JP2021148486A - X-ray inspection device and x-ray inspection method - Google Patents

Abstract

To provide an X-ray inspection device and X-ray inspection method that can cut down costs, and can achieve a highly accurate X-ray inspection using X-ray images of a plurality of energy bands.SOLUTION: An X-ray inspection device, which comprises: an X-ray generator 21; an X-ray detector 24; and a determination unit 33 that determines a quality state of a workpiece W to be inspected, has: an X-ray image storage unit 31 that stores X-ray image data from the X-ray detector 24; and an image creation unit 42 that has a quasi image generation model 41 quasi-generating X-ray image data on other energy band on the basis of a learning result of the X-ray image data on a plurality of different energy bands including a prescribed energy band about a learning object kind, and creates a quasi transmission image in other energy band by the quasi image generation model 41 on the basis of X-ray image data on the workpiece W to be inspected. The determination unit 33 is configured to implement a determination on the basis of X-ray image data on the prescribed energy band of the workpiece W to be inspected, and the quasi transmission image in other energy band created in the image creation unit 42.SELECTED DRAWING: Figure 1

Description

The present invention relates to an X-ray inspection apparatus and an X-ray inspection method, and more particularly to an X-ray inspection apparatus and an X-ray inspection method that employ a dual or multi-energy X-ray detection method.

An X-ray inspection device that inspects the quality condition required of the object to be inspected by X-rays, especially the cumulative amount of X-rays that irradiate the object to be inspected during a predetermined period is detected by an X-ray detector and transmitted. An X-ray inspection device that creates an image has been known for some time.

Further, as an inspection method in such an X-ray inspection apparatus, there is one that improves the foreign matter detection accuracy by acquiring two transmitted image data having different X-ray energies and performing subtraction (difference processing) of both transmitted image data. Are known.

Conventional X-ray inspection devices and X-ray inspection methods include, for example, an X-ray curing filter attached to one of two X-ray detectors that detect X-rays transmitted through an object to be inspected, and X-ray energy. There is one that acquires two transparent image data with different values and performs subtraction (difference processing) (see Patent Document 1).

In addition, the tube currents of the two X-ray tubes are made different so that the intensity of the X-rays emitted from both X-ray tubes to the object to be inspected is made different, or the tube voltages of the two X-ray tubes are changed to obtain the wavelength of the X-rays. By detecting the transparent image data of the object to be inspected with two detectors, synthesizing both transparent image data into one image, and then reducing the composite image, while making the corresponding transmission power different. Some are designed to reduce the misalignment of the two transparent images remaining in the composite image (see Patent Document 2).

Japanese Unexamined Patent Publication No. 10-318943 Japanese Unexamined Patent Publication No. 2012-122927

In the conventional X-ray inspection apparatus and X-ray inspection method as described above, in order to realize highly accurate X-ray inspection, X-ray inspection is performed using X-ray images of a plurality of energy bands and radiation qualities. You need to do it.

However, in order to realize such highly accurate X-ray inspection, conventionally, a plurality of X-ray sensors that generate a plurality of X-ray images having different energy bands (radio quality) and X-rays of a plurality of energy bands are used. Since an X-ray source or an X-ray quality conversion means to be generated is required, there is a problem that not only the inspection cost is high but also the X-ray inspection apparatus becomes large.

The present invention has been made to solve the above-mentioned conventional problems, can reduce costs, and can realize highly accurate X-ray inspection using X-ray images of a plurality of energy bands. It is an object of the present invention to provide an inspection device and an X-ray inspection method.

(1) In order to achieve the above object, the X-ray inspection apparatus according to the present invention has transmitted an X-ray generator that generates X-rays in a predetermined energy band that passes through the transported object to be inspected and the X-ray inspection device that has passed through the object to be inspected. An X-ray detector that detects X-rays, converts them into electrical signals, and outputs an X-ray detection signal for each transmission region, and a determination unit that determines the quality status of the object to be inspected based on the X-ray detection signal. The X-ray inspection apparatus including the X-ray image storage unit for storing X-ray image data corresponding to the X-ray detection signal for each transmission region of the object to be inspected from the X-ray detector, and a learning target product type. Based on the learning results of X-ray image data of a plurality of different energy bands including the predetermined energy band, X-ray image data of other energy bands corresponding to the X-ray image data of the predetermined energy band are pseudo-generated. An image creation unit having a pseudo image generation model and creating a pseudo transmission image in the other energy band of the inspected object by the pseudo image generation model based on the X-ray image data of the inspected object. The determination unit includes the X-ray image data of the object to be inspected stored in the X-ray image storage unit, and a pseudo-transmission image in the other energy band created by the image creation unit. Based on the above, the determination is executed.

