Article inspection device

Abstract

The present invention provides an article inspection device capable of reducing downtime for action confirmation and performing the action confirmation in parallel in an operation action for performing article inspection. The article inspection device includes: an image processing unit that performs image processing using a predetermined image processing algorithm on image pickup data of an article and outputs data for determining a quality state of the article; a determination unit that determines whether or not the quality state is good on the basis of the data; an operation display unit for displaying the determination result; an operation confirmation image generation unit for generating, as an operation confirmation image used in the operation confirmation processing unit, an image including at least an NG feature image in which the quality state of the article is determined to be defective, on the basis of the inspection image during the operation operation; and an operation confirmation processing unit that executes operation confirmation processing using a predetermined image processing algorithm in parallel with the operation of the image processing unit during the operation operation, and determines whether or not the operation of the image processing unit and the determination unit is normal.

Description

Technical Field

[0001] This invention relates to an inspection device for articles, and more particularly to an inspection device for articles that uses a prescribed image processing algorithm to inspect the quality status of articles by taking pictures of inspection images of a specified variety of articles. Background Technology

[0002] In the past, in the inspection of items, the image processing algorithm with a specified image processing filter or combination thereof corresponding to the inspection item was applied to the camera data of the inspected item, so that the specified quality status of the inspected item could be inspected with high precision.

[0003] In such an item inspection device, it is necessary to select and set an image processing filter corresponding to the specific inspection item of the item being inspected from multiple image processing filters that are individually manufactured and pre-stored in memory based on the characteristics of the item being inspected or the characteristics of the foreign object to be detected. Therefore, there is a technology that automates the selection and setting function of the image processing filters required for inspection, so that this selection and setting operation can be easily performed without depending on the operator's experience.

[0004] As such an article inspection device, for example, there is an article inspection device that includes: a detection unit that outputs a signal that reacts to the transported article; a foreign object detection unit that determines whether there is a foreign object in the transported article based on the signal from the detection unit; an operation confirmation processing unit that uses a foreign object sample to determine whether the operation of the detection unit is normal; and a display unit that displays the determination result of the foreign object detection unit and the determination result of the operation confirmation processing unit, wherein the operation confirmation processing unit determines whether the operation of the detection unit is normal based on whether the amplitude and phase of the output signal of the detection unit when testing and transporting a foreign object sample specified for each type falls within the allowable range (for example, see Patent Document 1).

[0005] Furthermore, as another article inspection device, there is a known article inspection device comprising: an inspection unit for inspecting conveyed articles; first and second pass detection sensors for detecting the passage of inspected articles moved into the inspection unit at different positions in the conveying direction; and a control unit capable of switching the mode of the inspection unit to a normal inspection mode and an operation confirmation mode. When the second pass detection sensor on the upstream side does not detect the passage of the inspected article and the first pass detection sensor on the downstream side detects the passage of the inspected article, the control unit performs an automatic mode switch from the normal inspection mode to the operation confirmation mode. On the other hand, if the detection interval of inspected articles detected by the first pass detection mechanism in the operation confirmation mode is more than a predetermined time, the device returns to the normal inspection mode (for example, see Patent Document 2).

[0006] Patent Document 1: Japanese Patent Application Publication No. 2009-31149

[0007] Patent Document 2: Japanese Patent Application Publication No. 2013-113784

[0008] However, in conventional item inspection devices as described above, the determination of whether the inspection action or detection signal of the item inspection device is normal is based on manual and physical actions such as inspectors placing samples for action confirmation on an inspection line with an inspection section and a conveyor section. Therefore, there are problems such as long operator operation time due to lack of data quantification and low efficiency in action confirmation.

[0009] Furthermore, if errors such as forgetting to return the sample used for confirmation occur, it will cause downtime on the production line, including the inspection line, resulting in a decrease in production efficiency.

[0010] Furthermore, the inspection confirmation of the item inspection device was performed not during the operation of the item inspection device, but in other operation confirmation modes or maintenance modes where the inspection was stopped. Therefore, it is impossible to accurately confirm the unevenness or deviation of product quality during the operation of the item inspection.

[0011] Furthermore, if the operation is confirmed not only during product switching but also periodically after the switching, the following problems will arise: the operation rate of the item inspection device will decrease not only due to the inspector's operation confirmation, but also due to the increased downtime or the labor and time required for managing the inspection results data. Summary of the Invention

[0012] The present invention was made in view of the unresolved problems described above, and its object is to provide an item inspection apparatus that can reduce the downtime for transitioning to the inspection stop state in order to perform action confirmation, and perform the action confirmation in parallel during the operation of item inspection.

[0013] (1) In order to achieve the above objective, the article inspection apparatus of the present invention is characterized by comprising: an image processing unit that performs image processing on an inspection image obtained by capturing an article being transported by a camera mechanism using a predetermined image processing algorithm and outputs judgment data on the quality state of the article; a judgment unit that determines whether the quality state of the article is good or bad based on the judgment data; a display unit that displays the result determined by the judgment unit; an action confirmation image generation unit that generates a defective image including defective features of the article as a action confirmation image based on the inspection image during the operation of the image processing unit, the judgment unit and the display unit; and an action confirmation processing unit 54 that performs action confirmation processing on the action confirmation image using the predetermined image processing algorithm in parallel with the operation of the image processing unit during the operation to determine whether the operation of the image processing unit and the judgment unit is normal.

[0014] According to this structure, motion confirmation processing is performed in parallel with the operation of the image processing unit during operation. Based on the inspection image during operation, the motion confirmation processing unit determines whether the operation of the image processing unit and the determination unit is normal. At this time, as the motion confirmation image used in the motion confirmation processing unit, a defective image is generated by the motion confirmation image generation unit, which includes at least the defective features that determine the quality state of the item as defective. Therefore, by using the motion confirmation image, which is generated during operation based on the inspection image of the inspected item that is mainly a qualified item and includes the defective features, to confirm whether the motion confirmation image is a defective item, it is possible to determine whether the operation of the image processing unit and the determination unit is normal. As a result, an item inspection device is created that can perform motion confirmation, such as checking whether the detection signal from the detection unit is appropriate, during operation without transitioning to an inspection stop state such as a motion confirmation mode.

[0015] Furthermore, the motion confirmation processing unit uses the same prescribed image processing algorithm as the determination unit. Therefore, based on the inspection image of the item that has been determined to be a qualified product by the determination unit as the inspection image of a qualified product, effective motion confirmation can be performed by using motion confirmation processing of a defective product image with an image of defective features added to that inspection image. However, the motion confirmation processing unit can use the same prescribed image processing algorithm as the determination unit, and execute in parallel during operation whether the determination of whether the inspection image of the item that has been determined to be a qualified product by the determination unit is qualified and whether the motion confirmation image of the defective product image is defective.

[0016] (2) In the article inspection device of the present invention, the structure can be configured such that the action confirmation image generation unit synthesizes multiple parts with different image features in the inspection image during the operation into multiple partial images of the defect feature part to generate the action confirmation image.

[0017] According to this structure, the action confirmation image synthesizes multiple partial images of different parts of the inspection image into defective feature parts, thereby making it less susceptible to the influence of uneven or biased quality status of the inspected item.

