Image processing device, image processing method, and image processing program

Abstract

To enable accurate detection of lithium-ion batteries from X-ray images of waste materials. [Solution] The present invention provides an image processing device for detecting whether or not an object has a battery based on an X-ray image of the object, comprising: an image acquisition unit that acquires an X-ray image of the object; and an image analysis unit that, based on the image data of the X-ray image, detects an image region with little variation in brightness among an image region that includes a dark area whose brightness value is less than a first threshold set in advance, as a candidate area where a battery may be present.

Description

Technical Field

[0001] The present invention relates to an image processing apparatus, an image processing method, and an image processing program.

Background Art

[0002] In a waste treatment system, when waste products or waste containing lithium-ion batteries are crushed, it may cause a fire. Since many of them, including smoke generation, spark generation, and fire, occur in the disassembly process or the crushing process, it is required to find out waste products etc. with built-in lithium-ion batteries before feeding them into the crusher.

[0003] Patent Document 1 discloses a method for excluding small household appliances with built-in rechargeable batteries. The technique described in Patent Document 1 pre-sets a reference case image showing the outer shape of a small household appliance and a reference battery image, and when the identification case image obtained by irradiating the target small household appliance with X-rays matches the reference case image, and further when the identification battery image and the reference battery image match, it is determined that the small household appliance waste has a built-in rechargeable battery.

[0004] Generally, since a lithium-ion battery is covered with metal, it blocks the transmission of X-rays and appears as a black silhouette in the X-ray image, and this black silhouette is determined as the lithium-ion battery part.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] However, since other components such as motors are also covered in metal, X-ray images show these other components as black silhouettes, making it difficult to distinguish the lithium-ion battery from the other components.

[0007] Furthermore, the amount of X-rays that penetrate a lithium-ion battery varies depending on the X-ray irradiation intensity. When the irradiation intensity is low, the lithium-ion battery appears as a black silhouette, and when the irradiation intensity is high, it appears as a light gray silhouette, making it difficult to identify the lithium-ion battery based solely on its brightness.

[0008] While there are conventional technologies that identify a black silhouette as a lithium-ion battery when there is a large difference in brightness (contrast) between the black silhouette and its surrounding area, the contrast may not be large enough depending on the material and thickness of the metal casing, which can lead to the lithium-ion battery being overlooked.

[0009] Furthermore, the same issues described above apply not only to lithium-ion batteries but also to other secondary batteries and primary batteries.

[0010] Therefore, in view of the above-mentioned problems, the present invention aims to provide an image processing apparatus, an image processing method, and an image processing program that can accurately detect batteries from X-ray images of objects to be discarded, by utilizing the structure of a battery in which the components are evenly filled inside a metal housing. [Means for solving the problem]

[0011] To solve these problems, the first image processing apparatus of the present invention is an image processing apparatus that detects whether or not an object has a battery based on an X-ray image of the object, and is characterized by comprising: (1) an image acquisition unit that acquires an X-ray image of the object; and (2) an image analysis unit that, based on the image data of the X-ray image, detects an image region with little variation in brightness among an image region that includes a dark area whose brightness value is less than a first threshold set in advance, as a candidate area where a battery may be present.

[0012] A second image processing method of the present invention is an image processing method for detecting whether or not an object has a battery based on an X-ray image of the object, characterized in that (1) an image acquisition unit acquires an X-ray image of the object, and (2) an image analysis unit detects, based on the image data of the X-ray image, an image region that includes a dark area whose brightness value is less than a first threshold set in advance, and which has little variation in brightness, as a candidate area where a battery may be present.

[0013] A third image processing program of the present invention is an image processing program that detects whether or not an object has a battery based on an X-ray image of the object, characterized in that a computer functions as (1) an image acquisition unit that acquires an X-ray image of the object, and (2) an image analysis unit that, based on the image data of the X-ray image, detects image regions with little variation in brightness among image regions that include dark areas where the brightness value is less than a first threshold set in advance, as candidate areas where a battery may be present. [Effects of the Invention]

[0014] According to the present invention, batteries (especially lithium-ion batteries) can be accurately detected from X-ray images of waste materials. [Brief explanation of the drawing]