With this configuration, in the present invention, a plurality of expensive X-ray sensors and X-ray sources are provided to generate a plurality of X-ray images having different energy bands, and an expensive X-ray quality is used to share the X-ray sources. There is no need to provide conversion means. Therefore, it is an X-ray inspection device that can be miniaturized at low cost. Moreover, since X-ray image data of other energy bands corresponding to the X-ray image data of a predetermined energy band are pseudo-generated based on the learning results of the X-ray image data of a plurality of energy bands, X-rays in each transmission region are generated. It is possible to accurately generate a transmitted image of X-rays having different energy bands from an X-ray image based on a detection signal, and it is possible to realize a highly accurate X-ray inspection by using the X-ray images of a plurality of energy bands.

(2) In the present invention, the image creating unit uses the pseudo-transmissive image of the object to be inspected in the other energy band as an image having an X-ray transmittance different from that of the X-ray image data in the predetermined energy band. It can be configured to be generated as.

When implemented in such a form, it is possible to accurately grasp the mode of the change when the X-ray transmittance changes according to the X-ray absorption characteristics of each type of the object to be inspected by learning in advance. Images with different line transmittances can be accurately generated in a pseudo manner.

(3) The image creating unit performs the predetermined energy band based on the X-ray image data stored in the X-ray image storage unit each time the transported object passes through a predetermined inspection section. In addition to creating the X-ray image in the above, the pseudo-transmission image in the other energy band may be created.

In this case, an X-ray image in a predetermined energy band of the inspected object is based on the X-ray image data stored and stored in the X-ray image storage unit while the inspected object being conveyed passes through a predetermined inspection section. And pseudo-transmission images in other energy bands will be created, and while appropriately and effectively utilizing the results of deep learning, etc. for the object to be inspected of a predetermined type, for each object to be inspected. It is possible to realize highly accurate X-ray inspection using X-ray images of a plurality of energy bands.

(4) In the present invention, the determination unit is used for the determination by the determination unit with respect to the X-ray image of the object to be inspected in the predetermined energy band and the pseudo-transmission image in the other energy band. It may have an image processing means for executing the predetermined image processing of the above.

When carried out in such a form, it is possible to obtain an image in which foreign matter is emphasized by performing difference processing or the like between an image of a predetermined energy band and an image of another energy band as a predetermined image processing.

(5) In the present invention, the pseudo image generation model has the X-ray attenuation characteristic of the object to be inspected with respect to the X-rays of the plurality of energy bands from the X-ray image data of the plurality of energy bands relating to the cultivar to be learned. It can be composed of learned differences.

When implemented in such a form, the pseudo-image generation model is trained in advance to learn the difference in the X-ray attenuation characteristics due to the difference in the absorption characteristics of the X-rays to be inspected with respect to the X-rays of a plurality of energy bands having different heights. An effective pseudo-transparent image can be quickly created based on the X-ray image data stored in the X-ray image storage unit.

(6) The X-ray inspection method according to the present invention detects an X-ray generation stage that generates X-rays in a predetermined energy band that passes through the transported object to be inspected and an X-ray that has passed through the object to be inspected and performs electricity. An X-ray inspection method including an X-ray detection step of converting into a signal and outputting an X-ray detection signal for each transmission region, and a determination step of determining the quality state of the object to be inspected based on the X-ray detection signal. The other energy band corresponding to the X-ray image data of the predetermined energy band is based on the learning result of the X-ray image data of a plurality of different energy bands including the predetermined energy band for the cultivar to be learned. A model creation stage for creating a pseudo-image generation model capable of pseudo-generating X-ray image data, and an X-ray image storage stage for storing X-ray image data corresponding to an X-ray detection signal for each transmission region of the object to be inspected. Based on the X-ray image data of the object to be inspected stored in the X-ray image storage step, an image for creating a pseudo-transmission image of the object to be inspected in the other energy band by the pseudo image generation model. Including the creation step, in the determination step, the X-ray image data of the object to be inspected stored in the X-ray image storage step and pseudo-transmission in the other energy band created in the image creation step. It is characterized in that the determination is executed based on the image.