[0018] (3) The image generation unit for action confirmation can also be configured as follows: the image generation unit for action confirmation synthesizes multiple defect feature images of the quality state of the item into defective images by combining the qualified product inspection image obtained by the determination unit in the operation of the item to determine the quality state of the item.

[0019] According to this structure, during normal operation, a motion confirmation image is generated that synthesizes the defective feature image based on the qualified product inspection image. By stably ensuring the defective features of the motion confirmation image, stable motion confirmation can be performed.

[0020] (4) The article inspection device of the present invention can also be configured as follows: it further includes a control unit, which has a first control function unit that controls the operation of the image processing unit, the determination unit and the display unit to determine the quality status of each article, and the control unit also has a second control function unit that controls the operation of the action confirmation processing unit and the action confirmation image generation unit to determine whether the operation of the image processing unit and the determination unit is normal during the operation of the first control function.

[0021] According to this structure, the control unit can, during the operation of determining the quality status of an item by controlling the actions of the image processing unit and the determination unit through the first control function unit, simultaneously execute an action confirmation process to determine whether the actions of the image processing unit and the determination unit, which are being inspected, are normal by controlling the action confirmation image generation unit and the action confirmation processing unit through the second control function unit. Here, the action confirmation process can be set to be processed in a background hidden from the display screen during the operation; however, in this case, if the action confirmation result is abnormal, a warning display or warning sound can be output.

[0022] (5) In the article inspection apparatus of the present invention, the structure can be configured such that the inspection image is an X-ray transmission image obtained by taking X-rays of the article.

[0023] According to this structure, similar to the X-ray inspection images based on the camera data of each item, as the action confirmation images used in the action confirmation processing unit during operation, the X-ray transmission images based on the camera data of each item are combined to form at least the defective feature section for defect determination of the quality status of the item, thereby making it easy to produce the action confirmation image of the defective product image.

[0024] Invention Effects

[0025] According to the present invention, an item inspection device is provided that can perform action confirmation, such as whether the detection signal from the detection unit is appropriate, during operation without transitioning to an inspection stop state such as an action confirmation mode. Attached Figure Description

[0026] Figure 1 This is a schematic structural diagram of an article inspection device according to one embodiment of the present invention.

[0027] Figure 2 This is an explanatory diagram of an example of an action confirmation image in an article inspection device according to an embodiment of the present invention.

[0028] Figure 3 This is a flowchart outlining the general processing steps of an action confirmation process performed during the operation of an article inspection device according to an embodiment of the present invention.

[0029] Figure 4 This is a schematic structural diagram of an article inspection device according to another embodiment of the present invention. Detailed Implementation

[0030] Hereinafter, the methods for carrying out the present invention will be described with reference to the accompanying drawings.

[0031] (One implementation method) Figures 1 to 3 An article inspection device according to one embodiment of the present invention is shown.

[0032] First, let's explain the structure.

[0033] like Figure 1 As shown, the article inspection apparatus 1 of this embodiment includes a conveying unit 10, an inspection unit 20, and a control unit 30. The inspection unit 20 irradiates the article P (hereinafter, simply referred to as article P) being inspected, which is conveyed by the conveying unit 10, with X-rays while simultaneously detecting image data corresponding to the distribution of transmitted X-rays. The quality status of article P is then inspected based on this detected image data. Furthermore, the quality status referred to here refers to the suitability of the quality or physical quantities required for article P as a product, such as the presence or absence of foreign matter, the presence or absence of missing items, the conformity of the shape / size / accommodation of the contents, and the distribution of density / thickness / volume or mass.

[0034] The conveying unit 10 is capable of winding the annular conveyor belt 11 onto multiple conveyor rollers 12 and 13 and conveying the item P through the upward section 11a of the conveyor belt 11. Figure 1 The conveyors that transport goods sequentially in the right direction are supported by a frame (not shown).

[0035] Inspection section 20 is an X-ray inspection section, which has an X-ray generator 21 (X-ray source) that generates X-rays that transmit through a specified energy band of the article P transported by the transport section 10 and an X-ray detector 23 disposed directly below the upward section 11a of the conveyor belt 11.

[0036] The X-ray generator 21 generates X-rays of wavelength and intensity corresponding to the tube current and tube voltage through a known X-ray tube 22, and is able to irradiate the article P in a predetermined inspection zone Zx on the conveyor belt 11 with fan-shaped beams of X-rays that extend in a direction orthogonal to the article transport direction of the conveyor section 10 through the X-ray window of the peripheral device (details not shown).

[0037] Although the detailed structure is not shown, the X-ray detector 23 is, for example, composed of an X-ray linear sensor camera and is positioned at a predetermined position in the transport direction corresponding to the X-ray irradiation position from the X-ray generator 21. The X-ray linear sensor camera is equipped with detection elements consisting of scintillators as phosphors and photodiodes or charge-coupled elements arranged in an array at predetermined intervals in the width direction of the transport path of the transport section 10, and performs X-ray detection at a predetermined resolution.

[0038] That is, the X-ray detector 23 can detect X-rays that have been irradiated by the X-ray generator 21 and transmitted through the article P in each specified transmission area corresponding to the detection element, convert them into an electrical signal corresponding to the transmission amount of the X-rays, and output an X-ray detection signal for generating an X-ray transmission image with the transmission direction of the X-rays as the observation direction.

[0039] Furthermore, the X-ray detector 23 performs a width scan corresponding to the conveying speed of the conveyor belt 11 and outputs X-ray detection signals sequentially. If the item detection sensor 28 detects that the item P is put onto the conveyor belt 11 and is conveyed toward the specified inspection interval Zx, the X-ray detection signal of the item P detected after a specified time is output.

[0040] The control unit 30 is an inspection control mechanism that controls the X-ray irradiation intensity or irradiation period in the inspection unit 20, or controls the X-ray detection cycle and detection period of each item P in the X-ray linear sensor of the X-ray detector 23 corresponding to the conveying speed of the item P, and also serves as a conveying control mechanism that controls the conveying speed or conveying interval of the item P based on the conveyor belt 11 in the conveying unit 10, but a detailed illustration of the conveying control mechanism is omitted.

[0041] The control unit 30 is configured to include, for example, a microcomputer (processor) having a CPU, ROM, RAM and I / O interface (not shown); a program device that readablely stores control programs for performing the functions of the multiple functional units described later in the ROM or auxiliary storage device and other recording media or downloads them from other computers via data communication; and a timer circuit, etc., and according to the control program stored in the ROM, etc., the CPU performs prescribed arithmetic processing while sending and receiving data with the RAM, etc., and executes the control programs of the multiple functional units.

[0042] Specifically, the control unit 30 includes an inspection image acquisition unit 31 that sequentially reads detection data Lx from the X-ray detector 23 for each predetermined period and acquires and outputs X-ray imaging data Dpx (hereinafter also referred to as imaging data Dpx) corresponding to the dose distribution of X-rays transmitted through the article P, an inspection processing unit 32 that performs a predetermined inspection based on the imaging data Dpx, and an image processing algorithm setting unit 35 that can be set according to the selected type of article P to the image processing algorithm used in the inspection processing unit 32.

[0043] Here, we will give an example of the case where the image data Dpx of each item P output from the image acquisition unit 31 is a constant radiation quality (energy, wavelength) determined according to the quality of item P. However, it is also possible to output a so-called dual-energy or multi-energy X-ray image with different X-ray radiation quality.

