11a of the article transport unit 10 in between. In this case, the X-ray generator 21 and X-ray detector 23 are positioned opposite each other while being separated vertically, but they may also be positioned so as to be separated both vertically and horizontally.

[0029] The X-ray generator 21, although not shown in detail, has, for example, an X-ray tube 22 inside a metal box, and the X-ray tube 22 is immersed in insulating oil for cooling inside the box. This X-ray generator 21 is an X-ray irradiation unit that irradiates X-rays toward a predetermined inspection area in the transport path 11a of the object to be inspected PF, and in this embodiment, it has an X-ray tube 22 that irradiates X-rays downward from the vertically above side toward the object to be inspected PF.

[0030] The X-ray tube 22 is positioned such that its axial direction is oriented approximately parallel to the predetermined transport direction. On its cathode side, a negative DC potential is applied to light up a filament to a high temperature, causing electrons to be emitted and focused by a focusing electrode. On its anode side, a positive DC potential is applied to the target, accelerating the electrons from the filament with a high voltage and causing them to collide with the target, thereby generating X-rays within a predetermined energy range from the target.

[0031] The X-rays generated by the X-ray tube 22 are irradiated in a fan-beam shape, directed downwards from the X-ray window at the bottom of the aforementioned box towards the inspection area into which the object to be inspected PF is transported, and spreading in a line scanning direction perpendicular to the transport direction. The anode of the X-ray tube 22 may be of a fixed or rotating type.

[0032] The X-ray imaging unit 20 also includes a filament power supply circuit (not shown) and a high-voltage circuit that applies a high voltage between the filament of the X-ray tube 22 and the target.

[0033] The X-ray detector 23, although not shown in detail, is a line sensor positioned, for example, directly beneath the transport path 11a of the inspection belt 11. It comprises a scintillator that absorbs and emits X-rays of a predetermined energy (wavelength, penetrating power), and a photodiode array consisting of N (e.g., several hundred) light-receiving elements arranged perpendicular to the transport direction of the object under inspection PF to receive light (scintillation light) from the scintillator. This X-ray detector 23 absorbs X-rays that are irradiated onto and transmitted through the object under inspection PF, causing it to emit light according to the transmission intensity of the X-rays, and outputs an electrical signal corresponding to the amount of light received by the photodiode at predetermined scanning cycles. It should be noted that the X-ray detector 23 is not limited to this indirect conversion method that indirectly converts X-rays into electrical signals; a direct conversion method is also possible.

[0034] The photodiode array of the X-ray detector 23 stores the photocurrents generated simultaneously by each of the N photodetectors for a predetermined storage time, and outputs a brightness detection signal Lx, which is a voltage signal, based on the charge corresponding to the product of the photocurrent and the storage time.

[0035] The inspection control unit 30, although its detailed configuration is not shown, is hardware-configured to include, for example, a microcomputer having a CPU, ROM, RAM, and I / O interface, an auxiliary storage device that stores control programs for performing various functions in a readable format in cooperation with the ROM, and timer circuits, driver circuits, etc. The CPU executes predetermined arithmetic processing and the control program while exchanging data with the RAM, etc., according to the software such as control programs and setting information stored in the ROM, etc. The hardware may also include FPGA (Field Programmable Gate Array) or DSP (Digital Signal Processor), etc. Furthermore, the various functions referred to herein are the functions of each functional unit and means for X-ray output control, X-ray image data generation, inspection control, and display output control, etc., as described below.

[0036] This inspection control unit 30 has a transport control function that controls the transport speed and transport interval of the object PF to be inspected by the inspection belt 11 in the item transport unit 10, and an inspection control function that controls the X-ray irradiation intensity and irradiation period in the X-ray imaging unit 20, and controls the X-ray detection period of the X-ray detector 23 and the detection period of each object PF to be inspected according to the transport speed of the object PF to be inspected. Note that the configuration of the transport control function unit is the same as known, so a detailed illustration thereof is omitted here.

[0037] As shown in Figure 1, the inspection control unit 30 comprises a relative mass measurement function unit 30A and a variety parameter matching function unit 30B. The relative mass measurement function unit 30A includes an inspection image acquisition unit 31, a belt surface correction unit 32, a relative mass measurement processing unit 33, and a determination unit 35, while the variety parameter matching function unit 30B includes a variety feature image recognition unit 53 and an inspection parameter variable setting unit 54.

[0038] Specifically, the inspection image acquisition unit 31 acquires a brightness detection signal Lx for each line scan from the photodiode array of the X-ray detector 23 and generates X-ray image data Dpx.

[0039] Furthermore, the belt surface correction unit 32 is configured to correct the value of the brightness detection signal Lx from the X-ray detector 23 to match the white reference value for each of the N photodetectors when the X-rays from the X-ray generator 21 pass only through the transport path 11a, which is the belt surface before the object PF to be inspected is brought in (when there is no object PF to be inspected), that is, to perform a light reception sensitivity correction on the belt surface (so-called shading correction).

[0040] The determination unit 35 determines, based on the measurement results of the relative mass measurement processing unit 33, whether the measured mass Wv of the object PF under inspection falls within a predetermined tolerance range relative to its actual weight reference value Wr.