[0015] [Figure 1] This is a configuration diagram showing a part of the waste treatment system according to the embodiment. [Figure 2] This is a flowchart showing the operation of the battery detection process by the information processing device according to the embodiment. [Figure 3] This is a diagram illustrating the X-ray image analysis process by an information processing device according to the embodiment (Part 1). [Figure 4] This is a diagram illustrating the X-ray image analysis process using the information processing device of the embodiment (part 2). [Figure 5] This is a diagram illustrating the X-ray image analysis process using the information processing device of the embodiment (part 3). [Figure 6] It is an image diagram (Part 4) for explaining the X-ray image analysis process by the information processing apparatus of the embodiment. [Figure 7] It is an image diagram (Part 5) for explaining the X-ray image analysis process by the information processing apparatus of the embodiment. [Figure 8] It is an image diagram (Part 6) for explaining the X-ray image analysis process by the information processing apparatus of the embodiment.

Mode for Carrying Out the Invention

[0016] (A) Embodiment Hereinafter, embodiments of an image processing apparatus, an image processing method, and an image processing program according to the present invention will be described in detail with reference to the drawings.

[0017] (A-1) Configuration of the Embodiment FIG. 1 is a configuration diagram showing a partial configuration of a waste treatment system according to an embodiment.

[0018] In FIG. 1, a waste treatment system 10 of the embodiment includes an information processing apparatus 1, an X-ray imaging apparatus 2, and a conveyor 3.

[0019] Generally, in a waste treatment facility, the waste that has been carried in is disassembled or finely crushed by a crusher, and the waste is appropriately treated according to the type of waste.

[0020] The waste treatment system 10 in FIG. 1 is a waste treatment system before disassembling or crushing the waste 5, irradiates the waste 5 with X-rays, and uses the X-ray image obtained by the X-rays passing through the waste 5 to detect whether a lithium-ion battery is built in. It is a system.

[0021] The conveyor 3 is a conveying device that conveys the carried-in waste 5. For example, a roller conveyor, a belt conveyor, etc. can be applied, and the conveyance destination of the conveyor 3 is a machine such as a crusher used in the next waste treatment process, an incinerator, or a storage for temporarily storing the waste 5 to proceed to the next process.

[0022] Here, the waste material 5 can be general waste, bulky waste, non-burnable waste, etc., but in this embodiment, in order to detect the presence or absence of lithium-ion batteries, the waste material 5 to be targeted is electrical appliances and electronic devices that may contain lithium-ion batteries. Of course, various types of waste are mixed together, and it is also possible to detect devices that may contain lithium-ion batteries from X-ray images obtained by transmitting X-rays through them.

[0023] The X-ray imaging device 2 is an imaging device that uses transmitted X-rays to image waste 5, and comprises an X-ray source 21, a detector 22, and an X-ray image generation unit 23. The X-ray source 21 is installed above the conveyor 3 and irradiates the waste 5 being transported by the conveyor 3 with X-rays. The detector 22 is installed on the side opposite the X-ray source 21 and detects the X-rays that have passed through the waste 5 (transmitted X-rays), and the X-ray image generation unit 23 generates an X-ray image of the waste 5. The detector 22 can be an image sensor such as a CCD (Charge-Coupled Device) sensor or a CMOS (Complementary Metal Oxide Semiconductor) sensor. The X-ray image of the waste 5 generated by the detector 22 is provided to the information processing device 1. The X-ray imaging device 2 is made of a housing formed of a material with high X-ray shielding ability to prevent external leakage of irradiated X-rays.

[0024] Information processing device 1 is a computer connected to the X-ray imaging device 2 that performs various processing tasks. Information processing device 1 may be installed as a server on the cloud, or it may be a computer mounted on the X-ray imaging device 2.

[0025] The information processing device 1 detects whether or not waste 5 (the object) contains a battery based on an X-ray image of the waste object, and comprises an image acquisition unit that acquires an X-ray image of the object, and an image analysis unit that, based on the image data of the X-ray image, detects image regions with little variation in brightness among image regions that include dark areas whose brightness value is less than a first threshold set in advance, as candidate areas where the battery may be located.

[0026] The information processing device 1 includes a control unit 11, a storage unit 12, an input unit 13, a display unit 14, and the like.