[0044] The inspection processing unit 32 includes: an image processing unit 33, which performs image analysis processing such as setting one or more predetermined filtering processes that allow the extraction of image features by reading the camera data Dpx output from the inspection image acquisition unit 31 and performing feature measurement to calculate the feature quantity of the extracted image features; and a judgment unit 34, which performs judgment processing to determine the quality status of the article P based on the feature quantity data extracted and measured by the image processing unit 33, such as whether there are foreign objects mixed in, whether there are missing items, and whether the shape, size, or container status of the contents is qualified.

[0045] The image processing algorithm setting unit 35 and the inspection processing unit 32 are connected for data communication, and the image processing algorithm used in at least the image processing unit 33 (including the determination processing algorithm used in the determination unit 34) can be updated / changed. Furthermore, the inspection processing unit 32 displays the determination result from the determination unit 34 to the operation display unit 41 (display). If there is an operation input requesting the selection or switching of a product from the operation display unit 41, it requests the image processing algorithm setting unit 35 to download and / or set a specific image processing algorithm Pgm corresponding to the requested product from the operation display unit 41.

[0046] More specifically, the image acquisition unit 31 is an image input unit that performs A / D conversion on X-ray detection signals from multiple detection elements of the X-ray detector 23, and outputs data of the cumulative transmittance per unit time for all detection element regions of the n (n is an integer greater than 1, for example, 640) detection element regions, for example as digital data representing the concentration level of gray from 0 to 1023 (hereinafter referred to as line scan).

[0047] Furthermore, the inspection image acquisition unit 31 has a data processing program and a working memory (not shown) that perform the following functions: when the line scan based on the X-ray detector 23 is repeated only a predetermined number of times corresponding to the inspection period of the article P, the image data Dpx of the article P is generated based on the detection data Lx of the line scan image sequentially written into the image memory, and output to the image processing unit 33 and the inspection image storage unit 51.

[0048] The image processing unit 33 of the inspection processing unit 32 can update and switchably set a specified image processing algorithm that combines image processing filters, etc., so as to perform a specified item inspection based on the camera data Dpx of the item P read from the inspection image acquisition unit 31.

[0049] The image processing filter included in the image processing algorithm of the image processing unit 33 is a processing procedure for extracting image features (e.g., edges, lines, corners, regions, density, texture) required for specified item inspection based on the photographic data Dpx of the item P. When the image processing algorithm includes a foreign object detection filter, it may be a feature extraction filter that performs edge detection processing to emphasize the outline of a foreign object in the item P. For example, it may be a differential filter such as a Sobel filter that performs differential processing based on a specified formula on the vicinity of the pixel of interest to emphasize the edge of the foreign object. Furthermore, the term "image processing filter" refers to preprocessing such as density correction or noise removal of the photographic data from the inspection image acquisition unit 31 to improve the detection accuracy of image features.

[0050] Furthermore, the feature measurement of the image features performed by the image processing unit 33 refers to the preprocessing or image processing required for the image of the photographic data Dpx of the item P read from the inspection image acquisition unit 31. For example, by performing the calculation of attributes (feature quantities that assign features to edges, regions, distances, positions, shapes, etc.) related to density features, color features, shape features, etc., the calculation of feature quantities representing the spatial relationship between such features, or the calculation of texture feature quantities related to spatial frequency distribution and directional components, the processing of the feature quantities required for the determination processing in the determination unit 34 is performed.

[0051] The determination unit 34 detects feature shapes or foreign objects detected in the article P based on the feature quantities extracted and measured by the image processing unit 33, or compares the area, outline length, concentration and other feature quantities of the detected object with the prescribed determination benchmark value, i.e., the limit, thereby performing determination processing on whether foreign objects or local feature shapes equivalent to defective parts that meet the determination conditions for defect determination are included in the article P.

[0052] Thus, the inspection and processing unit 32 applies a prescribed image processing algorithm, which combines multiple filtering processes, to the X-ray inspection image data Dpx of the article P that has undergone the prescribed type of processing or the X-ray inspection image data that has undergone the required preprocessing and image processing, and determines and inspects the prescribed quality status of the article P from the image.

[0053] The image processing algorithm setting unit 35 pre-stores at least a plurality of image processing algorithms for image determination that can be executed in the image processing unit 33 (and may also include determination processing algorithms that can be executed in the determination unit 34). Furthermore, the image processing algorithm setting unit 35 has the function of updating the image processing algorithm used in the determination unit 32 to a specific image processing algorithm Pgm corresponding to the new set variety when a switch / update to another variety is requested from the inspection processing unit 32 based on a variety selection or switching operation from the operation display unit 41.

[0054] The control unit 30 also has a first control function that operates as the aforementioned inspection control mechanism and transport control mechanism, and also has a second control function that confirms whether the inspection processing in the inspection processing unit 32 (i.e., the image processing in the image processing unit 33 and the determination processing in the determination unit 34) is operating normally during the operation of the item inspection device 1. Here, it is configured such that the operation confirmation control unit 50 that performs the second control function is built in.

[0055] The action confirmation control unit 50 can, for example, execute the judgment processing in the judgment and inspection processing unit 32 in real time to determine whether the action is normal.

[0056] The action confirmation control unit 50 is configured to include: an inspection image storage unit 51, which sequentially stores and holds the camera data Dpx of each item P acquired sequentially by the inspection image acquisition unit 31 as inspection images; an NG feature storage unit 52, which pre-stores image data of NG features that assign a defective quality state to item P; an action confirmation image generation unit 53, which performs digital image synthesis (hereinafter referred to as synthesis) on the inspection images of each item P from the inspection image storage unit 51 and the NG features stored in the NG feature storage unit 52 to generate image data including the action confirmation image Dpc containing the NG features; and an action confirmation processing unit 54, which performs the same inspection processing as the image processing of the image processing unit 33 and the determination processing of the determination unit 34 on the action confirmation image Dpc containing the above-mentioned NG features, and confirms whether the inspection processing in the inspection processing unit 32 is normal.

[0057] Here, in order to create an image Dpc for action confirmation including NG features, the camera data Dpx synthesized to each item P, i.e., the NG features in the inspection image, refer to image features that indicate that the quality or physical quantity of the item P required as a product is inappropriate. For example, image features that indicate defects such as foreign matter, lack of items, defects such as the shape / size / containment state of the contents, and defects such as the distribution of density / thickness / volume or mass.

[0058] In addition, in the following description, the NG feature detected by the motion confirmation processing unit 54 will be referred to as "NG detection", and the NG feature not detected by the motion confirmation processing unit 54 will be referred to as "NG not detected".

[0059] When the inspection processing unit 32 switches to a specific image processing algorithm Pgm corresponding to the set variety through the image processing algorithm setting unit 35, the action confirmation processing unit 54 reads the specific image processing algorithm Pgm from the image processing algorithm setting unit 35 in parallel.

[0060] Furthermore, during the operation of the item inspection device 1, in which the inspection processing unit 32 performs inspection processing based on the camera data Dpx of each item P and outputs the inspection result to the operation display unit 41, the action confirmation control unit 50 can, in parallel with this operation, use the image data of the camera data Dpx of each item P and the specified NG feature image Cn to generate an action confirmation image Dpc using the image data of the image data Dpx of each item P and the specified NG feature image Cn, and then perform the same image processing and judgment processing on the action confirmation image Dpc as the inspection processing in the inspection processing unit 32 using the action confirmation processing unit 54.