[0041] As shown in Figure 2, the relative mass measurement processing unit 33 includes a data conversion processing unit 33a, a volume measurement unit 33b, and a mass conversion unit 33c.

[0042] The data conversion processing unit 33a is a data conversion processing means that performs a conversion process (details will be described later) from the density data P of each transmission region in the X-ray image data Dpx to the density data Q(P) of an equivalent thickness image corresponding to the thickness t of the object PF under inspection in that transmission region, and has a conversion processing program and a working memory area for this purpose. The equivalent thickness referred to here is the X-ray equivalent thickness τ (=α·t) corresponding to the line absorption coefficient α and thickness t of each part of the object PF under inspection, and is the thickness equivalent to the density value (luminance value) of the X-ray image.

[0043] Here, the density data Q(P) of the equivalent thickness image is a density value (in this case, a brightness value) obtained by logarithmically transforming the density data P of the X-ray image using a conversion formula described later to correspond to the amount of X-ray absorption. The maximum brightness P0 occurs when there is no object PF under inspection, the value of the X-ray transmission is at its maximum, and the amount of X-ray absorption by the object PF under inspection is zero. The minimum brightness P1 occurs when the value of the X-ray transmission is at its minimum, and the amount of X-ray absorption by the object PF under inspection is at its maximum.

[0044] Furthermore, the phenomenon in which N0 photons of a predetermined energy are reduced to N by X-rays being irradiated toward the object PF under inspection and passing through the object PF (thickness t, linear absorption coefficient α) and the inspection belt 11 (thickness t0, linear absorption coefficient α0) can be approximated by the Lambert-Beer law by the following equation [1]. N=N0exp(-α·t-α0·t0) ···[1]

[0045] Furthermore, in an X-ray fan-beam optical system such as the one employed in this embodiment, the amount of light received I for each pixel (image) within the focal elevation angle (90°-θ) range of the X-ray detector 23, which consists of a scintillator-type CCD line sensor, is a value corresponding to the X-ray irradiation intensity in each transmission region. I(θ) = {1 / (1+tan)} 2 θ)}·I(0°) ···[2] This can be expressed as follows, but typically, the light-receiving sensitivity correction, i.e., the belt surface correction mentioned above, adjusts the detection sensitivity of the X-ray detector 23 so that the amount of light received I(θ) on the conveyor belt surface (background image area) before the goods are loaded, so that the output of all sensor elements of the X-ray detector 23 is adjusted to a predetermined white reference value (maximum brightness) with respect to the amount of light received I(θ) on the belt surface only before the goods are loaded.

[0046] Therefore, the amount of light received on the belt surface (background) before the loading of goods, after the light receiving sensitivity correction, becomes a constant value that is approximately N0exp(-α0·t0), and the amount of light received in each transmission region, which no longer depends on the angle θ due to the light receiving sensitivity correction of the X-ray detector 23, can be considered to be calculable by the following equation [3]. I'=I0exp(-α·t) ···[3]

[0047] Furthermore, α·t is a value that directly represents the amount of X-ray absorption by the material that has passed through the X-rays from the X-ray generator 21 until they are detected by the X-ray detector 23. By associating this with the density value of the X-ray absorption image, it is possible to create an image in which the density value of pixels is larger and the brightness is lower for materials with high X-ray absorption rates or for areas with thicker thickness in the direction of X-ray transmission.

[0048] Therefore, using the aforementioned light-receiving quantities I' and I0, and taking α·t as the equivalent X-ray thickness τ in each transmission region of the object PF under inspection, and taking γ as the correction index value corresponding to the X-ray intensity (penetrating power) that decays exponentially during material transmission, the image density value J(τ) of the X-ray transmission image can be expressed by the following equation [4]. J(τ)=(α·t) γ ={ln(I0)-ln(I')} γ ...[4]

[0049] The data conversion processing unit 33a has a linearity correction function that adjusts the correction index value γ within a predetermined range set according to the object PF under inspection, and corrects the data to improve the linearity of the density data of the equivalent thickness image corresponding to the thickness t of the object PF under inspection.

[0050] Specifically, the belt surface correction unit 32 and the data conversion processing unit 33a are configured to input the brightness value P0, which is the background density value in the X-ray image, the brightness value P1, which is the representative density of the foreground in the X-ray image, and the maximum density Qmax of the equivalent thickness image, respectively. Based on these settings, the conversion process to density data Q(P) of the equivalent thickness image in each transmission region of the object PF under inspection is performed according to the following equation [5].

[0051]

number

[0052] The density data Q(P) in this equivalent thickness image is the logarithmically transformed value corresponding to the density J(τ) in the distribution image of the equivalent thickness τ described above, but it has also been corrected for the light-receiving sensitivity on the belt surface as described earlier.

[0053] Furthermore, the correction index value γ is adjusted within a predetermined range depending on the type of object PF being inspected. By performing corrections by adding or subtracting the correction index value γ within this range, good linearity of the density data Q(P) of the equivalent thickness image with respect to the thickness t of the object PF in each transmission region can be ensured, even when X-ray inspection is performed on many food items as objects PF.