With this configuration, in the method of the present invention, an expensive X-ray sensor or X-ray source is used to generate a plurality of X-ray images having different energy bands, or an expensive X-ray is used to share the X-ray source. There is no need to use quality conversion means, and X-ray inspection can be performed at low cost. Moreover, since X-ray image data of other energy bands corresponding to the X-ray image data of a predetermined energy band are simulated based on the learning results of the X-ray image data of a plurality of energy bands, X-rays in each transmission region are generated. It is possible to accurately generate a transmitted image of X-rays having different energy bands from an X-ray image based on a detection signal, and it is possible to realize a highly accurate X-ray inspection by using the X-ray images of a plurality of energy bands.

(7) In the method of the present invention, in the image creation stage, the predetermined energy is based on the X-ray image data stored in the X-ray image storage unit while the object to be inspected passes through a predetermined inspection section. An X-ray image in the band is created, and a pseudo-transmission image in the other energy band having an X-ray transmission rate different from that of the X-ray image is created. The X-ray image in the predetermined energy band of the inspection object and the pseudo-transmission image in the other energy band may be configured to perform the predetermined image processing for the determination.

When carried out in such a form, it is possible to obtain an image in which foreign matter is emphasized by differential processing or synthesizing images of a predetermined energy band and another energy band, and to cover X-rays of a plurality of energy bands having different heights. The difference in the X-ray attenuation characteristics of the inspection object is learned in advance in the pseudo image generation model, and an effective pseudo transmission image can be quickly created based on the X-ray image data stored in the X-ray image storage unit. ..

According to the present invention, it is possible to provide an X-ray inspection apparatus and an X-ray inspection method capable of reducing costs and realizing highly accurate X-ray inspection using X-ray images of a plurality of energy bands.

It is a schematic block diagram of the X-ray inspection apparatus which concerns on one Embodiment of this invention. It is a block diagram which shows the schematic structure of the main part of the X-ray inspection apparatus which concerns on one Embodiment of this invention. It is a flowchart which shows the outline processing procedure of the X-ray inspection method carried out by the X-ray inspection apparatus which concerns on one Embodiment of this invention.

Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings.

As shown in FIGS. 1 and 2, the X-ray inspection device according to the embodiment of the present invention is configured as an X-ray foreign matter detection device that detects foreign matter mixed in the transported object (article) to be inspected. There is. Further, in this apparatus, as an X-ray inspection method according to an embodiment of the present invention, an X-ray inspection method capable of accurately detecting a foreign substance by acquiring two X-ray image data having different X-ray energies is adopted. doing.

First, the configuration of the X-ray inspection apparatus of this embodiment will be described.

As shown in FIGS. 1 and 2, the X-ray inspection apparatus 1 includes a transport unit 10, an X-ray inspection unit 20, and a control unit 30 thereof, and the inspected object W transported by the transport unit 10 on a conveyor. The X-ray inspection unit 20 irradiates the subject with X-rays, and inspects the quality state of the inspected object W based on the X-ray image data. 2) is detected. The quality condition referred to here is the suitability of the quality and physical quantity required for the product W to be inspected, for example, the presence or absence of mixed foreign matter, the presence or absence of missing items, and the pass / fail of the shape, size, storage state, etc. of the contents. , Density / thickness / volume or mass distribution, etc.

The transport unit 10 may wind the loop-shaped transport belt 11 around a plurality of transport rollers 12 and 13 and sequentially transport the object W to be inspected in the right direction in FIG. 1 by the traveling section 11a of the transport belt 11. It is a capable conveyor and is supported by a housing (not shown).

The X-ray inspection unit 20 has an X-ray generator 21 that generates X-rays in a predetermined energy band that passes through the inspected object W conveyed by the transfer unit 10, and the X-ray generator 21 is known. The X-ray tube 22 generates X-rays having a wavelength and intensity corresponding to the tube current and tube voltage, and also passes through the X-ray window portion 23a of the enclosure 23 and is a fan beam orthogonal to the transport direction of the transport unit 10. The shape of the X-ray can be applied to the object W to be inspected on the transport belt 11.

The X-ray inspection unit 20 further has an X-ray detector 24 arranged directly below the traveling section 11a of the transport belt 11.

