AUTOMATIC DETECTION OF MANIPULATED METALLIC OBJECTS IN X-RAY IMAGES

Description

[0001] The present invention generally relates to the non-destructive inspection of objects and the detection of potentially hazardous items contained therein. More specifically, the invention relates to the detection of otherwise harmless objects that have been manipulated to conceal hazardous substances, namely battery cells or accumulator cells that have been modified to conceal a hazardous substance, such as an explosive or drugs, from detection. background

[0002] The following introductory description serves only to improve understanding of the invention and should in no way be understood as admitted prior art unless it is expressly identified as such.

[0003] It has become known (e.g., C. Endt et al. in "The Size Matters," Süddeutsche Zeitung, May 17, 2017) that an inherently dangerous organic explosive concealed in an inherently harmless metallic object is difficult to detect in a transmission image, if not highly unlikely. This means that, similar to mimicry, the dangerous "explosive" object is currently barely distinguishable from its surroundings in an optically oriented X-ray image. This means, specifically, that explosives could be concealed in inherently harmless battery cells or accumulators of a mobile electronic device.

[0004] In the following, we will only refer to rechargeable cells, without excluding batteries as primary cells. Batteries in mobile devices, such as laptops, are now predominantly lithium-ion batteries (here referred to as lithium batteries for short). To conceal an explosive in one or more battery cells of a battery pack, either part of the interior or the entire interior of one or more battery cells is replaced with the explosive. This ensures that such a battery pack with manipulated or fake battery cells can still supply the mobile device with sufficient energy for a functional test.

[0005] X-ray-based inspections of hand luggage at airports predominantly produce only two-dimensional X-ray images of the items to be inspected. When inspecting luggage intended for the cargo hold of an aircraft, it is now common practice to use computed tomography techniques, which reliably detect explosives. Inspection systems at security checkpoints for hand luggage screening, which show an operator a two-dimensional X-ray image, should therefore be improved to detect concealed explosives.

[0006] US 2016 00 84 984 A1 discloses a system and method for detecting lithium batteries in inspection objects. This is sufficient if the sole objective is to detect the presence of lithium batteries or batteries that resemble lithium batteries. However, this is not a solution if the sole objective is to detect tampered battery cells. The application of this known system and method would require a closer examination of each mobile device with battery cells to rule out the possibility of explosives concealed within them, or a general prohibition against carrying the corresponding mobile devices on board an aircraft. Neither approach is practical.

[0007] US 2011 02 06 240 A1 also relates to the detection of potentially threatening objects that may be hidden inside objects, such as portable electronic devices. In this case, an object to be inspected is subjected to an imaging process using a line scanner (LS) or computed tomography (CT), and 2D LS or 3D CT data of the object to be inspected is obtained. When analyzing the LS or CT data, the first step is to identify an object to be examined in more detail, such as a laptop, e.g. in a piece of luggage, as an inspection object. This object can then be divided into partitions for further examination by generating one-dimensional eigenprojections of the CT data. Feature vectors of these partitions and the LS or CT image data are used to generate layout feature vectors.One or more layout feature vectors are then compared with training data for threat-containing and non-threat-containing items from a class of the depicted item to determine whether the depicted item contains a threat.

[0008] US 4,539,648 A discloses an imaging X-ray inspection system for inspection objects for detecting agricultural contraband contained in an inspection object (e.g., a piece of luggage or a package). The system is intended to selectively improve the visual representation of objects with a circular cross-section in order to better distinguish these objects from objects with a rectangular cross-section. Based on the assumption that the objects are surrounded by a material that has a different density absorption coefficient product than the objects themselves, it is proposed to calculate a gradient image of the spatially resolved intensity of the X-ray radiation transmitted by the inspection object in order to remove the edges of the objects with a rectangular cross-section in the X-ray image.

[0009] US 2014 / 185755 A1 discloses a system for scanning a shoe for illegal materials. The system includes an X-ray source for projecting X-rays onto the shoe, a detector array for detecting X-rays penetrating the shoe, and at least one metal detector coil for detecting metals in the shoe. The system generates an X-ray image of the shoe by processing the detected X-rays and the data obtained from the at least one metal detector coil. Summary of Revelation

[0010] An object of the present invention is to propose an X-ray inspection system and an X-ray inspection method by means of which the detection of manipulated, generally metallic objects, namely rechargeable battery cells or battery cells, in particular lithium cells, in which an organic hazardous substance, for example an explosive, has been concealed, is improved or at least made possible.