[0054] Furthermore, the data conversion processing unit 33a takes the density value Q(Pa) of the equivalent thickness image calculated by formula [5] based on the density value Pa (luminance value) of the X-ray image in a specific transmission region of the representative workpiece that serves as the standard for the workpiece PF, and sets it to the value of Qa (Qa = Q(Pa)) expressed by the following formula [6]. Using the calculated density value Q(Pa) of the equivalent thickness image and the density value Pa of the X-ray image in a specific transmission region of the representative workpiece (for example, the X-ray image density of the transmission region corresponding to the maximum thickness), the density value P2 in formula [6], which is the representative density value of the foreground in the X-ray image of the representative workpiece, is calculated in advance.

number

[0055] Then, using the calculated density value P2 (luminance value), the system updates the representative density value P1 of the foreground in equation [5], and then performs a conversion process from the density data P of the X-ray image in each transmission region to the density data Q(P) of an equivalent thickness image having a density corresponding to the thickness t of the object PF under inspection.

[0056] The representative workpiece PF to be inspected here is a typical good workpiece. This representative workpiece is passed through the inspection space once to calculate the density value P2 in equation [6], and this calculated value is substituted into P1 in equation [5] to perform the conversion process from the density data P of the X-ray image to the density data Q(P) of the equivalent thickness image for each workpiece PF to be inspected.

[0057] Furthermore, the data conversion processing unit 33a has a data lookup table (not shown) that defines the conversion conditions, and a processor that performs data conversion processing using this conversion table. The lookup table stores the density levels of the equivalent thickness image density data Q(P), which is the result of the above-described conversion processing, associated with the density levels of the density data P of the X-ray images at multiple stages.

[0058] The volume measurement unit 33b calculates the volume V of the object under inspection PF by summing the density data Q(P) of the equivalent thickness image corresponding to each of the multiple transmission regions over the entire measurement range of each object under inspection PF, and has a measurement processing program and a working memory area for this purpose.

[0059] Furthermore, the volume measurement unit 33b performs volume calculations only for transmission regions on the scanning line of the X-ray detector 23 where the density level of the equivalent thickness image density data Q(P) is equal to or greater than a predetermined noise cut threshold. By summing only the valid data from the equivalent thickness image density data Q(P) (hereinafter also referred to as slice data) obtained in each scan from the leading edge to the trailing edge in the transport direction of the object under inspection PF, the volume V of each object under inspection PF can be calculated.

[0060] The mass conversion section 33c includes a conversion processing program for converting the volume measurement value V for each transmission region measured by the volume measurement section 33b into a conversion value (mass) in mass units at a preset conversion ratio (conversion rate), and a mass conversion coefficient λ of the inspection object PF read from a variety parameter file (not shown). W It has a coefficient retaining memory area for storing things such as this, and a work memory area for conversion processing.

[0061] Here, when the volume measurement value V of a certain inspection object PF is multiplied by the mass conversion coefficient λ depending on the variety of the inspection object PF W the volume measurement value V of the inspection object PF can be converted into a value Wv in mass units. If the value converted into mass units is Wv, this conversion value Wv can be expressed as Wv = λ W ·V.

[0062] Also, an operation display section 40 is connected to the determination section 35 (which may be the mass conversion section 33c), and the mass conversion value Wv of the inspection object PF converted into mass units by the mass conversion section 33c, or alternatively the volume measurement result V of the inspection object PF before conversion, is output to the operation display section 40 respectively.

[0063] The determination section 35 further incorporates a program for determining whether or not the conversion value Wv of the volume measurement value V of the inspection object PF converted into a value in mass units is within a predetermined allowable range with respect to a predetermined mass reference value Wr, whereby obtaining pass / fail information OK / NG of the mass measurement result for each inspection object PF, and being able to output it to the operation display section 40 together with the mass conversion value W of the inspection object PF.

[0064] On the other hand, above the upstream conveyor 14 of the item transport unit 10 (inspection conveyor), a visible light camera 51 (morphological feature imaging unit) is provided. This camera takes an external image of the item PF to be inspected using visible light, for example, at a predetermined transport position of the item PF to be inspected, upstream of the position where the item detection sensor 28, such as a photoelectric sensor, detects the item PF to be inspected entering the X-ray imaging unit 20, and outputs external image data Dea (classification image data, see Figure 2) that includes classifiable morphological features of the item PF to be inspected. Although the visible light camera 51 uses visible light for external imaging, it is also conceivable to provide a near-infrared camera or a multispectral camera to visualize, from a different perspective, classifiable morphological features of the item PF that are difficult to grasp with X-ray imaging, particularly changes in the characteristics of its variety. Furthermore, the timing of imaging of the item PF to be inspected by the visible light camera 51 or an alternative morphological feature imaging unit is not particularly limited, and it is also conceivable to do so simultaneously with imaging by the X-ray imaging unit 20.

[0065] Classifiable morphological features of an object under inspection (PF) are those features that allow multiple objects under inspection (PF) to be classified as products, where each object is substantially the same in shape, and which include external features such as package color, images of contents, product type indications that show the difference between multiple types within that classification, or morphological features that can be identified by human senses other than sight, such as touch, and which contribute to classification and variety identification separately from identification information such as product identification labels.