[0027] The control unit 11 is a device having a CPU (Central Processing Unit), ROM (Read Only Memory), RAM (Random Access Memory), etc. The CPU loads various computer programs (for example, image processing programs, etc.) stored in the built-in ROM and memory unit 42 onto the RAM and executes them, thereby making the entire device function as the information processing device 1 of the present disclosure. The control unit 11 is not limited to the above configuration, and may be any processing circuit or arithmetic circuit equipped with a GPU (Graphics Processing Unit), FPGA (Field Programmable Gate Array), DSP (Digital Signal Processor), volatile or non-volatile memory, etc.

[0028] The control unit 11 includes an image acquisition unit 111, an image analysis unit 112, and an output unit 115.

[0029] The image acquisition unit 111 acquires the X-ray image to be analyzed.

[0030] The image analysis unit 112 analyzes the X-ray image acquired by the image acquisition unit 111 and includes a candidate area detection unit 113 and a battery portion detection unit 114. A detailed explanation of the image analysis unit 112, which includes the candidate area detection unit 113 and the battery portion detection unit 114, will be provided in the operation section.

[0031] The output unit 115 outputs the results analyzed by the image analysis unit 112 to a display unit such as a display.

[0032] The storage unit 12 can be a storage device such as an SSD (Solid State Drive) or a hard disk drive. The storage unit 12 stores the X-ray images generated by the X-ray imaging device 2. The storage unit 12 also stores various computer programs executed by the control unit 11 and data used when the computer programs are executed. The computer programs stored in the storage unit 12 include a battery detection program that determines whether or not lithium-ion batteries are included in the waste transported by the conveyor 3 described above.

[0033] The battery detection program is a single computer program, but it may also be a group of programs constructed from multiple computer programs, or it may be distributed across multiple computers and executed collaboratively by those computers.

[0034] The input unit 13 is an input device such as a touch panel, keyboard, or switch, and accepts various operations and data inputs from the operator. The control unit 11 performs appropriate control based on the various operation information provided by the input unit 13 and stores the input data in the storage unit 12 as needed.

[0035] The display unit 14 is equipped with a display device such as a liquid crystal display or an organic EL (Electro-Luminescence) display. The display unit 14 displays information that should be notified to workers, etc., in response to instructions from the control unit 11.

[0036] (A-2) Operation of the embodiment Next, the operation of the battery detection process by the information processing device 1 according to the embodiment will be described with reference to the drawings.

[0037] Figure 2 is a flowchart showing the operation of the battery detection process by the information processing device 1 according to the embodiment. Figures 3 to 9 are image diagrams illustrating the X-ray image analysis process in the information processing device 1. The explanation will follow the flow in Figure 2, using the images in Figures 3 to 9 as appropriate.

[0038] Here, when the X-ray imaging device 2 generates an X-ray image of the waste 5, the information processing device 1 analyzes the X-ray image in real time or near real time to determine whether or not a lithium-ion battery is built in. Alternatively, the X-ray image of the waste 5 may be temporarily stored in the storage unit 12, and the determination of whether or not a lithium-ion battery is built in may be made after a delay.

[0039] Furthermore, for the sake of simplicity, this explanation assumes that one object is captured in one X-ray image, and that one object contains one lithium-ion battery. However, even when multiple objects are captured in one X-ray image, or when one object contains multiple lithium-ion batteries, detection can be performed using the same process.

[0040] [Step S101] First, the waste material 5 is placed on the conveyor belt 3 and transported to a position where it can be photographed by the X-ray imaging device 2.

[0041] When the waste 5 is transported to a position where imaging is possible, the X-ray imaging device 2 uses an X-ray source 21 to irradiate the waste 5 with X-rays, a detector 22 detects the X-rays that have passed through the waste 5, and an X-ray image generation unit 23 generates an X-ray image. The X-ray image of the waste 5 is then provided to the information processing device 1.

[0042] In the information processing device 1, the image acquisition unit 111 acquires an X-ray image of the waste 5 (step S101).

[0043] Here, we assume that waste 5 is a robotic vacuum cleaner, and that an X-ray image of waste 5 (robotic vacuum cleaner) is generated as illustrated in Figure 3. Generally, X-ray images are monochrome, so we will use a monochrome image here as well. Note that even if a color X-ray image is used, the same effect can be obtained by performing the same processing as in this embodiment.