[0011] The object is achieved by the features of claim 1. Further embodiments and advantageous developments are defined in the respective subsequent dependent claims. Features and details defined in connection with the X-ray inspection method according to the invention naturally also apply in connection with the X-ray inspection system according to the invention, and vice versa. Therefore, reference is made to the disclosure of the individual aspects.

[0012] In order to find the technical solution according to the invention, several technical considerations were required on the part of the inventor.

[0013] A lithium cell typically has a cylindrical shape with a thin metal foil casing and an interior made of familiar functional materials (e.g., aluminum, metallic lithium oxide, a highly porous separator, carbon, and copper). If the interior of such a lithium cell is tampered with by filling the cell entirely or partially with an organic explosive, this tampered cell will produce a different X-ray image than that of a genuine lithium cell. Due to its predominantly metallic components, the X-ray attenuation of genuine lithium cells is higher than that of organic materials, such as explosives. Therefore, in a tampered cell, the X-ray image would be significantly different from that expected from a lithium cell due to the organic explosive content, and would therefore be noticeable to an operator in the X-ray image.

[0014] To compensate for the lack of attenuation properties of a manipulated cell, it can be constructed with a significantly thicker metal casing so that the resulting transmission-based X-ray image of the manipulated cell resembles that of a conventional lithium cell. If the metal casing is constructed with a uniform thickness, i.e., as a sleeve of uniform thickness, the resulting X-ray image can appear equally realistic in all viewing directions, even in an X-ray inspection system with multiple viewing directions (so-called multiview systems). This is problematic.

[0015] A cell manipulated in this way is referred to here as a "fake cell." A fake cell is defined as a manipulated rechargeable or battery cell in which a material foreign to the cell's actual function has been concealed within the cell and the cell has been manipulated in such a way that the cell appears as closely as possible to a real rechargeable or battery cell in an X-ray image.

[0016] The inventor(s) have discovered, in an X-ray image of the attenuation or intensity curve of detected X-rays through a fake cell, that it is possible to detect features characteristic of non-manipulated cells and, in particular, anomalies characteristic of fake cells. A particularly suitable attenuation or intensity curve of detected X-rays for analysis is one that runs, if possible, transversely through a battery cell, i.e., orthogonally to its longitudinal axis.

[0017] For example, it has been recognized that the metallic casing or sleeve of a fake cell causes a characteristic peak at the edges of the fake cell in an X-ray image in the attenuation curve or correspondingly in the intensity curve through the fake cell. This results from the fact that the metallic casing or sleeve is irradiated with X-rays tangentially and not perpendicularly at the cell edge. As a result, the X-ray radiation is attenuated to a much greater extent at the cell edge, particularly in the edge region defined by the metallic casing or sleeve, than in the area of the fake cell between the edge regions defined by the metallic casing or sleeve. This is because the radiation there passes perpendicularly essentially through the metallic casing or sleeve, and thus through less metal than in the edge regions, and the X-ray radiation is therefore effectively less attenuated. This effect is all the more significant the thicker the metallic cell casing, i.e. the metallic casing or sleeve, is.This is precisely the case with fake cells, since the lack of attenuation caused by an organic hazardous material, such as explosives, is supposed to be compensated by more metal in the cell casing.

[0018] Finally, during the development of the solution outlined above, it became clear that a key part of the task was also to find a sufficiently robust method for detecting fake cells. Real X-ray images contain a variety of interferences, i.e., the cells to be analyzed are not ideally isolated but are stacked differently, for example, laptop parts and cables can overlap the cells, and metal casings of fake cells and real cells can be of different sizes. The principle proposed here, which underlies the developed detection method, is explained using idealized examples. Therefore, the X-ray images and drawings shown could be misleading in isolation, as the cells analyzed are "exposed."

[0019] It should be noted that the inventive principle is not limited to detecting fake cells containing explosives. Rather, the principle is applicable to detecting any manipulated metallic objects in which an organic material has been concealed. The organic material could, for example, also be drugs. Therefore, the term "fake cell(s)" should not be understood in a restrictive sense, but should generally be used as a representative for any type of fake object manipulated according to the same principle.

[0020] The core idea of the invention is to use the above findings to find metallic fake cells in an X-ray inspection method, in particular in an evaluation method for finding fake cells in a transmission image of an inspection object, such as a two-dimensional X-ray image.