[0066] This visible light camera 51 is connected to a PLC (Programmable Logic Controller) 50, and the external image data Dea from the visible light camera 51 is input to the product parameter matching function unit 30B of the inspection control unit 30 via the camera image input unit 52 in the PLC 50. Although not shown in the figures, the PLC 50 may also be equipped with a tablet-type information terminal that functions as a programming tool and a setting input switcher.

[0067] The PLC 50 pre-stores control procedures as a program list for sequence control to control, for example, the X-ray irradiation drive of the X-ray imaging unit 20, the detection drive of the X-ray detector 23, the input of an item detection signal from the item detection sensor 28, and the transport drive of the item transport unit 10, via their respective drive circuits (not shown). Following these control procedures, the PLC 50 drives the transport drive circuit of the item transport unit 10, the X-ray irradiation drive circuit and detection drive circuit of the X-ray imaging unit 20, etc., while the X-ray inspection apparatus 1 is in operation. At the same time, it acquires item detection information from the item detection sensor 28 and appearance image data Dea from the visible light camera 51. This allows for timely X-ray imaging of each object under inspection PF passing through the inspection section of the X-ray imaging unit 20 during operation, and imaging of classifiable morphological features of the object under inspection PF with the visible light camera 51.

[0068] During operation of the X-ray inspection apparatus 1, the inspection control unit 30 uses the relative mass measurement function unit 30A to acquire a brightness detection signal Lx for each line scan from the photodiode array of the X-ray detector 23 to generate X-ray image data Dpx, and then performs the aforementioned relative mass measurement function based on the X-ray image data Dpx.

[0069] Meanwhile, the variety parameter matching function unit 30B of the inspection control unit 30 detects image feature quantities of external variety characteristics that perform classification and variety identification functions of the object PF under inspection, based on external image data Dea from the visible light camera 51, using the variety feature image recognition unit 53. When the image feature quantities Fc of external variety characteristics recognized by the variety feature image recognition unit 53 are sent to the inspection parameter variable setting unit 54, specific inspection parameters related to the relative mass measurement processing unit 33, which is responsible for the relative mass inspection function of the X-ray imaging unit 20, are variably set by the inspection parameter variable setting unit 54.

[0070] The term "product type" of the inspected object PF, as used herein, refers to product type information that allows for the identification of the specific components of the inspected object PF. For example, in the case of multiple types (product types) of products belonging to the same classification (same at the major classification level) with identical or similar external shapes, this corresponds to one of those types (a group of items of the same type at the minor classification level). Furthermore, the term "classification" as used herein refers to the product classification of multiple inspected objects PF that can be manufactured simultaneously on the same production line. This classification covers a group of products where multiple inspected objects PF include multiple types (product types) with different internal components, but their external shapes are identical or similar, and where they also share common classifiable morphological characteristics.

[0071] More specifically, when multiple items to be inspected PF are transported in a predetermined direction by the item transport unit 10, the visible light camera 51 captures an image of each of the multiple items to be inspected PF, and the appearance image data Dea of ​​each item to be inspected PF is output from the camera image input unit 52 of the PLC 50 to the product feature image recognition unit 53 of the inspection control unit 30.

[0072] Then, when the variety feature image recognition unit 53 detects the image feature quantity Fc of the variety characteristics of the external appearance of each object PF from the external image data Dea of ​​each object PF to be inspected, the inspection parameter variable setting unit 54, which takes in the image feature quantity Fc, recognizes the classification and variety (shown as type(Fc) in Figure 2) of each object PF to be inspected captured by the X-ray imaging unit 20 according to the image feature quantity Fc for each variety of the object PF recognized by the variety feature image recognition unit 53, and variably sets the specific inspection parameters for each variety related to the relative mass measurement function, such as the correction index γ and the reference value of the actual weight (measurement standard product weight) Wr.

[0073] To process such variable settings, the inspection parameter variable setting unit 54 stores and retains typical values ​​as set values ​​for inspection parameters common to multiple product types, such as the luminance P0, which is equivalent to the maximum luminance indicating the background image density in the X-ray image data Dpx; the luminance P1, which is equivalent to the minimum luminance indicating the foreground image density of the object under inspection PF, etc., in the X-ray image data Dpx; and the unsaturated maximum luminance Qmax, which is equivalent to the background image density, as shown in the first parameter setting table in Figure 3(a).

[0074] The variable inspection parameter setting unit 54 further stores and maintains pre-set values ​​for each product category of the object under inspection PF, such as the correction index value γ in the aforementioned

[11] equation and the weight of the measurement standard for that type, as shown in the second parameter setting table in Figure 3(b). Then, for example, when one of the multiple product types a, b, c, ..., j, ... of the object under inspection PF in Figure 3(b) is selected, the inspection parameters for that type, such as the correction index value γ and the weight Wr of the measurement standard for that type, are variably set. The weight Wr of the measurement standard here is the weight corresponding to the relative mass value Wv of the measurement standard for each type (variety) whose volume V and material are known, and the setting of this weight Wr determines the mass conversion coefficient λ for each type. W This will also be set for each type.