[0044] [Step S102] The candidate area detection unit 113 of the image analysis unit 112 analyzes the acquired X-ray image of the waste 5 to detect a first candidate area where a lithium-ion battery may be present (step S102).

[0045] First, the image analysis unit 112 performs preprocessing of the X-ray image. For example, the image analysis unit 112 may perform preprocessing such as image cropping, image size adjustment, and brightness adjustment.

[0046] The candidate area detection unit 113 identifies dark areas in the X-ray image 510 in Figure 3 where the brightness (e.g., lightness) value is less than a first threshold as white areas in Figure 4, and surrounds the region where these clusters of white areas exist with a white dotted frame to define it as the first candidate area.

[0047] In the X-ray image 520 of Figure 4, the candidate area detection unit 113 surrounds the region where the white mass exists with a white dotted frame, and designates each of the three regions R1, R2, and R3 as the first candidate area.

[0048] Here, various methods can be applied to determine the clusters of white areas in Figure 4, but one method that can be applied is to determine a cluster when the number of pixels in the dark areas (white areas in Figure 4) where the brightness value is less than the first threshold is greater than or equal to the pre-set first threshold.

[0049] Another method is to define a cluster when the proportion of pixels in the white area (i.e., the ratio of pixels) across the entire X-ray image is greater than or equal to a predetermined ratio. In any case, regions with a large number of white pixels are extracted and designated as the first candidate areas where lithium-ion batteries may be present.

[0050] [Step S103] The candidate area detection unit 113 generates a histogram for each of the first candidate areas showing the brightness distribution within that area (step S103).

[0051] For example, the candidate area detection unit 113 generates a histogram showing the distribution of brightness at the pixel level in the X-ray image.

[0052] Figures 5(A) to 5(C) show histograms of brightness for regions R1, R2, and R3. The horizontal axis represents the brightness value (e.g., the lightness value) with a range of 0 to 255, and the vertical axis represents the distribution value (e.g., the number of pixels of lightness).

[0053] [Step S104] The candidate area detection unit 113 determines whether the variation in brightness within the first candidate area is below a certain value (also called the second threshold). Then, it selects the areas within the first candidate area where the variation in brightness is below the certain value as the second candidate area (step S104).

[0054] This utilizes the structural characteristics of lithium-ion batteries. Generally, lithium-ion batteries are enclosed in a metal casing, and within the metal casing, components such as positive electrodes, negative electrodes, and separators are stacked or wound, with the components evenly packed without any gaps. In contrast, motors enclosed in a metal casing have metal shafts and brushes, but multiple components are arranged within the metal casing at intervals rather than evenly.

[0055] Therefore, when taking X-ray images, the amount of radiation transmitted is not constant depending on the difference in X-ray irradiation intensity, but lithium-ion batteries, whose components are evenly packed, produce a silhouette with less variation in brightness.

[0056] On the other hand, other components such as motors have a silhouette where dark areas appear due to the low amount of radiation transmitted through shafts, while bright areas appear due to the high amount of radiation transmitted through gaps, resulting in a silhouette with a relatively large mixture of dark and bright areas.

[0057] Therefore, in this embodiment, the variation in brightness within the first candidate area is analyzed, and if the variation in brightness is below a certain value, that is, if there are many dark areas in the first candidate area, it is determined to be a second candidate area where a lithium-ion battery may be present.

[0058] The candidate area detection unit 113 determines the degree of brightness variation within the first candidate area based on the brightness histogram for each of the first candidate areas.

[0059] For example, the candidate area detection unit 113 calculates the sum of the histogram values ​​within the first candidate area, that is, the total number of pixels present within the first candidate area.

[0060] Next, the candidate area detection unit 113 accumulates the distribution values ​​(histogram values) from the darkest to the lowest in the brightness histogram for each of the first candidate areas, and sets the point where the accumulated value reaches 10% brightness (brightness value) as [hist_low].

[0061] Furthermore, the candidate area detection unit 113 accumulates histogram values ​​starting from the brighter areas, and sets the point where the accumulated value reaches 20% (brightness value) as [hist_high].

[0062] Then, if the value of [hist_high]-[hist_low] is less than or equal to a certain value, it is determined that the first candidate area contains a portion where a lithium-ion battery may exist; otherwise, it is determined that a lithium-ion battery does not exist in the first candidate area.