[0021] A first aspect of the present invention relates to a method for locating a manipulated metal object, in which a non-metallic substance is concealed, in two-dimensional (2D) transmission data, in particular 2D X-ray data, of an inspection object containing the object. The metal object can, for example, be manipulated such that the manipulated metal object produces a transmission image comparable to a corresponding non-manipulated metal object, for example a 2D X-ray image. 2D transmission data is data that has preferably been acquired by means of X-raying the inspection object in one spatial dimension along a second spatial dimension and detecting the radiation not absorbed by the inspection object. In the case of 2D X-ray data, these are the attenuation values or intensity values for X-rays, always recorded for one (or more) row(s) along a transport direction.Of course, 2D transmission data can also be derived from three-dimensional (3D) transmission data recorded for spatial elements of the inspection object and used for the method presented here.

[0022] The method comprises the following steps: determining an area containing a metal object in the X-ray data; providing an attenuation curve for or intensity curve of detected X-rays (with which the inspection object was irradiated) along a line through the metal object from the X-ray data; evaluating whether the attenuation curve or intensity curve shows a characteristic anomaly in a predetermined area, for example at edges or in edge regions of the metal object; and if the attenuation curve or intensity curve in the predetermined area shows the characteristic anomaly, triggering an alarm function.

[0023] Preferably, determining the region in the x-ray data containing the metal object comprises: determining that a region in the x-ray data comprises metal based on whether the region causes substantially minimum attenuation of the x-rays and / or on which atomic number or effective atomic number (Z-value or Z-effective) is assigned to a material in the x-ray data.

[0024] Preferably, providing the attenuation curve or the intensity curve along the line through the metal object from the X-ray data comprises: determining a longitudinal direction of the metal object; and creating the attenuation curve or intensity curve along the line through the metal object. Preferably, the line is aligned such that it runs orthogonally to the determined longitudinal direction.

[0025] Preferably, evaluating whether the attenuation curve or, correspondingly, the intensity curve in the predetermined range exhibits a characteristic anomaly comprises at least one of the following steps (a)-(d). It is clear that the intensity curve is essentially complementary to the attenuation curve, and thus all steps for the attenuation curve can be applied accordingly to the intensity curve, taking this into account.

[0026] Step (a): Determine whether the attenuation value increases along the attenuation curve from the edge of the metal object to the center of the metal object.

[0027] Step (b): Determine whether the attenuation curve has a jump at the edges or in the edge regions of the metal object.

[0028] Step (c): Determine whether the attenuation curve has an attenuation maximum in the edge region of the metal object and decreases from there towards the center of the metal object and / or is at a low level compared to the edge regions.

[0029] Step (d): Determine a first derivative of the attenuation curve with respect to location and determine whether the derivative shows a peak in the edge regions of the metal object.

[0030] In connection with the measures explained above, but also for the entire present document, it should be noted that "edge area" is understood here to mean an area starting from the edge of the metal object in the direction of the center of the metal object, whereby the edge area is preferably defined as approximately 10% of the distance starting from the edge to the center.

[0031] The attenuation curve or, correspondingly, the intensity curve as profile signals can be evaluated to recognize or detect a characteristic anomaly using various methods, e.g., machine learning, in which "artificial" knowledge from experience is incorporated into the system. In principle, the system learns patterns and regularities in the training data based on presented examples of fake cells to be recognized, with the knowledge of the examples being generalized after the learning phase. As a result, the system can not only recognize the presented examples but also evaluate unknown, i.e., new, data. The principles of machine learning are familiar to experts in this field and therefore need not be explained in detail here.

[0032] Preferably, triggering an alarm function comprises at least one of the following steps (i)-(iii).

[0033] Step (i): Overlaying the attenuation curve or the intensity curve as well as the line along which the attenuation curve or the intensity curve is present into an X-ray image of the inspection object.

[0034] Step (ii): Initiating a manual check of the inspection object.

[0035] Step (iii): Activation of a visual and / or audible alarm at an inspection facility where the method is applied.

[0036] Preferably, the metal object is an object with a metal sleeve or metal casing. Particularly preferably, the manipulated metal object is a manipulated rechargeable battery cell or battery cell (fake cell), for example, a manipulated lithium battery cell.

[0037] A second aspect of the present invention relates to a processing device for evaluating fluoroscopy images, wherein the processing device comprises a computer unit. The computer unit is configured to carry out a method according to the first aspect of the present invention.

[0038] Preferably, the processing device comprises an output unit configured to execute an alarm function.

[0039] Preferably, the computer unit has a communication interface by means of which the computer unit can be networked with one or more inspection systems for the non-destructive inspection of inspection objects for data communication in order to receive X-ray images from the one or more inspection systems for a (preferably visual) evaluation by means of the data communication.

[0040] A third aspect of the present invention relates to an inspection system which is configured for transporting inspection objects through the inspection system, for carrying out an imaging inspection method for non-destructive inspection of the inspection objects and for providing radiographic images of the inspection objects to a processing device according to the second aspect of the present invention and is connected to the processing device for data communication.