Although the details of the X-ray detector 24 are not shown, for example, a detection element composed of a scintillator which is a phosphor and a photodiode or a charge coupling element is arranged in an array in a predetermined pitch in the width direction of the transport path of the transport unit 10. The X-ray line sensor camera is arranged so as to detect X-rays at a predetermined resolution, and is arranged at a predetermined position in the transport direction corresponding to the X-ray irradiation position from the X-ray generator 21. ..

That is, the X-ray detector 24 detects X-rays irradiated from the X-ray generator 21 and transmitted through the object W to be inspected for each predetermined transmission region corresponding to the detection element, and corresponds to the amount of the X-rays transmitted. It is possible to output an X-ray detection signal for each transmission region by converting it into an electric signal.

The control unit 30 controls the transport control means for controlling the transport speed and the transport interval of the object W to be inspected by the transport belt 11 in the transport unit 10, and controls the X-ray irradiation intensity and the irradiation period in the X-ray inspection unit 20. It includes an inspection control means for controlling the X-ray detection cycle by the X-ray line sensor of the X-ray detector 24 according to the transport speed of the object W to be inspected, the detection period of each object W to be inspected, and the like. Is omitted.

The control unit 30 also has an X-ray image storage unit 31 that captures an X-ray detection signal from the X-ray detector 24 at predetermined intervals and stores an X-ray image during the inspection period of each object W to be inspected, and an X-ray image storage unit 31. An image processing unit 32 that executes image processing for a predetermined determination process based on the image data captured in the line image storage unit 31, and a predetermined determination process based on the image data processed by the image processing unit 32. For example, it has a determination unit 33 that executes a process of determining the presence or absence of a foreign substance, and a display unit 34 that can display and output the determination result of the determination unit 33.

The X-ray image storage unit 31 is an image input unit, for example, A / D-converts X-ray detection signals from a plurality of detection elements of the X-ray detector 24, and corresponds to the size of the detection element in the X-ray detector 24. For each predetermined unit transport time, data on the cumulative transmission amount within the unit time is obtained, for example, 0 for all the detection element regions of the number n (n is an integer larger than 1, for example, 640) of the detection elements. An operation (hereinafter referred to as line scanning) of writing to a memory as digital data having a density level representing gradations from to 1023 is executed.

Further, the X-ray image storage unit 31 sequentially generates transmission density data written in the memory as X-ray image data when the line scanning is repeated a predetermined number of times according to the inspection period of each object W to be inspected. It has a data processing program and a working memory (not shown) that exert a function of outputting to the image processing unit 32.

The image processing unit 32 and the determination unit 33 cooperate with the ROM, for example, a microcomputer having a CPU, a ROM, a RAM, and an I / O interface (not shown) and a control program for exerting each function of a plurality of processing units described later. It is configured to include an auxiliary storage device that is readable and stored, a timer circuit, and the like, and a predetermined arithmetic process is performed while the CPU sends and receives data to and from the RAM and the like according to a control program stored in the ROM and the like. Is executed and the control program is executed at the same time.

As shown in FIG. 2, the image processing unit 32 has a filter processing means 32a that performs a predetermined filter process on the X-ray image for each object W to be inspected taken from the X-ray image storage unit 31. The filter processing means 32a is a feature extraction filter that enhances the contour (edge) of the foreign matter on the X-ray image for each object W to be inspected, for example, a differential filter such as a Sobel filter, and is predetermined in a region near the pixel of interest. The edge of the foreign matter is emphasized by performing differential processing based on the calculation formula.

The image processing unit 32 further has a foreign matter detection program that performs differential processing on the dual energy X-ray image of the object W to be measured to detect the presence or absence of mixed foreign matter. As described in, for example, Japanese Patent Application Laid-Open No. 2010-91483, this foreign matter detection program produces an equivalent thickness image pair of a low energy equivalent thickness image L (see FIG. 2) and a high energy equivalent thickness image H as dual energy X-ray images. The first process to be generated, the second process to obtain the separation matrix by applying the independent component analysis so as to separate the component of the foreign matter mixed in and the component of the object to be measured for the equivalent thickness image pair, and the separation matrix A third process for obtaining the weight parameter of the difference process based on the two separation vectors that are elements, and a fourth process for separating the component image of the mixed foreign matter from the equivalent thickness image pair by the difference process using the weight parameter. It is possible to execute the fifth process of detecting the foreign matter by performing the threshold processing on the component image of the mixed foreign matter.