[0061] That is, the motion confirmation image generation unit 53 generates a motion confirmation image Dpc based on the inspection images, i.e., the camera data Dpx of each item P, during the operation of the item inspection device 1. This image includes at least the defective feature (NG feature) of the quality state of item P, which is determined to be defective. This image is used as the motion confirmation image in the motion confirmation processing unit 54.

[0062] Furthermore, during the operation of the item inspection device 1, which operates the image processing unit 33, the determination unit 34, and the operation display unit 41, the action confirmation processing unit 54, in parallel with the operation of the image processing unit 33 and the determination unit 34 of the inspection processing unit 32, uses the prescribed image processing algorithm Pgm and the image data Dpx of each item P to synthesize the image data of the action confirmation image Dpc with NG characteristics. At the same time, it performs the same image processing and determination processing as the inspection processing in the inspection processing unit 32. As a result, it is possible to reliably determine whether the processing result is NG detection when foreign objects with NG characteristics are mixed in or other abnormalities are generated in each inspection image Dpx processed by the inspection processing unit 32.

[0063] In this embodiment, the inspection image, i.e., the photographic data Dpx of each item P, is an X-ray transmission image obtained by photographing item P with X-rays. Therefore, as... Figure 2 As shown, during the operation of the item inspection device 1, the action confirmation image generation unit 53 synthesizes multiple partial images, including NG feature parts C1, C2, C3, C4, C5, and C6 (partial images of defective feature parts), from multiple parts with different image features in the inspection image, i.e., the camera data Dpx of each item P. For example, it performs pixel value synthesis and conversion according to a prescribed formula to generate an image of suspected defective product, i.e., the action confirmation image Dpc.

[0064] In this case, the action confirmation image generation unit 53 can generate an image of suspected defective product, namely the action confirmation image Dpc, by synthesizing multiple NG feature units C1, C2, C3, C4, C5, and C6 of the product P whose quality status is determined to be qualified product in the determination unit 34 during the operation, and the product P whose quality status is determined to be defective.

[0065] However, the motion confirmation image generation unit 53 does not necessarily need to synthesize NG feature images Cn for each item P in real time to create a simulated defective image, i.e., a motion confirmation image Dpc. It does not need to synthesize NG feature images Cn from the camera data Dpx of the defective product; instead, it can directly output the image obtained from the defective product determination as the motion confirmation image Dpc and execute the motion confirmation processing in the motion confirmation processing unit 54. In this case, the location information of the defective portion of the defective product can be stored and used as the result of motion confirmation. Preferably, the motion confirmation processing is performed after the limits of the determination unit 34 change. Furthermore, if the inspection image DPx obtained from the defective product determination is an image of a product with foreign objects, the motion confirmation processing unit 54 can use the data output from the determination unit 34 to create the foreign object location information of the image of the product with foreign objects.

[0066] Furthermore, the motion confirmation image generation unit 53 does not necessarily need to output motion confirmation images Dpc for all items P. After sequentially creating simulated defective images (i.e., motion confirmation images Dpc) by synthesizing NG feature images Cn from the video data Dpx of the specified number of items P that have passed the quality inspection, it can directly output the motion confirmation image Dpc as the motion confirmation image Dpc without synthesizing NG feature images Cn from the next video data Dpx that has passed the quality inspection or defective inspection, and then perform the motion confirmation processing in the motion confirmation processing unit 54. In addition, it is of course possible to establish a correspondence between each item P and its video data Dpx based on the time of inspection of each item P and the time of the video data Dpx, or based on other identification information.

[0067] In this embodiment, the product that becomes the object of article P is not limited to products with the same thickness and composition, but is a product whose brightness varies depending on the shape, composition, or location, as well as the brightness of the X-ray transmission image, i.e., the photographic data Dpx. Specifically, article P is, for example, a food product or a product containing a variety of ingredients cooked as contents in a package, container, or bag of a specified shape, with a specified net content or weight.

[0068] In the photographic data Dpx of article P, several areas with different image characteristics are included, such as between multiple contents, near the edges of contents, areas where the brightness of the inspection image corresponding to the X-ray transmittance of the contents is relatively lower than the average brightness of the surrounding area, and areas where the brightness of the inspection image corresponding to the X-ray transmittance of the contents is relatively higher than the average brightness of the surrounding area. Previously, when an operator attached a sample (test piece) of the defective features to be inspected to article P for operational verification, these areas with different image characteristics were considered preferred locations for photographing, representing the positions of the sample containing the defective features.

[0069] As described above, in the image data Dpx obtained by X-ray imaging of the article P of this embodiment, which has brightness deviations, for example, when foreign matter is mixed in, the brightness will change depending on the location of the foreign matter, even if it is the same foreign matter. Consequently, the detection sensitivity will also change depending on the influence of the brightness of the surrounding products. In this case, if the foreign matter or a test piece equivalent to the foreign matter is manually attached to the product that becomes article P to collect X-ray images, it is sometimes difficult to know the location of the contents from the appearance of the product, such as in packaged products or products with internal cavities. It is often difficult to collect images for action confirmation based on a uniform change in the location of the foreign matter.

[0070] Therefore, as Figure 2 As illustrated, in the action confirmation image Dpc of this embodiment, for the image data Dpx obtained by taking X-ray pictures of each article P, multiple NG feature parts C1, C2 located between or near the edges of multiple contents of article P, multiple NG feature parts C3, C5 located in the part of the inspection image with relatively low brightness relative to the surroundings corresponding to the X-ray transmission amount of the contents, and multiple NG feature parts C4, C6 located in the part of the inspection image with relatively high brightness relative to the surroundings corresponding to the X-ray transmission amount of the contents, are combined into an NG feature image Cn.

[0071] That is, in this embodiment, when synthesizing NG feature images Cn from the image data Dpx obtained by X-ray imaging of each article P, by uniformly distributing multiple NG feature parts C1 to C6 on multiple parts with different image features, such as the high-brightness part, low-brightness part, and between adjacent contents or near the edges of contents of the product that is the article P, it is not necessary to perform the previous operation of attaching foreign object samples or test pieces to the product of the article P for image acquisition. Instead, it is possible to quickly / accurately and in real time produce an action confirmation image Dpc that is more effective than image judgment.

[0072] In the motion confirmation image Dpc, the motion confirmation image generation unit 53 sets rules for synthesizing / configuring multiple NG feature parts C1 to C6 at multiple locations with different image features based on the image data Dpx obtained by X-ray imaging of the product that becomes article P. The motion confirmation image generation unit 53 can easily create images for AI learning or sensitivity verification, for example.

[0073] Furthermore, the composite position of the multiple NG feature parts C1 to C6 of the item P at this time is, for example, part of multiple small image regions of a predetermined size (number of pixels) whose image features differ from each other, such as the bright and dark parts of the product, the vicinity of the product edges, and the areas between adjacent contents within the product. These are extracted as multiple item image regions with image features suitable for NG feature synthesis using a search size with a preset number of pixels. On the other hand, the composite position of the multiple NG feature parts C1 to C6 themselves can, for example, be set to the centroid position of each NG feature image.