[0063] In other words, when the difference in brightness between the brightness values ​​of [hist_high] and [hist_low] is less than or equal to a certain value, it is determined that the difference in brightness between the dark and bright areas is small, and therefore the variation in brightness is small.

[0064] On the other hand, when the difference in brightness is greater than a certain value, it can be determined that the difference in brightness between the dark and bright areas is large, and therefore the variation in brightness is considered to be relatively large.

[0065] The aforementioned constant value is a threshold for determining the variation in brightness, and can be a value set in advance, such as a constant value like 5 or 10.

[0066] Furthermore, in the example above, [hist_low] was defined as the cumulative histogram value from the darkest area being 10% brightness, and [hist_high] was defined as the cumulative histogram value from the brightest area being 20% ​​brightness. However, the values ​​of 10% or 20% are not limiting; other values ​​may be used as long as the brightness of the dark and bright areas in the X-ray image can be identified.

[0067] Figures 5(A), 5(B), and 5(C) show histograms of the brightness of regions R1, R2, and R3, respectively. In this example, the constant value is assumed to be "5".

[0068] In region R1 of Figure 5(A), [hist_low]=5 and [hist_high]=9. In this case, the brightness difference value "4" is less than or equal to the constant value "5" ([hist_high]-[hist_low]≦constant value), so the brightness variation is small and the candidate area detection unit 113 determines that there is a part of region R1 where a lithium-ion battery may exist.

[0069] In region R2 of Figure 5(B), [hist_low]=5 and [hist_high]=9. In this case, the brightness difference value "4" is less than or equal to the constant value "5" ([hist_high]-[hist_low]≦constant value), so the brightness variation is small and the candidate area detection unit 113 determines that there is a part of region R2 where a lithium-ion battery may exist.

[0070] In region R3 of Figure 5(C), [hist_low]=14 and [hist_high]=91. In this case, the brightness difference value "77" is greater than the constant value "5" ([hist_high]-[hist_low]>constant value), so the brightness variation is large and the candidate area detection unit 113 determines that there is no lithium-ion battery in region R3.

[0071] In this step, out of regions R1 to R3, regions R1 and R2 are selected as second candidate areas, but region R3 is not selected as a second candidate area.

[0072] [Step S105] The candidate area detection unit 113, for each second candidate area, uses a brightness value offset from the brightness value of the peak of the brightness histogram as a third threshold, and selects the region containing parts darker than the third threshold (white parts) as the third candidate area (step S105). The candidate area detection unit 113 surrounds the region containing the white parts with a white dotted line frame.

[0073] Here, the peak brightness value is the brightness value at which the distribution value is maximum in the brightness histogram, and is denoted as [hist_peak].

[0074] The offset is a brightness value that is a predetermined distance from the peak brightness value and is slightly brighter than the peak brightness value. For example, if a 10% margin is taken from the peak brightness value, the offset value is set to 110%. This sets the value obtained by multiplying the peak brightness value by 110% ([hist_peak] × 110%) as the third threshold, and the area that is darker than this (the white area in the figure) is considered the battery area.

[0075] Furthermore, as mentioned above, the offset value may be a ratio to the peak brightness value, and the value of this ratio may be variable. Also, for example, the offset value may be a pre-set value such as "5" or "10".

[0076] Referring to Figures 4, 5, and 6, the process by which the candidate area detection unit 113 selects a third candidate area when the offset value is set to 110% will be explained.

[0077] In Figure 5(A), the histogram of brightness in region R1 shows that [hist_peak] = 6. Therefore, a value of 6 × 110% is used as the third threshold, and areas darker than the third threshold are identified.

[0078] Here, by identifying the areas darker than the threshold, the white mass in region R1 in Figure 4 is separated into two, and each mass is enclosed in a white dotted frame, making the two regions R11 and R12 the third candidate areas (see Figure 6).

[0079] In Figure 5(B), the histogram of brightness in region R2 shows that [hist_peak] = 6. Therefore, a value of 6 × 110% is used as the third threshold. Areas darker than the third threshold are identified and enclosed with a white dotted line frame, making region R2 the third candidate area.

[0080] Following this step, as illustrated in Figure 6, three candidate areas can be identified in the X-ray image 530: region R11, region R12, and region R2.