The determination unit 33 obtains a binary image of the object contained inside the object W to be inspected by, for example, binarizing the difference-processed X-ray image, and from this binary image, the area and contour length of the object. And, by calculating an extraordinary amount such as the sum of concentrations and comparing the extraordinary amount with a predetermined judgment reference value, it is possible to judge whether or not a foreign substance satisfying the judgment condition is contained in the object W to be inspected. There is.

As shown in FIG. 2, the control unit 30 further uses image data of a large number of X-ray images of a plurality of energy bands having different heights acquired in advance for the cultivar to be learned, and a large number of images among the large number of images. A generation model 41 that is composed of a learner that learns the difference in feature quantities between a plurality of energy bands common to features and can output modeling information associated with the energy bands, and an X-ray image storage unit 31. For image data of a predetermined image size, based on modeling information associated with the energy band supplied from the generation model 41, X-ray image data of another energy band is obtained based on the X-ray image data of the predetermined energy band. It has a pseudo image generation unit 42 (image creation unit) for pseudo-creating.

The generation model 41 includes an X-ray image when the X-ray inspection unit 20 irradiates the X-ray inspection unit 20 with X-rays in a predetermined energy band in advance to the object W of the predetermined learning target type, and the X-ray inspection unit 20. By changing the X-ray irradiation conditions of the above and using an X-ray filter, we learned the X-ray image when X-rays of other energy bands that are different in height from the predetermined energy band are irradiated. Based on the result, modeling information capable of pseudo-generating an X-ray image of another energy band based on the X-ray image of the object W to be inspected in a predetermined energy band stored in the X-ray image storage unit 31 is provided in advance. Has been generated.

Specifically, the generation model 41 describes the X-ray images of each learning target for a large number of pre-labeled X-ray images of the test object W of the same type for each of the energy bands of the plurality of energy bands. A small area that can be filtered (feature extraction) is set inside, and the convolution process that compresses the image data of that small area into one feature amount (for example, a multi-gradation image density value) is performed in the X-ray image. With feature extraction, a layer image of a single convolution layer in which the information in the filter is convoluted is created repeatedly while sliding the small area in sequence, and a layer image of a one-layer pooling layer is created by further compressing this. , Feature extraction of combinations from random weighted combinations of nodes (values of features mentioned above) between adjacent layer images is executed in multiple layers by changing filters, movement ranges, etc., unless it is region-based. It constitutes a multi-layer (n-layer) neural network that can automatically extract changes in features that are difficult to extract regardless of the displacement of the object to be inspected in the target image.

Further, the generation model 41 has a plurality of relationships between the feature quantities of each layer of the neural network and the feature quantities of the nodes of the adjacent layers based on a large number of learning target images created as samples for each predetermined type of the object W to be inspected. The difference in the feature amount between a plurality of energy bands in which the energy band of the X-ray irradiated to the object W is different, for example, the transmission amount ratio and the transmission rate ratio of the X-ray is detected. Or, when a predetermined energy band is changed to another energy band by specifying an arithmetic expression such as a weighting coefficient of a feature amount or an activation function in a combination of connections of specific nodes between multiple layers of layer images. It can be output as the above-mentioned modeling information for creating a pseudo image (pseudo transparent image) of the obtained X-ray image.

The pseudo image generation unit 42 has a neural network for image generation corresponding to the multi-layer neural network in the generation model 41, for example, a hostile generation network using parameters such as the transmission ratio of each trained layer. When image data of a predetermined energy band is input from the line image storage unit 31, X-rays of another energy band are obtained based on the input image of the predetermined energy band and the above-mentioned modeling information from the generative model 41. The work of pseudo-generating image data and checking the pseudo degree of each layer using an arithmetic expression such as an activation function that has been learned for X-ray images of other energy bands is repeated so as to improve the degree of pseudo. It is supposed to run.

The pseudo image generation unit 42 creates image data of another energy band as an image having a different X-ray transmission rate from the X-ray image data of a predetermined energy band based on the feature data of the multilayer neural network model. As a pseudo image (an image of each layer of multiple layers or one composite image) presumed to have been obtained when the X-ray image from the X-ray image storage unit 31 was an X image of another energy band. It is designed to output.