[0075] In this example, the visible light camera 51 is a standalone morphological feature imaging unit that captures external features. However, it may also be mounted on an external inspection device positioned before the X-ray imaging unit 20. The external image data Dea of ​​the object under inspection, transmitted and output from the external inspection device, can be received and input to the inspection control unit 30 of the X-ray imaging unit 20.

[0076] In the variety parameter matching function unit 30B of the inspection control unit 30, the variety feature image recognition unit 54 uses an object detection method using a neural network to detect variety-specific image feature quantities Fc from the appearance image data Dea of ​​the object under inspection PF, which can identify the content component types of the object under inspection PF. Based on the variety-specific image feature quantities Fc, the unit recognizes the variety of the object under inspection PF, and according to the content component types of the object under inspection PF identified by that variety, it selects specific inspection parameters for relative mass inspection in the X-ray imaging unit 20, such as the correction index γ and the mass conversion coefficient λ mentioned above. W The reference value for actual weight (measurement reference weight) Wr is dynamically set to a variable value.

[0077] The object detection method using a neural network referred to herein is one in which an object detection algorithm is acquired through AI learning using a neural network, such as the object detection methods known as YOLO (You Only Look Once) or SSD (Single Shot Detector). By detecting the image feature quantity Fc, described later, from the appearance image data Dea of ​​the object under inspection PF using this object detection method, the product classification and variety of the object under inspection PF can be recognized. However, the object detection method referred to in this invention is not limited to the object detection method using a neural network. As another variation, it is also conceivable to use an object detection method that detects the image feature quantity Fc from the appearance image data Dea of ​​the object under inspection PF using rule (logic)-based techniques such as pattern matching.

[0078] The variety feature image recognition unit 54 may also have an additional code recognition function that detects an image of a predetermined two-dimensional code from the external image data Dea of ​​the object under inspection PF. In this case, the variety feature image recognition unit 54 can be configured to perform a predetermined code reading process to read the two-dimensional code, recognize the variety of the object under inspection PF, and variably set specific inspection parameters for relative mass inspection in the X-ray imaging unit 20 according to the type of content component of the object under inspection PF identified by that variety.

[0079] The operation display unit 40 is a touch panel type, such as an LCD (Liquid Crystal Display), and combines the functions of a display means and an operation input means. The function of the display means is to display the operating status and setting information of the X-ray imaging unit 20, and other various information required in relation to X-ray inspection, on the display screen. The function of the operation input means is to manually perform various touch panel operations, such as selecting a display screen, switching the operating mode between inspection mode, setting mode, or other modes, and inputting various parameters in setting mode, and to input requested information in accordance with the user's operation input.

[0080] This operation display unit 40 is not limited to a touch panel integrated with the X-ray inspection device 1, but may also be provided in the form of a portable tablet-type information terminal or other similar device, or may be additionally installed in the form of a separate display and control panel from the touch panel integrated with the X-ray inspection device 1.

[0081] Next, I will explain how it works.

[0082] In this embodiment, first, for example, an automatic setting sequence sets the item transport section conditions and the tube voltage of the X-ray tube 22 in the X-ray generator 21 based on the setting parameters stored in the first parameter setting table and other inspection parameters common to multiple product types.

[0083] Next, under predetermined X-ray output conditions of the X-ray generator 21, the X-ray detector 23, which detects X-rays within a predetermined energy range, performs the aforementioned light-receiving sensitivity correction (shading correction) on the belt surface to adjust the brightness detection signal Lx, i.e., the electrical signals corresponding to the amount of light received from the N photodiode arrays for each scanning period, to a white reference value according to the type of object PF under inspection and the inspection conditions.

[0084] Next, predetermined inspection controls are performed according to the selected product classification or the selection of varieties included in that product classification.

[0085] In this inspection control process, first, the classification of products manufactured on the same production line at the same time for the product under inspection PF is selected and set, for example, by its classification number, or one of several similar varieties of product under inspection PF included in that product classification is selected and set, for example, by its symbol or number.

[0086] If this selection is made, then, as information on multiple similar varieties of the inspected object PF included in that classification, the second parameter setting table, for example shown in Figure 3(b), is read according to the selected classification, and a list is read of the product types that can be identified by the variety characteristic image for multiple varieties included in the same classification, the correction index value γ in equation

[11] corresponding to that identification type, and the weight of the measurement standard product for that type.

[0087] Next, among the multiple objects to be inspected PF that are transported sequentially into the inspection space while moving apart in a predetermined direction, a specific object to be inspected PF that has reached the appearance imaging position is imaged by the visible light camera 51, and the appearance image data Dea output from the visible light camera 51 is taken up by the variety feature image recognition unit 53 in the variety parameter matching function unit 30B of the inspection control unit 30 via the camera image input unit 52 in the PLC 50.

[0088] Then, based on the captured external image data Dea, the variety feature image recognition unit 53 detects image feature quantities Fc of external variety characteristics that perform classification and variety identification functions for the object PF under inspection, and these image feature quantities Fc are sent to the inspection parameter variable setting unit 54.