[0081] [Step S106] Next, the battery detection unit 114 analyzes the shape of the dark area within the third candidate area to determine whether or not the battery is included within the third candidate area (step S106).

[0082] Here, since lithium-ion batteries are generally rectangular or cylindrical (including button-type batteries), the battery detection unit 114 uses this to select the dark areas (white areas) within the third candidate area that are quadrangular (including rhombuses, trapezoids, etc.) or circular (including ellipses). This eliminates candidate areas that are easily mistaken for lithium-ion batteries, such as the metal parts of a kitchen knife or an iron.

[0083] Depending on the orientation of the lithium-ion battery contained in waste 5, for example, if a cylindrical lithium-ion battery is photographed sideways, it will appear as a quadrilateral in the X-ray image, and if a cylindrical lithium-ion battery is photographed from the plane, it will appear as a circle in the X-ray image. Also, if a rectangular lithium-ion battery is photographed at an angle, depending on the angle, it may appear as a rhombus or a trapezoid.

[0084] An example of a method for determining the battery portion by the battery portion detection unit 114 will be explained with reference to the X-ray image 540 in Figure 7.

[0085] For example, the battery detection unit 114 rotates the dark area suspected to be the lithium-ion battery area by 180° for each third candidate area, using the center of the dark area as the point of point symmetry. If the rotated dark area matches or nearly matches the original dark area, the unit determines that the dark area is the lithium-ion battery area.

[0086] This method utilizes the fact that when a lithium-ion battery appears relatively clearly as a quadrangular or circular shape in an X-ray image, rotating the dark area predicted to represent the battery symmetrically is expected to result in a shape that is almost identical to the original dark area.

[0087] Furthermore, instead of limiting the scope to point symmetry, line symmetry can also be applied to the area around a line passing through the center of the region. If the shape of the dark area after the movement matches the shape of the original dark area, it may be determined to be a lithium-ion battery.

[0088] For example, as another determination method, the battery detection unit 114 determines that the dark area is a cylindrical lithium-ion battery when, for each third candidate area, the vertical and horizontal dimensions of the rectangle surrounding the dark area are approximately the same, and the area ratio of the dark area to the area of ​​the rectangle is close to approximately 0.78.

[0089] This method utilizes the fact that when a cylindrical lithium-ion battery is reflected as a quadrilateral, the ratio of its vertical length to its horizontal length remains constant. Of course, this is not limited to this example; the ratio of the vertical and horizontal lengths can be known in advance based on the standardized shape of the lithium-ion battery, so by pre-setting the standardized vertical-to-horizontal ratio, it is possible to accommodate other standards as well.

[0090] [Step S107] The battery detection unit 114 determines whether the outer shape of the third candidate area matches the outer shape of a standardized battery (step S107). If they match, the battery detection unit 114 determines that the third candidate area is a region where a lithium-ion battery may exist.

[0091] Steps S105 and S106 allow for highly accurate identification of the lithium-ion battery portion. To further improve accuracy, the battery portion detection unit 114 determines whether the external shape of the region including the dark area matches the standardized external shape of a lithium-ion battery. As a result, as shown in Figure 8, the region R12 in the X-ray image 550 can be detected as a lithium-ion battery.

[0092] Here, since there are multiple standardized external shapes for lithium-ion batteries, all types of external shapes are stored in advance. Then, it is determined whether any of the standardized external shapes among the multiple types match the external shape of the third candidate area.

[0093] Of course, one could determine if the outer diameter matches, but since the shape may differ depending on the orientation of the lithium-ion battery as seen in the X-ray image, the following method may also be used.

[0094] For example, if the shape of the dark area is rectangular, the area of ​​each type of non-standard shape may be determined in advance, and minimum and maximum values ​​for the non-standard shapes may be set. Then, it may be determined whether the area of ​​the third candidate area falls within the range of the minimum and maximum values.

[0095] For example, if the shape of the dark area is circular, it may be possible to determine whether the diameter of the dark area matches the diameter of the non-standard shape.

[0096] (A-3) Effects of the Embodiment As described above, according to this embodiment, lithium-ion batteries can be accurately detected by utilizing the structural characteristics of lithium-ion batteries in which the components are evenly filled within the housing, and by detecting regions with little variation in brightness in the X-ray image of waste.