[0074] The number or position of the composites in the NG feature units C1 to C6 can be changed according to the number or category of multiple small image regions with different image features. However, at least one image composite position can be set in multiple small image regions with common image features, and at least one image composite position can be set in multiple small image regions with different image features.

[0075] Specifically, the action confirmation control unit 50 synthesizes the digital X-ray image Dpx obtained by taking an X-ray image of a product that is currently set as an article P into an NG feature image Cn that is equivalent to a foreign object, which is pre-stored in the NG feature storage unit 52, and makes the conditions for X-ray taking and X-ray detection (hereinafter referred to as X-ray detection conditions) in the image data Dpx and the NG feature image Cn that is equivalent to a foreign object substantially consistent or effectively approximate.

[0076] That is, in the inspection image storage unit 51, the image data Dpx of the product of the currently set type of article P is associated with and stored in the X-ray detection conditions, such as the staged setting values ​​of the tube voltage and tube current of the X-ray tube 22 corresponding to the X-ray output (tube voltage × tube current) of the X-ray generator 21. On the other hand, in the NG feature storage unit 52, multiple NG feature images Cn corresponding to the same foreign object sample are generated for each foreign object sample under different X-ray detection conditions, such as the staged setting values ​​of the tube voltage and tube current of the X-ray tube of the X-ray generator 21. These images are associated with each X-ray detection condition and stored in the NG feature storage unit 52. Thus, when selecting an NG feature image Cn, it is possible to match or effectively approximate the X-ray detection conditions of the image data Dpx of article P.

[0077] Furthermore, considering the difference between the X-ray detection conditions of the photographic data Dpx of article P and the X-ray detection conditions of the NG feature images Cn such as test pieces stored in the NG feature storage unit 52, the image density of the reference NG feature image Cn stored in the NG feature storage unit 52 can be made substantially the same as the image density when photographed under the same X-ray detection conditions as the photographic data Dpx of article P. This allows for density conversion processing of the NG feature image Cn, thus enabling handling of various X-ray detection conditions. In any case, the difference in variety is directly reflected in the photographic data Dpx of article P, while the NG feature image Cn is unaffected by the variety.

[0078] Furthermore, the NG feature image Cn, which corresponds to a foreign object, can be pre-stored in the NG feature storage unit 52 before the operation of the item inspection device 1 begins, for example, as an image of the foreign object as a test piece (which may be during production). The user of the item inspection device 1 can also create an NG feature image Cn corresponding to a known foreign object sample by taking a picture of it with the inspection unit 20 and detecting it with X-rays. Of course, the NG feature storage unit 52 can read the NG feature image Cn as image data from a removable recording medium, or obtain a known NG feature image Cn from other item inspection devices or a management computer installed within the company via data communication. It goes without saying that the NG feature image Cn is not limited to an image corresponding to a foreign object.

[0079] When determining the locations of multiple parts with different image features (the composite location of the NG feature image Cn) in the image data Dpx obtained by taking X-rays of a product that becomes an article P, features of each part of the image region of the product that becomes an article P, such as the high-brightness and low-brightness parts, the vicinity of the product's edges, and the contents of the product, can be extracted by using the set of brightness differences in the image as feature quantities. Alternatively, features can be extracted by focusing on the distribution of brightness gradient directions in the image. Furthermore, the extraction can be performed by using a mechanical method that automatically determines the conversion process from image to feature quantity by a computer.

[0080] Here, the X-ray transmission image of article P, as an image corresponding to the distribution of X-ray absorption in article P, can be obtained by taking the difference between the logarithmically transformed values ​​of the X-ray transmission amount detected when article P is not on conveyor belt 11 and the X-ray transmission amount detected when article P is on conveyor belt 11. Similarly, the NG feature image Cn, such as that of foreign objects, can also be obtained by taking the difference between the logarithmically transformed values ​​of the X-ray transmission amount detected when article P or foreign objects are not on conveyor belt 11 and the X-ray transmission amount detected when foreign objects are on conveyor belt 11. The operation steps for acquiring this image can be set to be approximately the same as those described in paragraphs 0036 to 0046 of Japanese Patent Application Publication No. 2009-168740.

[0081] For example, the wavelength of X-rays depends on the tube voltage of X-ray generator 21 (the voltage applied to the X-ray tube), and the amount of X-ray irradiation depends on the tube current of X-ray generator 21. Therefore, the shorter the wavelength of X-rays (the higher the tube voltage), the smaller the amount of X-ray absorption. Moreover, the X-ray transmission image or X-ray absorption image of object P represents a two-dimensional distribution of the amount of X-rays transmitted through the object or a two-dimensional distribution based on the amount of X-ray absorption of the object. If the amount of X-ray irradiation is set as Io (the intensity of X-rays), the amount of X-ray transmission is set as I (the intensity after transmission through the object), the X-ray absorption rate is set as μ, the thickness of the transmitted object is set as d, and log is set as the natural logarithm, then according to Beer-Lambert's attenuation law (I / Io = e^(-μd)), the amount of X-ray absorption T holds the following equation (1).

[0082] T=(log Io-log I)=μd...Equation (1)

[0083] Furthermore, if the wavelength of the X-ray is set as λ, the density of the object is set as ρ, the atomic number is set as Z, and the constant is set as C, then the absorption rate μ of the X-ray has the following relationship (2).

[0084] μ=λ 3 ρZC……Equation (2)

[0085] Equation (1) indicates that the X-ray absorption is the difference between the X-ray irradiation Io (after logarithmic transformation) and the X-ray transmission I, where the X-ray irradiation Io is the X-ray transmission when the X-ray absorption is zero. That is, the X-ray transmission detected when there is no transported object (i.e., the object being inspected) on the conveyor belt is called the X-ray irradiation Io. Therefore, from equations (1) and (2), it can be seen that by changing the wavelength of the X-rays, the degree of variation in X-ray absorption varies for each item (item P and foreign object) with different densities and atomic numbers. In the item inspection device 1, this can be used to set the X-ray output conditions (tube voltage and tube current of the X-ray generator 21) in a way that makes item P and foreign object appear different.

[0086] Furthermore, if the image of the foreign object sample (X-ray absorption image of the foreign object) detected based on the specified X-ray detection conditions (tube voltage and tube current of X-rays) stored in the NG feature storage unit 52 is set as S1, and the X-ray absorption image of the article P obtained based on the X-ray detection conditions (tube voltage and tube current of X-rays) of the X-ray generator 21 set corresponding to the article P is set as S2, then the composite image S becomes as shown in the following formula (3).

[0087] S = αS1 + S2 (where α is the conversion factor) ... Equation (3).

[0088] The conversion factor α in the above formula (3) is, for example, the ratio between the intensity of the foreign object sample image obtained according to each X-ray detection condition (tube voltage and tube current of X-ray) and the intensity of the NG feature image Cn of the foreign object sample detected based on the specified X-ray detection conditions (tube voltage and tube current of X-ray tube 22) stored in the NG feature storage unit 52, and can be stored in the conversion table according to each X-ray detection condition (tube voltage and tube current of X-ray).

[0089] The NG feature storage unit 52 stores X-ray images of multiple foreign object samples of different sizes detected based on preset X-ray detection conditions as various NG feature images Cn, such as foreign object images C1 to C6.