[0041] A fourth aspect of the present invention relates to an inspection system having at least one processing device according to the second aspect of the present invention, which is spatially separated and connected to at least one inspection system according to the third aspect of the present invention for data communication, wherein radiographic images of inspection objects inspected at the at least one inspection system are transmitted to the processing device for visual evaluation.

[0042] A fifth aspect of the present invention relates to a computer program product comprising a computer program having software means for carrying out a method according to the first aspect of the present invention when the computer program is executed on a computer, in particular on a processing device according to the second aspect of the present invention.

[0043] A sixth aspect of the present invention relates to a data carrier containing a computer program product according to the fifth aspect of the present invention.

[0044] A seventh aspect of the present invention relates to a data stream comprising electronically readable control signals that can interact with a programmable computer such that when the computer executes the electronically readable control signals, the computer performs a method according to the first aspect of the present invention. Preferred embodiments

[0045] Further advantages, features, and details of the invention will become apparent from the following description, in which exemplary embodiments of the invention are described in detail with reference to the drawings. The features mentioned in the claims and in the description may be essential to the invention individually or in any combination. Likewise, the features mentioned above and those further explained here may be used individually or in combinations. Parts or components with similar functions or that are identical are sometimes provided with the same reference numerals. The terms "left," "right," "top," and "bottom" used in the description of the exemplary embodiments refer to the drawings in an orientation with a normally legible figure designation or normally legible reference numerals.The embodiments shown and described are not intended to be exhaustive, but rather are exemplary in nature to illustrate the invention. The detailed description is intended to inform those skilled in the art; therefore, well-known structures and methods are not shown or explained in detail in order not to obscure the understanding of the present description. Figure 1 shows a simplified block diagram of an inspection system with a processing device for evaluating X-ray images of inspection objects. Figure 2 shows an inspection system with several (n) inspection systems of the Figure 1at a control point and several (m) evaluation devices at a spatially separated evaluation point. Figure 3 shows a comparison of the attenuation curve of X-rays along a line through genuine lithium cells (left) and through manipulated lithium cells, so-called fake cells, (right) to illustrate and explain the principle proposed here for detecting fake cells. Figure 4 shows another example of the attenuation curve along a line through non-manipulated lithium cells. Figure 5 shows another example of the attenuation curve across fake cells installed in a laptop.Figure 6A shows, from top to bottom, a simplified cross-section of a non-manipulated lithium cell. Below that, a section of a 2D X-ray image of the lithium cell. Below that, the intensity curve of detected X-rays along a line through the X-ray image, and below that, the curve of the derivative of the intensity curve with respect to the location variable. Figure 6B shows, from top to bottom, a simplified cross-section of two manipulated lithium cells lying next to each other, i.e., fake cells. Below that, a section of a 2D X-ray image of the fake cells. Below that, the intensity curve along a line through the X-ray image, and below that, the curve of the derivative of the intensity curve with respect to the location variable. Figure 7 shows a flowchart of an inventive method for detecting fake cells in inspection objects.

[0046] Figure 1shows a simplified block diagram of an inspection system 300 with an evaluation device 100 for evaluating fluoroscopy images B1 ( Figure 3 ), B2 ( Figure 4 ), B3 ( Figure 5 ), B4 ( Figure 6A ) and B5 ( Figure 6B ) of inspection objects O1, O2, O3. The inspection objects O1, O2, O3 in the Figure 1 are shown only symbolically for simplicity.

[0047] The evaluation device 100 can essentially comprise a processing unit 110, an input unit 120 and an output unit 130 as a workstation for an operator.

[0048] The output unit 130 is configured to provide the operator with fluoroscopy images, for example the fluoroscopy images B1 ( Figure 3 ), B2 ( Figure 4 ) and B3 ( Figure 5) of the inspection objects O1, O2, O3 in order to check the contents of the inspection objects O1, O2, O3 for the presence of target objects. Target objects can be dangerous objects such as weapons, explosives, dangerous liquids, and / or contraband and / or drugs, etc. In the present context, the focus is particularly on finding dangerous substances, particularly explosives, concealed in otherwise harmless metallic objects. As already described elsewhere, it is possible to manipulate or falsely reproduce an otherwise harmless metallic object in such a way that the manipulated or false object (fake object) is represented in a two-dimensional X-ray image as similar to or almost identical to the reproduced genuine, i.e. non-manipulated, object. This can accordingly impair the detection rate.In order to improve or facilitate the detection of such counterfeit items, suitable measures are described here using the example of manipulated or fake battery or accumulator cells.