[0097] (B) Other embodiments Although various modified embodiments have been mentioned in the embodiments described above, the present invention can also be applied to the following modified embodiments.

[0098] (B-1) In the above-described embodiment, an example was given of detecting an image region containing a lithium-ion battery embedded in the object. However, this method is not limited to lithium-ion batteries and can also be applied to detecting other secondary batteries or primary batteries.

[0099] Even with batteries other than lithium-ion batteries (other secondary batteries, primary batteries), some have battery elements densely and evenly packed within a metal casing, and it is expected that the transmitted X-ray dose will be approximately the same. Therefore, if other batteries can be detected by utilizing the structural characteristics of such batteries, the detection method can be applied not only to lithium-ion batteries but also to other secondary batteries or primary batteries.

[0100] (B-2) Variation (Part 1) For example, the brightness of an X-ray image may change due to minute differences in the X-ray irradiation energy of the X-ray imaging device 2, caused by ambient temperature or aging. Therefore, in addition to the configuration of the embodiment described above, the image analysis unit 112 may also include an automatic adjustment unit that corrects the brightness value of the X-ray image as a preprocessing step before X-ray image analysis. An example of the processing of the automatic adjustment unit of the image analysis unit 112 is shown below.

[0101] First, with no object (object to be inspected) present, the X-ray imaging device 2 sets the amount of X-ray irradiation energy and acquires an X-ray image. Here, the X-ray imaging device 2 sets the amount of X-ray irradiation energy emitted by the X-ray source 21 so that the average value of the brightness values ​​of all pixels in the X-ray image becomes the initial setting value (for example, 240 in the case of an image with a brightness gradation of 0 to 255).

[0102] Next, the X-ray imaging device 2 takes an image with the object in place. As a preprocessing step, the automatic adjustment unit of the image analysis unit 112 corrects the overall contrast of the X-ray image so that the average value of the brightness of the background portion of the X-ray image where the object is not visible becomes the initial setting value (for example, 240).

[0103] For example, if the original brightness value of all pixels in an X-ray image is y and the corrected value is x, the automatic adjustment unit corrects the brightness value according to the following formula. Here, we assume that the average brightness value of all pixels in the background (referred to as the "average brightness value of the background") is 250. x = y × (average brightness of the background / initial setting) = y × (250 / 240) ... (1)

[0104] This allows for the correction of the brightness value of each pixel in the X-ray image even if there are minute differences in the amount of X-ray irradiation energy from the X-ray imaging device 2, enabling accurate analysis of brightness variations. As a result, the detection accuracy of lithium-ion batteries is also improved.

[0105] Note that the processing example of the automatic adjustment unit of the image analysis unit 112 is not limited to this, and other methods can also be applied. Another method is briefly described below.

[0106] For example, reference pieces (brightness reference pieces) are placed at predetermined intervals (e.g., every 30 cm) on both sides of the conveyor 3, in positions where they appear in the X-ray image, to serve as a reference for the brightness value in the X-ray image. The reference pieces are made of the same material, are the same size, and have the same thickness. They are placed at equal intervals so that they can be identified as reference pieces when they appear in the X-ray image. In addition, the brightness of the reference pieces that appear in the X-ray image at a certain amount of X-ray irradiation energy (e.g., a fourth threshold; e.g., 120) is known in advance. Since the brightness of the reference pieces that appear in the X-ray image is a fixed value (the fourth threshold), even minute differences in the amount of X-ray irradiation energy directly affect the authentication accuracy.

[0107] The automatic adjustment unit corrects the overall contrast of the X-ray image using a fourth threshold for the brightness of the reference piece, which is known in advance, and the brightness of the reference piece as it appears in the actual X-ray image.

[0108] (B-3) Variation (Part 2) For example, to reduce over-detection or misrecognition of small parts such as screws, the image analysis unit 112 may be configured to detect the image region of the lithium-ion battery according to the size of the waste object 5. An example is given below.

[0109] First, the image analysis unit 112 separates the background region from the region of the object (waste 5) in the X-ray image.