[0090] Furthermore, the NG feature storage unit 52 has a storage mechanism that stores NG feature images Cn for each type of foreign object, i.e., each foreign object sample of each shape (including size) or material (e.g., metal, glass, etc.), when irradiated with X-rays under the aforementioned X-ray detection conditions. The NG feature image Cn consists of the following images: an X-ray absorption image converting the X-ray transmittance to X-ray absorption when irradiating a single foreign object under pre-set X-ray detection conditions; and an X-ray absorption image converting the X-ray transmittance to X-ray absorption when irradiating foreign objects of different sizes, shapes (e.g., cubes, spheres, lines, etc.) and of the same material (e.g., metal, glass, resin, bone, etc.) placed at predetermined intervals under specified X-ray detection conditions.

[0091] Thus, the control unit 30 includes an inspection processing unit 32, which controls the operation of the image processing unit 33, the determination unit 34, and the operation display unit 41 to determine the quality status of each item P as a first control function unit, and also includes an action confirmation control unit 50, which controls the operation of the action confirmation image generation unit 53 and the action confirmation processing unit 54 to determine whether the operation of the image processing unit 33 and the determination unit 34 is normal during the period when the inspection processing unit 32 performs the first control function.

[0092] The labels of each image processing algorithm stored in the image processing algorithm setting unit 35 include at least the number (identification number) of the image processing algorithm, but may also include the name of the algorithm, the name of the variety of the object to be inspected, or the inspection items of the inspection content in the algorithm.

[0093] Furthermore, the image processing algorithm setting unit 35 may include: an algorithm storage unit that pre-stores multiple image processing algorithms; and an algorithm setting unit capable of performing update setting processing to update the image processing algorithm set and stored in the image processing unit 33 to any image processing algorithm Pgm stored in the algorithm storage unit.

[0094] Next, the actions will be explained.

[0095] In the article inspection apparatus of this embodiment configured as described above, before the operation of the article inspection apparatus 1 begins or during the period when article inspection is stopped, the NG feature storage unit 52 is operated as follows: NG feature images Cn are read from a recording medium, or known NG feature images Cn are obtained from other article inspection apparatuses or management computers installed within the company via data communication, thereby pre-storing existing test piece images as NG feature images Cn in the NG feature storage unit 52. It is self-evident that additional storage operations for NG feature images Cn in the NG feature storage unit 52 are possible.

[0096] Next, if the operation of the item inspection device 1 for inspecting item P begins, the item inspection of the sequentially inserted item P is performed while the inspection screen is displayed on the operation display unit 41. On the other hand, if... Figure 3 The action confirmation process shown is usually not displayed on the screen, but is performed as a background process in parallel with the item inspection process.

[0097] like Figure 3 As shown, firstly, the image processing algorithm used in the operation of the item inspection device 1 is determined by using an algorithm number corresponding to the currently set variety (step S11).

[0098] Next, based on the X-ray detection conditions (X-ray output of X-ray generator 21 (tube voltage × tube current)) at the time of imaging the image data Dpx of the item P input to the inspection image storage unit 51, in order to synthesize any one of the multiple NG feature images Cn of the image data Dpx of the item P pre-stored in the NG feature storage unit 52 for the foreign object image group of the various foreign object samples stored in the NG feature storage unit 52, any one of the multiple NG feature images Cn whose X-ray detection conditions are substantially consistent with or effectively approximate with the X-ray detection conditions at the time of imaging the image data Dpx of the item P is extracted from the multiple NG feature images Cn whose X-ray detection conditions are different from each other at multiple stages, and is used as the NG feature image for synthesis (step S12).

[0099] Next, the detection signal from the item detection sensor 28 is detected, and the camera data Dpx of item P is output from the inspection image acquisition unit 31 (or, further, by checking the items that have completed a predetermined number of inspections or by the passage of a predetermined time, etc.) to determine whether the conditions for confirming the normal operation of the inspection processing in the inspection processing unit 32 are met (step S13).

[0100] At this time, if the action confirmation condition is met (the case where "yes" is met in step S13), the image generation unit 53 performs action confirmation on the image data Dpx of the item P input to the inspection image storage unit 51. Figure 2 Multiple parts of the image of the object P that have different image features are combined into an NG feature image Cn under the same X-ray detection conditions as when it was photographed, for example, as multiple foreign object images C1 to C6, and an action confirmation image Dpc is generated (step S14).

[0101] Next, during the operation of the item inspection device 1, which operates in the image processing unit 33, the determination unit 34, and the operation display unit 41, the same image processing (step S15) is performed simultaneously with the operation of the image processing unit 33 and the determination unit 34 of the inspection processing unit 32. This is done by using the image data of the NG feature image Cn, which is synthesized from the prescribed image processing algorithm Pgm and the camera data Dpx of each item P.

[0102] Next, based on the image processing result, an action confirmation process (step S16) is performed to determine whether the processing result of NG detection can be reliably obtained through the same determination process as determination unit 34.

[0103] Furthermore, if the processing result is determined to be "NG detection" (the case of "0" in step S16), then, based on checking whether the operation of the item inspection device 1 has stopped (step S17), the processing after step S13 is repeated until the operation of the item inspection device 1 stops.

[0104] On the other hand, if the processing result is determined to be "NG not detected" (the case of "×" in step S16), then, while the inspection image of item P is displayed on the operation display unit 41, a warning is issued that a certain defect may occur during the operation of the item inspection, or that "NG not detected" will occur again within a certain inspection period (prescribed number of inspections), and a warning output is issued that the operator needs to perform manual action confirmation or inspection of the judgment standard (step S18). Then, the action stop confirmation in step S17 and / or the processing after step S13 are repeated.

[0105] In this embodiment, the inspection processing unit 32 determines whether the inspection processing is a normal operation according to each prescribed action confirmation cycle, and performs the determination processing in real time sequentially or according to each prescribed number of inspections.

[0106] Next, the function will be explained.

[0107] In this embodiment, during the operation of the item inspection device 1, the action confirmation processing in the action confirmation control unit 50 is executed in parallel with the operation of the image processing unit 33 of the inspection processing unit 32, and the action confirmation processing unit 54 of the action confirmation control unit 50 determines whether the operation of the image processing unit 33 and the determination unit 34 is normal based on the inspection image, i.e., the camera data Dpx, during the operation.

[0108] At this time, the motion confirmation image Dpc generated by the motion confirmation image generation unit 53 is used by the motion confirmation processing unit 54 as a defective image of multiple parts with different image features, which are synthesized from the NG feature image Cn, i.e., defective feature parts C1 to C6, into the camera data Dpx of the article P. Therefore, during operation, the motion confirmation image Dpc, which is generated based on the camera data Dpx (inspection image) of the article P that is mainly qualified, including defective feature parts C1 to C6, is used to confirm whether the motion confirmation image Dpc is a defective judgment. This allows it to determine whether the operation of the image processing unit 33 and the judgment unit 34 of the inspection processing unit 32 is normal. As a result, without transitioning the article inspection device 1 to a motion confirmation mode or other inspection stop state, it is possible to accurately and in real time, sequentially or according to a predetermined number of inspections, simultaneously perform motion confirmation on whether the detection signals from the inspection unit 20 are appropriate during its operation.