[0049] An important part of the task was to find a sufficiently robust method, since a variety of disturbances can occur in the real image, e.g. the cells can be stacked differently, laptop parts and cables can overlap, the sleeves and cells can be of different sizes and types; in this respect, the images / drawings could be misleading in themselves, since the cells are "exposed" there; The input unit 120 in the Figure 1is essentially configured to input operator inputs to control the evaluation device 100. Furthermore, in response to a displayed X-ray image B1, B2, B3, the operator can input the result of the visual evaluation, such as, for example, that the inspection object O1, O2, O3 is "safe" or that the contents of the inspection object O1, O2, O3 require further inspection, particularly manual examination. A more detailed examination can also be automatically triggered by the processing unit 110 via an alarm function, for example, to examine the affected object for traces of explosives in a more detailed examination (explosive trace detection, ETD). Such a measure is time-consuming and costly and should therefore only be carried out in cases with justified cause.

[0050] The fluoroscopy images B1 ( Figure 3A ), B2 ( Figure 4a ) and B3 ( Figure 5A) of the inspection objects O1, O2, O3 were inspected by the inspection system 200 ( Figure 1 ) using an imaging fluoroscopy technique. In the exemplary embodiment, the inspection system 200 is an X-ray inspection system, as known, for example, from DE 101 49 254 A1. The X-ray inspection system can have one or more radiation planes.

[0051] The inspection system 200 is essentially configured for transporting the inspection objects O1, O2, O3 through the inspection system 200 by means of a transport device T, for example, a conveyor belt. Inside the inspection system 200, the inspection objects O1, O2, O3 are non-destructively X-rayed in a conventional manner. As a result, the inspection system 200 generates two-dimensional X-ray data of the inspection objects O1, O2, O3, from which X-ray images can be generated at the evaluation device 100 and displayed on the display unit 130 for the operator.

[0052] The evaluation device 100 is in the embodiment as an inspection system 300 of the Figure 1spatially separated from the inspection system 200 via a network connection 310 for data communication. For this purpose, the computer unit 110 has a first communication interface 112, and the inspection system 200 has a second communication interface 212, by means of which the computer unit 110 is data-communicatively networked with the one shown, but also with further inspection systems 200 for the non-destructive inspection of inspection objects via the network connection 310. Via the network connection 310, acquired X-ray data from inspection objects O1, O2, O3 inspected at the inspection system 200 are transmitted to the evaluation device 100 for evaluation.In addition to a plurality of inspection systems 200, the inspection system 300 can also have a plurality of evaluation devices 100, so that the workload of evaluating the fluoroscopy data can be optimally distributed among the existing evaluation devices 100 (such a system is described in the . Figure 2 shown). Of course, the evaluation device 100 can in principle be designed in direct spatial proximity to the inspection system 200 or as an integral part of the inspection system 200.

[0053] Figure 2 shows an inspection system 300 with several (n) inspection systems 200.1, 200.2, ..., 200.n of the Figure 1 , which are arranged at a control point 250, and several (m) evaluation devices 100.1, ..., 100.m of the Figure 1, which are located at an evaluation point 150 spatially separated from the inspection systems. The inspection systems 200.1, 200.2, ..., 200.n of the control point 250 are networked via a data network 320 with the evaluation devices 100.1, ..., 100.m of the evaluation point 150 for electronic data exchange in a conventional manner. Of course, several control points 250 can also be networked with one or more evaluation points 150.

[0054] The control point 250 can, for example, be a checkpoint at the entrance to a security area of an airport. At the checkpoint, the hand luggage of airline passengers is checked in a conventional manner using the inspection systems 200.1, 200.2, ..., 200.n of the control point 250. In this case, the x-ray data generated by one of the inspection systems 200.1, 200.2, ..., 200.n in the exemplary embodiment is x-ray data from, for example, a piece of luggage as the object of inspection of a passenger. For evaluation, e.g. with regard to possible security risks or other objects of interest, the x-ray data is transmitted via the network 320 to one of the evaluation devices 100.1, ..., 100.m. There, x-ray images B1-B3, among other things, are generated from the x-ray data and are visually examined by an operator.

[0055] When an inspection object contains an electronic mobile device with battery cells, the operator is particularly challenged. The operator must determine whether the mobile device's battery cells may have been tampered with, thus requiring additional, time-consuming and costly inspections. If more effective evaluation of the X-ray data at the evaluation device 100 can avoid unnecessary additional checks, unnecessary delays at the inspection point 250 can be avoided, thus improving the entire inspection process. Furthermore, additional costs are avoided.