The control unit 30 is based on the X-ray image data of a plurality of types of the object W to be inspected (for example, data in n rows and k columns) when the energy band is changed to a plurality of energy bands having different heights. The difference in the X-ray transmission characteristics of these multiple energy bands, that is, the difference in the absorption characteristics of the X-ray object W due to the difference in the X-ray energy bands and the corresponding X-ray attenuation characteristics, is described in a multi-layered two-dimensional system. Another X based on the image data of a predetermined X-ray energy band from the X-ray image storage unit 31 by the pseudo image generation unit 42 that has been trained by the generation model 41 of the convolutional neural network configuration and uses the learning result. It is configured to be able to create a pseudo-transmission image of the line energy band.

The pseudo-transmissive image of the other X-ray energy band from the pseudo-image generation unit 42 is taken into the image processing unit 32 together with the X-ray image of the predetermined energy band from the X-ray image storage unit 31, and is therefore described above. It is used for the difference processing of the dual energy X-ray image, the component image of the mixed foreign matter is separated from the equivalent thickness image pair, and the component image of the mixed foreign matter is threshold-processed to execute the foreign matter detection processing.

Then, the determination unit 33 binarizes the difference-processed image between the X-ray image data from the X-ray image storage unit 31 and the pseudo-transparent image from the pseudo-image generation unit 42 into the inside of the object W to be inspected. A binary image of an object such as a foreign substance contained therein is acquired, and from the binary image, for example, three types of extraordinary traces of the area, contour length, and sum of densities of the object are calculated, and the extraordinary trace value is a predetermined judgment reference value. By comparing with, it is possible to determine whether or not the foreign matter is contained in the object W to be inspected.

As described above, in the present embodiment, every time the object W to be inspected passes through a predetermined inspection section, an X-ray image in a predetermined energy band is generated based on the X-ray image data stored in the X-ray image storage unit 31. Along with the creation, a pseudo-transmission image in another energy band is also created, and the X-ray image in the predetermined energy band of the object W to be inspected and the pseudo-transmission image in the other energy band are differentially processed to determine. Predetermined image processing for determination in unit 33 is executed.

Next, the X-ray inspection method according to the embodiment of the present invention executed by using the X-ray inspection apparatus 1 of the present embodiment will be described, and its operation will be described.

FIG. 3 shows a schematic processing procedure of image processing in the X-ray inspection method according to the embodiment of the present invention, which is executed by the control unit 30 of the X-ray inspection apparatus 1 of the present embodiment.

First, the object W to be inspected (for example, packed sausage in FIG. 2) is transported by the transport unit 10 at a predetermined speed and a transport interval, and at the same time, with respect to the object W to be inspected passing through a predetermined inspection section. , X-rays are emitted from the X-ray inspection unit 20 with a preset irradiation intensity.

At this time, when the entry of the inspected object W into the inspection section is detected by an article detection sensor (not shown), the X-ray generator 21 of the X-ray inspection unit 20 irradiates the inspected object W with X-rays and transmits them. The X-rays are detected for each predetermined transmission region by the line scanning of the X-ray detector 24, and an X-ray detection signal corresponding to the amount of the X-ray transmission is transmitted from the X-ray detector 24 to the X-ray image storage unit 31. It is sequentially taken in (step S11).

Then, while the line scanning of the X-ray detector 24 is repeated a predetermined number of scans, the original image (hereinafter, the original image) in which the transmission amount sequentially written in the X-ray image storage unit 31 is used as the density value of each pixel. The X-ray image data of) is generated and output to the image processing unit 32 and also output to the pseudo image generation unit 42.

Next, the pseudo image generation unit 42 simulates the X-ray image data of another energy band based on the input image of the predetermined energy band from the X-ray image storage unit 31 and the above-mentioned modeling information from the generation model 41. It is generated (step S12).

Next, the image processing unit 32 uses the first energy image L, which is an X-ray image of a predetermined energy band from the X-ray image storage unit 31, and a pseudo transmission image of another X-ray energy band from the pseudo image generation unit 42. The difference processing of the dual energy X-ray image described above is executed using the second energy image H, and when there is a foreign substance mixed in, the component image is separated (step S13).

The X-ray image that has been differentially processed by the image processing unit 32 is binarized, a predetermined extraordinary amount is calculated, and the extraordinary amount is compared with a predetermined judgment reference value to execute a pass / fail judgment based on the presence or absence of foreign matter. (Step S14).