[0089] In this case, if the image feature quantity Fc is the same as the variety feature of the object PF being inspected immediately before, the specific inspection parameters applied to the relative mass measurement processing unit 33, which is responsible for the relative mass inspection function of the line inspection device 20, are maintained.

[0090] On the other hand, if the image feature quantity Fc detected by the variety feature image recognition unit 53 is not the same as the variety feature of the inspected object PF of the immediately preceding inspection target, and is of a different type, then specific inspection parameters for the relative mass measurement processing unit 33, which is responsible for the relative mass inspection function of the line inspection device 20, are variably set by the inspection parameter variable setting unit 54.

[0091] In other words, the relative mass measurement processing unit 33 updates the parameters for inspection control, which are pre-set and stored according to the product variety to be inspected, to target products of the type (variety) corresponding to the image feature quantity Fc detected by the variety feature image recognition unit 53, so that they are the target of relative mass measurement and inspection.

[0092] As described above, in the X-ray inspection apparatus 1 of this embodiment, the object to be inspected PF is imaged by X-ray transmission at a predetermined transport position to generate an X-ray inspection image Dpx, and relative mass inspection of the object to be inspected PF is performed based on the X-ray inspection image Dpx. At that time, the object to be inspected PF is imaged by the visible light camera 51 at a predetermined X-ray imaging position, and when appearance image data Dea is output from the visible light camera 51, the variety feature image recognition unit 53 detects variety-specific image feature quantities Fc that can identify the content component types of the object to be inspected PF based on the appearance image data Dea, and according to the variety-specific image feature quantities Fc, the inspection control unit 30 variably sets specific inspection parameters related to the relative mass measurement function of the X-ray inspection apparatus 1.

[0093] Therefore, for example, if multiple items to be inspected (PF) are substantially the same in shape, and the package color, images of contents, product type indications, etc., which indicate the difference between each type, can be visually observed as part of the appearance, separate from identification information such as product identification labels, then it becomes possible to identify the products from the same perspective as consumers, and the inspection parameters can be changed or updated according to the identification results.

[0094] As a result, even when the package shapes of multiple similar types of inspected items PF are identical or similar, and the content ingredient labeling of the inspected items PF in the manufacturing line is difficult to see from above or the side, and especially when the size of the inspected items PF is small enough to be held with fingertips and the content ingredient labeling is even more difficult to identify, effective product identification and rapid and accurate setting of inspection parameters based on the identification results can be easily performed without cumbersome setup work.

[0095] Furthermore, the image feature quantity Fc for each type of inspected object PF can be a combination of multiple image features, including the color of the package, images of the contents, and product type indications. Therefore, it can be easily detected from the appearance image data Dea using object detection methods, and the type of information is not limited, unlike methods for reading identification codes or textual information.

[0096] Furthermore, in this embodiment, when multiple objects PF to be inspected are transported in a predetermined direction, the visible light camera 51 captures an image of each object PF and outputs an appearance image data Dea of ​​each object PF. The variety feature image recognition unit 53 of the inspection control unit 30 detects the variety-specific image feature quantity Fc of each object PF from the appearance image data Dea of ​​each object PF. The variety feature image recognition unit 54 of the inspection control unit 30 can dynamically change the specific inspection parameters for each variety related to the relative mass inspection function of each object PF in the X-ray inspection device 1 according to the recognized variety-specific image feature quantity Fc of each object PF.

[0097] Therefore, when multiple inspected objects PF of the same classification but different varieties are transported one after the other, before the relative mass inspection is performed by the X-ray inspection device 1 based on the X-ray inspection image Dpx of the subsequent inspected object PF, the specific inspection parameters for the relative mass inspection of the preceding variety, which were previously identified and set by the variety feature image recognition unit 53 according to the variety-specific image features of the preceding first inspected object DF (preceding variety), can be changed to specific inspection parameters for the relative mass inspection of the second inspected object PF (successor variety) newly recognized by the variety feature image recognition unit 53. Thus, even immediately after a variety change or when multiple varieties pass through the inspection line 11 together, the setting of specific inspection parameters for each variety can be performed quickly, accurately, and easily without any hassle.

[0098] In addition, in this embodiment, the variety feature image recognition unit 53 detects variety-specific image feature quantities Fc from the appearance image data Dea of ​​the object under inspection PF using an object detection method, and the inspection control unit 30 can variably set specific inspection parameters for relative mass inspection in the X-ray inspection device 1 according to the variety-specific content component types of the object under inspection PF recognized based on the variety-specific image feature quantities Fc by the inspection parameter variable setting unit 54.

[0099] Therefore, since the object detection method can detect image features Fc for each variety from the appearance image data Dea, the position and number of image elements (targets of object detection) that contribute to variety identification can also be effectively utilized for variety identification.

[0100] Furthermore, if the variety feature image recognition unit 53 is equipped with a function to detect a two-dimensional code image from the external image data of the object under inspection PF, the inspection control unit 30 can perform a predetermined code reading process to read the two-dimensional code to recognize the variety of the object under inspection PF, and can variably set specific inspection parameters related to the relative mass inspection function of the X-ray inspection device 1 according to the content component type of the object under inspection PF identified by that variety. Therefore, the variety can be identified from the external image data Dea, as well as from the two-dimensional code as part of the image feature quantity Fc for each variety, and specific inspection parameters can be variably set using the identification result.