[0056] Figure 3 shows a first embodiment of the improvement proposed here for the evaluation, especially the detection, of possible fake objects in inspection objects.

[0057] Figure 3The upper part (a) shows a first X-ray image B1 of a tray W as a first inspection object O1. The tray W contains two battery packs AP1, AP2, each consisting of 9 cells.

[0058] The first battery pack AP1 consists of an arrangement of a total of 9 lithium cells LZ, three of which are connected in series and the three units with series-connected lithium cells are connected in parallel to each other.

[0059] The second battery pack AP2 is essentially identical to the first battery pack AP1 in terms of the circuitry of the lithium cells. However, a unit with three cells connected in series consists of three manipulated lithium cells, referred to here as fake cells (FZ). This means that the second battery pack AP2 has only two-thirds the capacity of the first battery pack AP1, but is fundamentally functional. Consequently, the second battery pack AP2 cannot be detected solely through a functional test of the mobile device containing this battery pack AP2.

[0060] As discussed elsewhere and in the Figure 3As can be seen, a real lithium cell LZ has a cylindrical shape with a thin shell made of a metal foil and consists inside of the known functional materials, such as an aluminum electrode coated with metallic lithium oxide and a copper electrode coated with carbon, wherein the lithium oxide layer and the carbon layer are both arranged in an electrolyte and separated from each other by means of a highly porous separator.

[0061] If the interior of a manipulated fake FZ cell is filled entirely or partially with an organic explosive, such a fake FZ cell would produce a noticeably different X-ray image than that of a genuine LZ lithium cell. Due to its predominantly metallic components, the X-ray attenuation of the genuine LZ lithium cell is higher than the attenuation of a filling consisting of an organic material, such as an explosive, as is the case with the fake cell. In the case of a fake FK cell, the X-ray image would thus be clearly distinguishable from the X-ray image of a genuine LZ lithium cell due to the organic explosive content, and would therefore be generally easy for an operator to recognize as conspicuous in X-ray image B1.

[0062] To compensate for the lack of damping properties of a fake FZ cell, it can be designed with a much thicker metal sheath, the material thickness of which is adjusted so that the resulting transmission-based X-ray image of this fake FZ cell looks like that of an ordinary non-manipulated LZ lithium cell.

[0063] If the metal casing is made of uniform thickness, i.e., in the form of a sleeve or casing of uniform thickness, the resulting X-ray image can appear equally realistic in all viewing directions, even in a multi-view X-ray inspection system. This is problematic because fake cells could remain undetected and potentially be used to smuggle explosives into secure areas.

[0064] The inventor(s) have recognized that fake cells FZ exhibit a characteristic anomaly in the attenuation curve along a line through the fake cell FZ in an X-ray image, which allows the possible presence of a fake cell FZ to be automatically detected. Such an anomaly in the attenuation curve is particularly easy to detect along a line L that runs essentially transversely or orthogonally to a longitudinal direction LR of the fake cell(s) FZ.

[0065] The metallic cell shell of the fake cell FZ causes a characteristic peak P1, P2 in the attenuation curve along line L through the fake cells FZ in the X-ray image B1, which peaks appear at the left edge R1 and at the right edge R2 of the fake cell FZ. The peak P1, P2 arises because the metal sleeve or metal casing is irradiated tangentially rather than perpendicularly at the cell edge R1, R2. As a result, the X-ray radiation passes through more metal at the cell edge R1, R2 and is correspondingly attenuated significantly more than in the area of the fake cell between the cell edges R1, R2. In the area between the cell edges R1, R2, the metal sleeve or metal casing is irradiated essentially perpendicularly, effectively passing through less metal. Accordingly, the X-ray radiation is less attenuated. This effect and thus the detectable anomaly are all the more significant the thicker the metal sleeve or metal casing of the fake cell FZ is.

[0066] This allows fake cells (FZ) to be detected by searching for the anomaly described above. The anomaly is difficult to detect with the human eye, particularly due to screen resolution, and has therefore not been detected until now.

[0067] In any case, the inventors have recognized that such anomalies occur particularly in fake FZ cells compared to non-manipulated lithium cells, since the lack of attenuation, especially of a hazardous organic material hidden within them, such as an explosive, has to be compensated for by more metal in the cell casing.

[0068] In the lower part (b) of the Figure 3The above finding is plotted using the attenuation curve D(r) along line L with the position variable r. For optimal analysis, the attenuation curve D(r) is preferably examined perpendicular to the longitudinal direction LR of the cells of a battery pack AP1, AP2. It should be noted, however, that the principle described here also applies, i.e., an anomaly indicating tampering can be detected if the curve is examined at an angle (i.e., obliquely) to the longitudinal direction LR.