As described above, in the present embodiment, in order to acquire the original image of the predetermined energy band, the X-ray generation step of generating the X-ray of the predetermined energy band that passes through the conveyed object W and the object W to be inspected. An X-ray detection step that detects X-rays that have passed through and outputs an X-ray detection signal for each transmission region, and a judgment step that determines the quality status of the object W to be inspected based on the X-ray detection signal. And, an X-ray inspection method including.

Further, in the present embodiment, an X-ray image storage step (step S11 described above) for storing X-ray image data corresponding to the X-ray detection signal for each transmission region of the object W to be inspected, and a plurality of different types of learning target varieties. From the X-ray image data of the energy band, the difference in the X-ray attenuation characteristics of the object W to be inspected with respect to the X-rays of the plurality of energy bands is learned, and the X-ray image data of the other energy bands corresponding to the X-ray image data of the predetermined energy band. Based on the X-ray image data of the object W to be inspected stored in the model creation stage for creating the pseudo-image generation model 41 capable of pseudo-generating X-ray image data and the X-ray image storage stage, the inspected object is inspected by the pseudo-image generation model. In the determination step (steps S13 and S14 described above), which includes an image creation step (step S12 described above) for creating a pseudo-transmissive image in another energy band of the object W, the subject stored in the X-ray image storage step. The presence or absence of foreign matter is determined based on the X-ray image data of the inspection object W and the pseudo-transmission image in another energy band created in the image creation stage.

In the present embodiment configured in this way, an expensive X-ray sensor or X-ray source is used to generate a plurality of X-ray images having different energy bands, or an expensive X-ray source is used in common. It is not necessary to use a simple X-ray quality conversion means, and X-ray inspection can be performed at low cost. Moreover, since X-ray image data of other energy bands corresponding to the X-ray image data of a predetermined energy band can be simulated based on the learning results of the X-ray image data of a plurality of energy bands, X-rays in each transmission region can be generated. X-ray inspection that can accurately generate transmitted images of X-rays with different energy bands from X-ray images based on the detection signal, and can detect foreign substances with high accuracy by subtraction method etc. using these X-ray images of multiple energy bands. Will be realized.

In particular, in the present embodiment, X-rays in a predetermined energy band are obtained based on the X-ray image data stored in the X-ray image storage stage while the object W to be inspected passes through the predetermined inspection section in the image creation stage. In addition to creating an image, a pseudo-transmission image in another energy band having an X-ray transmission rate different from that of the X-ray image is created. Predetermined image processing for determination can be performed on the image and the pseudo-transparent image in other energy bands.

Therefore, it is possible to obtain an image in which foreign matter is emphasized by performing difference processing or synthesizing images of a predetermined energy band and another energy band, and X-rays of an inspected object W with respect to X-rays of a plurality of energy bands having different heights. The difference in attenuation characteristics is trained in advance in the pseudo image generation model, and an effective pseudo transmission image can be quickly created based on the X-ray image data stored in the X-ray image storage stage.

As described above, the present embodiment provides an X-ray inspection apparatus and an X-ray inspection method capable of reducing costs and realizing highly accurate X-ray inspection using X-ray images of a plurality of energy bands. be able to.

In one embodiment described above, the X-ray inspection device 1 is used as an X-ray foreign matter detection device, and the image processing by the image processing unit 32 for determination by the determination unit 33 is performed by the energy subtraction method for two energy bands. Although the difference processing of the X-ray image is used, the present invention can be applied to an X-ray detection device other than the X-ray foreign matter detection device, and is also applied to a multi-energy X-ray detection method device instead of the dual energy detection method. It is possible. Needless to say, the image processing in the image processing unit 32 can be replaced with other conventionally known image processing depending on the content of the inspection.

Further, the image creation unit for creating the pseudo-transparent image referred to in the present invention uses a learner having a multi-layer neural network configuration using AI (artificial intelligence) or an image generation model, but the object to be inspected. Of course, the learning content may be added or changed, or the type of image used for image processing may be increased or decreased depending on the type, transport form, inspection item, and the like.

As described above, the present invention can provide an X-ray inspection apparatus and an X-ray inspection method capable of reducing costs and realizing highly accurate X-ray inspection using X-ray images of a plurality of energy bands. Is. The present invention is useful for all X-ray inspection devices and X-ray inspection methods that employ dual or multi-energy X-ray detection methods.