[0101] (Comparison and verification of inspection performance) Example 1 As the objects to be inspected (PF), multiple types of jelly-like pet food were enclosed in long, flat tubes formed by vapor deposition of packaging film and sealed on three sides, according to the product specifications that can be identified by the product classification shown in Figure 3(b). These were transported to the X-ray imaging unit 20 in a mixed state of multiple varieties within the same classification, and just before each object to be inspected (PF) reached the X-ray imaging position, relative mass measurement and mass inspection were performed while variably setting specific inspection parameters for each variety related to the relative mass measurement function according to the variety (one of types a, b, or c).

[0102] Figure 4(a) shows the results (difference from actual weight) of relative mass measurement and mass inspection performed on multiple types of the test object PF in Example 1, for example, six types, with 10 test samples for each type that differed slightly in actual weight (g). The cross marks (×) in the figure represent the maximum value among the 10 relative mass measurement results for each test sample, and the black circles (●) in the figure represent the minimum value among the 10 relative mass measurement results for each test sample, respectively, as the difference between the actual weight and the relative mass inspection result for each test sample.

[0103] Comparative Example 1 Figure 4(b) shows the results (difference from actual weight) of relative mass measurement and mass inspection performed on six types of inspected material PF, similar to Example 1, by mixing 10 test samples for each type, each with slightly different actual weights (g). However, regardless of the change in the type of inspected material PF, the results show that relative mass measurement and mass inspection were performed while maintaining constant specific inspection parameters for each type related to the relative mass measurement function.

[0104] (Comparison of measurement accuracy) Although the test samples of multiple types of inspected objects PF used for relative mass measurement in Example 1 and the test samples of multiple types of inspected objects PF used for relative mass measurement in Comparative Example 1 were the same, in Example 1, where specific inspection parameters for each type of inspected object PF were changed according to the change in the type of inspected object PF reaching the X-ray imaging position, the variation in the difference between each relative mass measurement result and the actual weight was kept small, as shown in Figure 4(a). In contrast, in Comparative Example 1, where the specific inspection parameters for each type of inspected object PF were kept constant regardless of the change in the type of inspected object PF reaching the X-ray imaging position, the variation in the difference between each relative mass measurement result and the actual weight was large, as shown in Figure 4(b).

[0105] As is clear from these results, according to this embodiment, even when inspecting multiple types of inspected objects PF with different contents on the same line, it is possible to provide an X-ray inspection device 1 that can perform highly accurate relative mass inspection without the trouble of changing the product settings.

[0106] (Other embodiments) Figure 5 shows an article inspection system equipped with an X-ray inspection apparatus according to another embodiment of the present invention. Since this embodiment has a configuration similar to the X-ray inspection apparatus 1 of one embodiment, in Figure 5, components similar to those of one embodiment will be described below using the same reference numerals as the corresponding components of one embodiment shown in Figures 1 to 3.

[0107] In the X-ray inspection system 5 of this embodiment shown in Figure 5, the visible light camera 51 is mounted on the visual inspection device 2, which is positioned in front of the X-ray inspection device 1. Visual image data Dea of ​​the object to be inspected PF is transmitted and output from the visual inspection device 2 and received and input to the inspection control unit 30 of the X-ray inspection device 1.

[0108] In this embodiment as well, although multiple types of inspected objects PF1, PF2, and PF3 with different internal components are mixed and inspected on the same inspection line, the line inspection device 1 can change specific inspection parameters for each type of inspected object PF1, PF2, and PF3 according to the change in type of inspected object PF1, PF2, and PF3. Therefore, highly accurate relative mass inspection can be performed without the trouble of changing the type settings.

[0109] Furthermore, in this embodiment, when the appearance image data Dea of ​​multiple types of objects to be inspected PF1, PF2, and PF3 transmitted and output from the appearance inspection device 2 is received and input to the inspection control unit 30 of the X-ray inspection device 1, the product type feature image recognition unit 53 detects the image feature quantity Fc for each product type of objects to be inspected PF1, PF2, and PF3 from the appearance image data Dea, and the inspection parameter variable setting unit 54 of the inspection control unit 30 variably sets specific inspection parameters for relative mass inspection in the X-ray inspection device 1 according to the image feature quantity Fc for each product type. Therefore, it is not necessary to set up a dedicated appearance imaging unit in the X-ray inspection device 1, which reduces costs.

[0110] In the embodiments described above, the X-ray detector 23 was an indirect conversion type having a scintillator, but it may also be a direct conversion type, such as a photon counting type X-ray detector. Furthermore, while a visible light camera 51 was given as an example of a morphological feature imaging unit, it is conceivable that a near-infrared camera or a multispectral camera, etc., may be provided in order to visualize changes in the variety of the inspected object PF, which are difficult to grasp with X-ray imaging, from a different perspective. Moreover, in cases where there are differences in detailed shape that allow for variety identification, it is also conceivable that changes in the variety of the inspected object PF, which are difficult to grasp with the X-rays used for relative mass measurement, may be imaged using X-rays with a different energy intensity than those used for relative mass measurement, in a way that enables variety detection.