[0069] In any case, in the lower part (b) the Figure 3 For the left battery pack AP1, it is clearly visible that the damping curve D(r) increases, as expected, from the cell edges to the center M of the battery pack AP1 and decreases from the center M towards the cell edge.

[0070] For the fake cells FZ of the second battery pack AK2 right in the lower part (b) of the Figure 3It is clearly visible that the attenuation curve D(r) at the cell edges R1, R2 of the fake cells FZ exhibits a characteristic peak P1, P2 at the respective local maximum values Dmax1 and Dmax2, i.e., a distinct jump. Subsequently, the attenuation curve D(r) noticeably decreases toward the center M of the battery pack AP2 and the fake cell FZ, or rather, it runs at a significantly lower level. This is anomalous compared to the attenuation curve D(r) for the genuine lithium cells LZ and is referred to here as a possible characteristic anomaly for detecting fake cells FZ.

[0071] In an initial consideration based on the structure of typical fake cells, one approach was to attempt to calculate metal components from the X-ray data in the area of detected battery cells. However, this proved extremely difficult even in simple scenes due to interference with the X-ray signal, which is attenuated by the organic material. The solution proposed and discussed here uses precisely the metal component of a fake cell to detect an anomaly caused by it (anomaly detection), which also works in difficult scenes. One particularly difficult scene to detect is battery cells installed in an electronic device, such as a laptop, which is located in a bag along with many other objects.

[0072] The Figures 4 and 5 each show a further example of the principle proposed here, therefore the Figures 4 and 5no longer explained in detail, but only in the essential points.

[0073] In the upper part (a) of the Figure 4 A second X-ray image B2 shows a tray W as the second inspection object O2. Tray W contains a third battery pack AP3 with six genuine, i.e., non-manipulated, lithium cells LZ arranged side by side.

[0074] In the lower part (b) of the Figure 4 is the attenuation curve D(r) for the X-ray radiation along the line L in the upper part (a) of the Figure 4 by the third battery pack AP3 with the position variable r. The expected attenuation curve D(r) across the six lithium cells LZ is clearly visible. The attenuation increases from the cell edge to the cell center M, where the X-rays must pass through the most metallic cell material, and decreases again from there toward the edge.

[0075] In the upper part (a) of the Figure 5This is a third X-ray image B3 of a box BO as the third inspection object O3. Inside the box BO is a laptop LT, an example of a mobile device, containing a fourth battery pack AP4 with nine adjacently arranged battery cells. Three of the battery cells have been tampered with to conceal an organic material inside them; these cells are fake cells FZ to be discovered.

[0076] In the lower part (b) of the Figure 5 is again the attenuation curve D(r) for the X-ray radiation along the line L through the fourth battery pack AP4 with the position variable r in the upper part (a) of the Figure 5The fake cells FZ are shown transversely to the longitudinal direction LR. Here, too, the expected attenuation curve D(r) across the fake cells FZ is clearly visible. The attenuation peaks at the respective cell edges R and decreases from the respective cell edge R to the respective cell center M of the respective fake cell FZ, and increases again from the respective center M of the fake cell FZ to the edge R.

[0077] Based on the above-mentioned Figures 3-5 According to the findings explained, fake cells FZ can be reliably and automatically identified using a detectable anomaly.

[0078] Figure 6Ashows from top to bottom a simplified cross-section Q1 of a non-manipulated lithium cell LZ, below it a section of a 2D X-ray image B4 of the lithium cell LZ, below it the intensity curve I(r) along the line L through the X-ray image B4 and below it the course of the derivative of the intensity curve I(r) with respect to the location variable (r).

[0079] The non-manipulated lithium cell LZ of the Figure 6Ahas a thin metal shell H1, while the interior is essentially uniformly filled with the functional materials of the lithium cell LZ with a high metal content up to the center M. This results in an intensity curve I(r) starting from the right edge RR or the left edge RL to the center M of the unmanipulated lithium cell LZ, with intensity values decreasing towards the center. Since the attenuation for the X-rays is essentially determined by the thickness of the material passing through them and the cross-section Q1 is circular and thus thickens towards the center, a typical "signature" results in the first derivative of the detected intensity curve I(r) of the X-rays. The first derivative dl(r) / dr of the intensity curve I(r) shows a peak at a left inflection point WL and a right inflection point WR of the intensity curve I(r), which peak is not located at the edge or in the edge region of the lithium cell LZ.Thus, with the findings presented here, even non-manipulated lithium cells LZ can be positively identified or confirmed.