[0109] Furthermore, in this embodiment, the motion confirmation image generation unit 53 synthesizes NG feature parts C1 to C6 (multiple partial images of defective feature parts) from multiple parts with different image features in the inspection image based on the camera data Dpx during the operation of the article inspection device 1, thereby generating a motion confirmation image Dpc. Therefore, the motion confirmation image Dpc is less affected by deviations in the quality state of the article P or deviations in the X-ray transmission amount.

[0110] Furthermore, in this embodiment, the motion confirmation image generation unit 53 synthesizes an image of multiple NG feature sections C1 to C6, which are determined as defective features by the image data Dpx (qualified inspection image) obtained from the quality status of the item P by the determination unit 34 during operation, into a defective image, namely, the motion confirmation image Dpc. Therefore, during normal operation, the motion confirmation image Dpc, which synthesizes the defective feature section from the qualified inspection image, is generated. By stably ensuring the defective features in the motion confirmation image, stable motion confirmation can be performed. Furthermore, when a defective item is generated, the inspection image of the defective item can also be used for motion confirmation.

[0111] Furthermore, in this embodiment, a control unit 30 is provided, which has a first control function unit that controls the operation of the image processing unit 33, the determination unit 34, and the operation display unit 41 to determine the quality status of each item P. This control unit 30 also includes an action confirmation control unit 50, which, during the operation of the first control function, controls the operation of the action confirmation processing unit 54 and the action confirmation image generation unit 53 to determine whether the operation of the image processing unit 33 and the determination unit 34 is normal. Therefore, during the operation of determining the quality status of item P by controlling the operation of the image processing unit 33, the determination unit 34, and the operation display unit 41 through the first control function unit, i.e., the inspection processing unit 32, the control unit 30 can simultaneously execute action confirmation processing, which controls the operation of the image processing unit 33 and the determination unit 34 to determine whether the operation of the image processing unit 33 and the determination unit 34 is normal, by controlling the operation of the action confirmation image generation unit 53 and the action confirmation processing unit 54 through the second control function unit, i.e., the action confirmation control unit 50, even when the display screen is hidden during the operation. Furthermore, when the action confirmation result becomes abnormal, it can output notifications such as warning display or warning sound.

[0112] Furthermore, in this embodiment, the inspection image based on the camera data Dpx is an X-ray transmission image obtained by taking an X-ray of an item P. Therefore, in the action confirmation image generation unit 53, the camera data Dpx of each item P in operation, i.e., each inspection image, is used to synthesize multiple NG feature parts C1 to C6 as NG feature images Cn, thereby making it easy to generate an action confirmation image Dpc of defective product images.

[0113] Thus, in this embodiment, an item inspection device 1 can be provided that can perform action confirmation, such as whether the detection signal from the detection unit is appropriate, during operation without transitioning to an inspection stop state such as an action confirmation mode.

[0114] Furthermore, as described above, the action confirmation processing unit 54 in this embodiment can use the same prescribed image processing algorithm as the determination unit 34, and in the operation of the article inspection device 1, it performs in parallel the action confirmation determination of whether the inspection image of the article P, which has been determined to be a qualified article by the determination unit 34, is a qualified article, and the action confirmation determination of whether the action confirmation image Dpc of the defective article is a defective article. Moreover, the action confirmation processing unit 54 can generate a defective image Dpc from a predetermined number of inspection images among multiple inspection images of the article P, and without generating defective images for the remaining inspection images, it uses the action confirmation image Dpc equivalent to a qualified article to perform action confirmation processing to determine whether it is an undetected NG (Not Found).

[0115] exist Figures 1 to 3In one embodiment shown, the action confirmation control unit 50 is built into the control unit 30, but the action confirmation control unit 50 may also be provided as a control device or management device independent of the control unit 30. This other embodiment will be described next.

[0116] (Other implementation methods)

[0117] Figure 4 An article inspection device according to another embodiment of the present invention is shown.

[0118] In addition, Figure 4 In the middle, to and Figure 1 The same structures shown in one embodiment are represented by the same symbols, and repeated detailed descriptions are omitted.

[0119] like Figure 4 As shown, the article inspection device 2 of this embodiment includes: a conveying unit 10 and an inspection unit 20, configured in the same manner as in the first embodiment; and an inspection control unit 130 and an action confirmation control unit 150, configured to communicate data with each other and to perform functions similar to those of the control unit 30 in the first embodiment, and to be configured independently.

[0120] The inspection control unit 130 has a hardware structure similar to that of the control unit 30 in one embodiment. As multiple functional units, it includes: an inspection image acquisition unit 31, which sequentially reads detection data Lx from line scan images of each predetermined period from the X-ray detector 23, acquires and outputs imaging data Dpx corresponding to the dose distribution of X-rays transmitted through the article P; and an inspection processing unit 132, which performs a predetermined inspection based on the imaging data Dpx, and appends the inspection results of each article P as attribute information to the inspection images, i.e., the imaging data Dpx, of each article P as imaging data based on the inspection results, which are then designated as either qualified product imaging data Dpx1 or defective product imaging data Dpx2. Output to the operation display unit 41 and the action confirmation control unit 150; the image processing algorithm setting unit 35 can variably set the image processing algorithm used in the inspection processing unit 132 according to the selected variety of item P; the inspection image storage unit 51 stores qualified product images with the same hardware structure as the control unit in one embodiment; and the action confirmation processing unit 54, with the same hardware structure as the control unit in one embodiment, reads the specific image processing algorithm Pgm from the image processing algorithm setting unit 35 in parallel with the inspection processing unit 32 when it switches to the specific image processing algorithm Pgm corresponding to the set variety through the image processing algorithm setting unit 35.

[0121] The action confirmation control unit 150 includes: a qualified product image database 151a, which stores image data Dpx1 of qualified products from the inspection processing unit 132 based on the inspection results; an NG image database 151b, which stores image data Dpx2 of defective products from the inspection results; and a foreign object image database 151c, which reads and stores NG feature images, i.e., foreign object images Dsc1, Dsc2, and Dsci, via downloading from a recording medium or network. Additionally, in Figure 4 In Chinese, the database is denoted as DB.

[0122] The motion confirmation control unit 150 further includes: an NG feature learning unit 152, which learns NG features suitable for compositing into a qualified product image, such as features of a foreign object image, by reading image data from the qualified product image database 151a, the NG image database 151b, and the foreign object image database 151c, and generates an NG feature image generation model NM; and a motion confirmation image generation unit 153, which uses a qualified product image Dpx1 from the inspection image storage unit 51 of the inspection control unit 130 and the NG feature image generation model NM from the NG feature learning unit 152 to generate and output a motion confirmation image Dpc that is the same as the motion confirmation image generated by the motion confirmation image generation unit 53 in one embodiment.

[0123] Here, the NG feature image generation model NM generated by the NG feature learning unit 152 is, for example, a model that has learned NG features through two neural networks (generation and recognition) constituting a generative adversarial network (GAN), or a model that has learned NG features through a distributed model that uses two nodes on the network (each equipped with a GPU) to segment the task of the learning object. By using the NG feature image generation model NM, it is possible to provide the function of generating a new NG feature image Cn that does not exist in the foreign object image DB and the NG image DB, synthesized by the image generation unit 153 for action verification from the qualified product image Dpx1 from the inspection image storage unit 51. In addition, even if performance is maximized during the learning stage, and the AI ​​that synthesizes poorly from the qualified product image cannot output a realistic image, action verification can be performed using the actual defective product image captured.