[0111] As described above, the present invention provides an X-ray inspection apparatus that can perform highly accurate relative mass inspection without the need to change product settings, even when multiple types of inspected objects with different internal components are mixed together on the same line for inspection. The present invention is useful for an X-ray inspection apparatus capable of measuring at least the relative mass of an inspected object based on the dose (transmission amount) of X-rays transmitted through the inspected object and the distribution of that transmission amount, and for article inspection systems equipped therewith in general. [Explanation of symbols]

[0112] 1. X-ray inspection device 2. Visual Inspection Device 5. X-ray Inspection System 10. Goods transport section 11 Inspection belt 11a Conveyor path 12,13 Laura 14 Front Conveyor 15. Downstream conveyor 20 X-ray imaging section 21 X-ray generator 22 X-ray tube 23 X-ray detector 28. Item detection sensor 30 Inspection Control Unit 30A Relative Mass Measurement Function Unit 30B Variety Parameter Adaptation Function Unit 31. Inspection Image Acquisition Unit 32. Belt surface correction unit (belt surface light reception sensitivity correction unit) 33 Relative Mass Measurement Processing Unit 33a Data conversion processing unit 33b Volume measurement section 33c Mass conversion section 35 Judgment section 40 Operation display section 50 PLCs (Programmable Logic Controllers) 51. Visible light camera (external image imaging unit, variety identification morphology imaging unit) 52 Camera image input section 53 Variety characteristic image recognition unit 54 Variable Inspection Parameter Setting Unit 54 Variety Feature Image Recognition Unit a,b,c,j Type (variety) Dea (Degree Image Data) - Image data for classification, images of classifiable morphological features. DPX X-ray examination images Fc Image Features (Classifiable Characteristic Morphology) Lx Brightness detection signal P: Brightness value (density value of the X-ray image in each transmission region) P0 luminance value (maximum luminance, density value of the equivalent thickness image in the X-ray image) P1 Representative density of the foreground (luminance value, representative density of the foreground in an X-ray image) P2 Density Value (Representative density and brightness values ​​of the foreground in the X-ray image of a representative workpiece) PF, PF1, PF2, PF3 - Items under inspection Qmax Equivalent Thickness Maximum Density Q(P) Equivalent thickness image density data Q(Pa) Equivalent thickness image density value (density value of the equivalent thickness image of a representative workpiece) Wr Reference Value (Reference value for actual weight) Wv mass measurement γ Correction Index Value (Test Parameter)

Claims

1. An X-ray inspection apparatus that performs relative mass inspection of an object to be inspected (PF) transported in a predetermined direction, based on an X-ray inspection image (Dpx) captured by X-ray transmission at a predetermined transport position, A morphological feature imaging unit (51) captures an image of the object to be inspected and outputs classification image data (Dea) that includes classifiable morphological features of the object to be inspected, A variety feature image recognition unit (53) detects variety-specific image feature quantities (Fc) that can identify the content component types of the object to be inspected from the classification image data, An inspection control unit (30) that variably sets specific inspection parameters (γ, Wr) for the relative mass inspection according to the image feature quantities for each variety recognized by the variety feature image recognition unit, An X-ray inspection apparatus characterized by comprising the following features.

2. The X-ray inspection apparatus according to claim 1, characterized in that when a plurality of objects to be inspected are transported in the predetermined direction, the morphological feature imaging unit images each of the plurality of objects to be inspected and outputs classification image data for each object to be inspected, the variety feature image recognition unit detects the image feature quantities for each variety of each object to be inspected from the classification image data of each object to be inspected, and the inspection control unit dynamically sets specific inspection parameters for each variety of each object to be inspected in the relative mass inspection of each object to be inspected in the X-ray inspection apparatus according to claim 1.

3. The X-ray inspection apparatus according to claim 1 or 2, characterized in that the variety feature image recognition unit detects variety-specific image feature quantities from the classification image data that can identify the content component types of the object to be inspected using an object detection method, the inspection control unit recognizes the variety of the object to be inspected based on the variety-specific image feature quantities, and variably sets specific inspection parameters for the relative mass inspection according to the content component types of the object to be inspected identified by the variety.

4. The X-ray inspection apparatus according to claim 1 or 2, characterized in that the variety feature image recognition unit detects an image of a two-dimensional code from the classification image data, the inspection control unit performs a predetermined code reading process to read the two-dimensional code to recognize the variety of the object to be inspected, and sets specific inspection parameters for the relative mass inspection variably according to the type of content component of the object to be inspected identified by the variety.

5. An article inspection system comprising an X-ray inspection apparatus according to claim 1 or 2, The article inspection system is characterized in that the morphological feature imaging unit is mounted on a visual inspection device (2) positioned in front of the X-ray inspection device, and classification image data including classifiable morphological features of the object to be inspected, transmitted and output from the visual inspection device, is received and input to the inspection control unit of the X-ray inspection device.

Images