[0110] For example, generally speaking, when comparing the area ratio within an X-ray image, the area of ​​the object is larger than the area of ​​the background. Therefore, it is assumed that the area ratio of the background within the X-ray image is less than 50%. The image analysis unit 112 uses the brightness histogram of the entire X-ray image to determine the peak value as the brightness of the background, and extracts the area darker than that brightness as the object. This allows the background area and the object area to be separated. Although the example uses the brightness histogram of a single X-ray image, it is not limited to this, and for example, a histogram accumulated from all frames of one minute may also be used.

[0111] Next, the image analysis unit 112 selects the lithium-ion battery portion of the X-ray image to be analyzed, taking advantage of the fact that the size of the lithium-ion battery differs depending on the size of the object.

[0112] For example, if the object is large, such as a laptop computer or a robotic vacuum cleaner, the lithium-ion battery inside will also be large. Conversely, if the object is small, such as a cordless earphone or a small drone, the lithium-ion battery will be small.

[0113] Therefore, if the object is large, the image analysis unit 112 will analyze only the black regions above a certain size in the X-ray image, and conversely, if the object is small, it will select only the black regions below a certain size for analysis. Such analysis target selection processing may be added to the processing flow shown in Figure 2.

[0114] Information regarding the type and size of the object is input into the information processing device 1 during X-ray analysis. In addition, information regarding the size of the lithium-ion battery that may be embedded in the object, and the size of the black area to be analyzed corresponding to the size of that lithium-ion battery, are pre-registered in the information processing device 1. When the size of the object is input during analysis, the size of the corresponding black area to be analyzed can be read out.

[0115] Furthermore, the objects on conveyor belt 3 are not limited to just one in a single image; multiple objects, such as both a laptop and cordless earphones, may be captured. Even in such cases, the size and range of the lithium-ion battery can be set for each object area according to the size of the object. [Explanation of Symbols]

[0116] 1: Information processing equipment, 2: X-ray imaging equipment, 3: Conveyor, 5: Waste, 10: Waste disposal system, 11: Control unit, 12: Memory unit, 13: Input unit, 14: Display unit, 21: X-ray source, 22: Detector, 23: X-ray image generation unit, 42: Memory unit, 111: Image acquisition unit, 112: Image analysis unit, 113: Candidate area detection unit, 114: Battery part detection unit, 115: Output unit.

Claims

1. In an image processing device that detects whether or not an object has a battery installed based on an X-ray image of the object, An image acquisition unit that acquires the X-ray image of the object, Based on the image data of the X-ray image, an image analysis unit detects image regions with little variation in brightness among the dark areas where the brightness value is less than a first threshold set in advance, as candidate areas where the battery may be located. An image processing apparatus characterized by comprising:

2. The aforementioned image analysis unit, Based on the brightness frequency distribution within the aforementioned image region, an image region is selected in which the difference between the brightness value at which the cumulative frequency from the darker side reaches a predetermined value and the brightness value at which the cumulative frequency from the brighter side reaches a predetermined value is less than or equal to a second threshold. In the selected image region, an image region is detected as the candidate area by extracting a dark area from a brightness value that is offset by a predetermined value from the brightness value that shows the peak. The image processing apparatus according to feature 1.

3. The image analysis unit determines whether the shape of each candidate area approximates the shape of the battery. The image analysis unit moves the center of the candidate area as the center point of point symmetry or line symmetry, and determines the candidate area based on the agreement between the dark area after the move and the dark area before the move. The image processing apparatus according to claim 2.

4. The image analysis unit determines each candidate area based on the agreement between the shape of the candidate area and the shape of each battery standard type that has been set in advance. The image processing apparatus according to claim 2.

5. In an image processing method for detecting whether or not an object has a battery installed based on an X-ray image of the object, The image acquisition unit acquires the X-ray image of the object, The image analysis unit detects, based on the image data of the X-ray image, image regions that include dark areas where the brightness value is below a first threshold set in advance, and where the variation in brightness is small, as candidate areas where the battery may be located. An image processing method characterized by the following:

6. An image processing program that detects whether or not an object has a battery installed based on an X-ray image of the object, Computers, An image acquisition unit that acquires the X-ray image of the object, Based on the image data of the X-ray image, an image analysis unit detects image regions with little variation in brightness among the dark areas where the brightness value is less than a first threshold set in advance, as candidate areas where the battery may be located. An image processing program characterized by its ability to function in this way.

Images