[0124] In this embodiment, during the operation of the item inspection device 2, the action confirmation processing in the action confirmation control unit 150 is executed in parallel with the operation of the image processing unit 33 of the inspection processing unit 132. Based on the inspection image, i.e., the camera data Dpx, during the operation, the action confirmation control unit 150 and the action confirmation processing unit 54 of the inspection control unit 130 determine whether the operation of the image processing unit 33 and the determination unit 34 of the inspection processing unit 132 is normal.

[0125] At this time, the motion confirmation image generation unit 153 generates a motion confirmation image Dpc as the NG feature image Cn, for example... Figure 2 The defective feature sections C1 to C6 shown are synthesized into multiple parts with different image features in the image data Dpx of the article P, forming a defective product image. Therefore, by using the action confirmation image Dpc generated from the image data Dpx (inspection image) of the qualified article P, it is possible to confirm whether the action confirmation image Dpc is a defective determination, and accurately determine whether the operation of the image processing unit 33 and the determination unit 34 of the inspection processing unit 132 is normal. As a result, similar to the case of the first embodiment, it is possible to accurately and in real time confirm whether the detection signal from the inspection unit 20 is appropriate during its operation without transitioning the article inspection device 1 to an inspection stop state such as the action confirmation mode. Furthermore, the action confirmation image Dpc is less affected by deviations in the quality state of the article P or deviations in the X-ray transmission amount.

[0126] Furthermore, in the above embodiments, it is set to synthesize an NG feature image from the latest qualified product image to create an action confirmation image, and to perform action confirmation on whether the NG detection function of the item inspection device 1 is normal. However, if the inspection result is a qualified product, it can be set to store the qualified product image, and to perform action confirmation by creating an action confirmation image based on any stored qualified product image or the latest qualified product image at a predetermined time. Furthermore, the defective portion of the defective product image can be added to the NG feature storage unit and used for NG synthesis. In the case where the defective product image is an image of a product with foreign objects, it is preferable to provide a foreign object location information generation unit to add the foreign object image and its location information to the NG feature storage unit. Moreover, while this may lead to redundancy in processing, it is also possible to consider synthesizing an NG feature image to perform action confirmation processing regardless of whether it is a qualified or defective product image, and to determine if the action is defective by comparing the inspection result with the action confirmation result.

[0127] As explained above, the article inspection apparatus of the present invention provides an article inspection apparatus that reduces downtime for transitioning to an inspection stop state for action confirmation, and that can perform the action confirmation in parallel during the article inspection operation. This invention is useful in all article inspection apparatuses that apply a prescribed image processing algorithm to inspection images obtained by photographing a specified variety of inspected articles to check the quality status of the inspected articles.

[0128] Symbol Explanation

[0129] 1, 2 - Item inspection device; 10 - Conveying unit; 11 - Conveyor belt; 11a - Upward section; 12, 13 - Conveying rollers; 20 - Inspection unit; 21 - X-ray generator; 22 - X-ray tube; 23 - X-ray detector; 28 - Item detection sensor; 30 - Control unit; 31 - Inspection image acquisition unit; 32 - Inspection processing unit (inspection control unit); 33 - Image processing unit; 34 - Judgment unit; 35 - Image processing algorithm setting unit; 41 - Operation display unit (display unit); 50 - Action confirmation control unit (second control function unit); 51 - Inspection image storage unit (qualified product image storage unit); 52 - NG feature storage unit (feature storage unit for foreign objects and other defective parts); 53 - Action confirmation image generation unit; 54 - Action confirmation processing unit; 130 - Inspection control unit; 132 - Inspection processing unit; 150 - Action confirmation control unit; 151a - Qualified product image database (qualified product image database). Image DB), 151b-NG Image Database (NG Image DB), 151c-Foreign Object Image Database (Foreign Object Image DB), 152-NG Feature Learning Unit, 153-Action Confirmation Image Generation Unit, C1, C2, C3, C4, C5, C6-NG Feature Unit (multiple part images that become defective feature units), Cn-NG Feature Image (Foreign Object Image), Dsc1, Dsc2, Dsci-Foreign Object Image (NG Feature Image), Dpc-Action Confirmation Image, Dpx-Video Data (Inspection Image, X-ray Transmission Image, X-ray Photography Image Data), Dpx1-Video Data (Video Data of Qualified Products, Inspection Image, Qualified Product Image), Dpx2-Video Data (Video Data of Defective Products, Inspection Image, Defective Product Image), Lx-Detection Data, NM-NG Feature Image Generation Model, P-Item (Inspected Item), Pgm-Specific Image Processing Algorithm.

Claims

1. An item inspection device, characterized in that, have: The image processing unit (33) performs image processing using a prescribed image processing algorithm on the inspection image (Dpx) obtained by the camera mechanism (20) of the transported article (P) and outputs data for judging the quality status of the article. The determination unit (34) determines the quality status of the item based on the determination data. The display unit (41) displays the result determined by the determination unit; The action confirmation image generation unit (53) generates a defective image as an action confirmation image based on the inspection image during the operation of the image processing unit, the determination unit and the display unit. The defective image includes the defective feature part of the quality state of the item that is determined to be defective. and The action confirmation processing unit (54) performs action confirmation processing on the action confirmation image using the prescribed image processing algorithm in parallel with the operation of the image processing unit during the running operation to determine whether the operation of the image processing unit and the determination unit is normal.

2. The article inspection device according to claim 1, characterized in that, The action confirmation image generation unit synthesizes multiple parts with different image features in the inspection image during the running action into multiple partial images of the defect feature part to generate the action confirmation image.

3. The article inspection device according to claim 2, characterized in that, The image generation unit for action confirmation synthesizes multiple defect feature images of the item into defective product images by combining the qualified product inspection images obtained by the determination unit during the operation to determine the quality status of the item as qualified products.

4. The article inspection device according to claim 2, characterized in that, It also has: The control unit (30) has a first control function unit that controls the operation of the image processing unit, the determination unit, and the display unit to determine the quality status of each item. The control unit also has a second control function unit (50) that controls the actions of the action confirmation processing unit and the action confirmation image generation unit during the period when the first control function is performed, and determines whether the actions of the image processing unit and the determination unit are normal.

5. The article inspection device according to claim 2, characterized in that, The inspection image is an X-ray transmission image (Dpx) obtained by taking X-rays of the item.

6. The article inspection device according to claim 1, characterized in that, The image generation unit for action confirmation synthesizes multiple defect feature images of the item into defective product images by combining the qualified product inspection image (which is determined to be a qualified product by the determination unit during the operation) to form a defective product image.

7. The article inspection device according to claim 1, characterized in that, It also has: The control unit (30) has a first control function unit that controls the operation of the image processing unit, the determination unit, and the display unit to determine the quality status of each item. The control unit also has a second control function unit (50) that controls the actions of the action confirmation processing unit and the action confirmation image generation unit during the period when the first control function is performed, and determines whether the actions of the image processing unit and the determination unit are normal.

8. The article inspection device according to claim 1, characterized in that, The inspection image is an X-ray transmission image (Dpx) obtained by taking X-rays of the item.

Images