1 X-ray inspection device 10 Conveying unit 11 Conveying belt 11a Running section 12, 13 Conveying roller 20 X-ray inspection unit 21 X-ray generator 22 X-ray tube 23 Enclosure 23a X-ray window 24 X-ray detector 30 Control Unit 31 X-ray image storage unit 32 Image processing unit 32a Filter processing means 33 Judgment unit 34 Display unit 41 Generation model (pseudo image generation model)
42 Pseudo image generation unit (image creation unit)

Claims (7)

An X-ray generator that generates X-rays in a predetermined energy band that passes through the transported object to be inspected, and an X-ray detection signal that detects the X-rays that have passed through the object to be inspected and converts them into electrical signals for each transmission region. In an X-ray inspection apparatus including an X-ray detector that outputs an X-ray detector and a determination unit that determines the quality state of the object to be inspected based on the X-ray detection signal.
An X-ray image storage unit that stores X-ray image data corresponding to an X-ray detection signal for each transmission region of the object to be inspected from the X-ray detector.
Based on the learning results of X-ray image data of a plurality of different energy bands including the predetermined energy band for the cultivar to be learned, X-ray image data of other energy bands corresponding to the X-ray image data of the predetermined energy band. Has a pseudo-image generation model that pseudo-generates the Equipped with an image creation unit
The determination unit is based on the X-ray image data of the object to be inspected stored in the X-ray image storage unit and the pseudo-transmission image in the other energy band created by the image creation unit. An X-ray inspection apparatus characterized by performing a determination. The image creating unit is characterized in that a pseudo-transmission image of the object to be inspected in the other energy band is generated as an image having a different X-ray transmittance with respect to the X-ray image data of the predetermined energy band. The X-ray inspection apparatus according to claim 1. Each time the transported object passes through a predetermined inspection section, the image creating unit performs X in the predetermined energy band based on the X-ray image data stored in the X-ray image storage unit. The X-ray inspection apparatus according to claim 1 or 2, wherein a line image is created and the pseudo-transmission image in the other energy band is created. The determination unit executes predetermined image processing for the determination by the determination unit on the X-ray image in the predetermined energy band and the pseudo-transmission image in the other energy band of the object to be inspected. The X-ray inspection apparatus according to any one of claims 1 to 3, further comprising an image processing means. The pseudo image generation model is obtained by learning the difference in the X-ray attenuation characteristics of the object to be inspected with respect to the X-rays of the plurality of energy bands from the X-ray image data of the plurality of energy bands relating to the cultivar to be learned. The X-ray inspection apparatus according to any one of claims 1 to 4, wherein the X-ray inspection apparatus is characterized. An X-ray generation stage that generates X-rays in a predetermined energy band that passes through the transported object to be inspected.
An X-ray detection step that detects X-rays that have passed through the object to be inspected, converts them into electrical signals, and outputs an X-ray detection signal for each transmission region.
An X-ray inspection method including a determination step of determining the quality state of the object to be inspected based on the X-ray detection signal.
Based on the learning results of X-ray image data of a plurality of different energy bands including the predetermined energy band for the cultivar to be learned, X-ray image data of other energy bands corresponding to the X-ray image data of the predetermined energy band. At the model creation stage to create a pseudo-image generation model that can be pseudo-generated,
An X-ray image storage step for storing X-ray image data corresponding to the X-ray detection signal for each transmission region of the object to be inspected, and an X-ray image storage step.
An image creation step of creating a pseudo-transmission image of the object to be inspected in the other energy band by the pseudo image generation model based on the X-ray image data of the object to be inspected stored in the X-ray image storage step. And, including
In the determination step, the determination is made based on the X-ray image data of the object to be inspected stored in the X-ray image storage step and the pseudo-transmission image in the other energy band created in the image creation step. An X-ray inspection method characterized by performing. In the image creation stage, an X-ray image in the predetermined energy band is created based on the X-ray image data stored while the object to be inspected passes through the predetermined inspection section, and the X-ray image is created. Creates the pseudo-transmission image in the other energy band having a different X-ray transmission rate.
In the determination step, prior to the determination, a predetermined image processing for the determination is executed on the X-ray image in the predetermined energy band and the pseudo-transmission image in the other energy band of the object to be inspected. The X-ray inspection method according to claim 6, wherein the X-ray inspection method is performed.

Images