[0080] Figure 6B shows in comparison to Figure 6B from top to bottom a simplified cross-section Q2 of two manipulated lithium cells lying next to each other, i.e. fake cells FZ, below it a section of a 2D X-ray image B5 of these fake cells FZ, below it the intensity curve I(r) along the line L through the X-ray image B5 and below it the curve of the derivative dl(r) / dr of the intensity curve I(r) with respect to the position variable (r).

[0081] The manipulated fake cells FZ of the Figure 6B has a thin metal shell H1 compared to the Figure 6AA thicker metal casing H2, while the interior is essentially uniformly filled up to the center M with an organic material (e.g., explosives) concealed in the fake cell FZ. The thickness of the metal casing H2 defines a left edge region RB1 and a right edge region RB2.

[0082] As noted elsewhere, the edge region is understood here as an area that extends from the edge RL, RR of the cell FZ towards the center M of the cell FZ and whose thickness is at most approximately 10% of the distance from the edge RL, RR to the center M; this definition is generally valid for all examples here.

[0083] In the case of the fake cell, the right edge area RB1 and the left edge area RB2 result in comparison to the non-manipulated lithium cell LZ. Figure 6Aa significant dip in the intensity curve I(r), whereby the intensity curve of the detected X-rays is almost uniform towards the center M of the fake cell FZ due to the lower attenuation properties of the organic material, since the total (effective) thickness of the metal sleeve H2 through which the X-rays have to pass hardly changes. This results in a significantly different signature for the fake cell FZ in the first derivative of the intensity I(r). Particularly significant are the clearly recognizable peaks in the edge areas RB1, RB2, which, compared to the peaks at the inflection points WL, WR, Figure 6A at the edge of the cell and thus clearly in the edge area of the cell, while the turning points are clearly closer to the center of a non-manipulated lithium cell LZ.

[0084] Figure 7shows a flowchart of a method for detecting a manipulated metal object, such as a fake cell FZ of the Figures 3A and 5A , in which a non-metallic substance, such as explosives, is concealed, in two-dimensional X-ray data of an inspection object O1 or O3 containing the metal object FZ. The method comprises the following steps.

[0085] A step S10 with determining a region containing the metal object FZ in the X-ray data. Subsequently, a step S20 with providing an attenuation curve I(r) across the metal object FZ from the X-ray data (cf. Figures 3A-5B ). Subsequently, a step S30 with evaluation of whether the attenuation curve I(r) at edges R1, R2 of the metal object FZ shows a characteristic anomaly P1, P2 (cf. Figures 3B , 4B , 5B). Finally, a step S40 with, if the attenuation curve I(r) at edges R1, R2 of the metal object FZ shows the characteristic anomaly P1, P2, with triggering of an alarm function.

[0086] The alarm function may include at least one of the following steps: (i) Displaying the attenuation curve I(r) and the line along which the attenuation curve I(r) is present in an X-ray image B1, B2, B3 of the inspection object O1, O2, O3 to enable an operator to perform a better visual evaluation; (ii) Initiating a manual inspection of the inspection object O1, O2, O3. For this purpose, the system can be automatically controlled so that the inspection object in question is automatically removed from the access of third parties and transported to a follow-up inspection point. (iii) Initiating a visual and / or acoustic alarm on an inspection system 200 at which the method is applied. This can also ensure the attention of the operator and other security personnel.

[0087] Finally, it should be noted that the principle of the invention is not limited to the detection of fake cells (FZ) containing explosives. Rather, the principle is applicable to the detection of any manipulated metallic objects in which an organic material has been concealed. This could also include drugs, for example. Therefore, in the following, we will not refer to fake cells in a restrictive manner, but rather to fake objects.

[0088] The core idea of the invention is to use the above finding to find metallic fake cells in an X-ray inspection method, in particular in an evaluation method for finding fake cells in a transmission image of an inspection object, such as a two-dimensional X-ray image.

Claims

1. A method for detecting a tampered accumulator or battery cell (FZ) in which a non-metallic substance is concealed and which has been tampered with such that a resulting transmission-based two-dimensional X-ray image of the tampered accumulator or battery cell (FZ) looks like that of a corresponding non-tampered accumulator or battery cell, in two-dimensional X-ray data of an inspection object (O1, O2, O3) containing the battery cell (FZ), the method being performed by a processing apparatus and comprising the following steps: - (S10) determining a region containing a battery cell in the X-ray data; characterized by - (S20) providing the attenuation curve (D(r)) for or intensity curve (I(r)) of detected X-rays along a line (L) through the accumulator or battery cell from the X-ray data; - (S30) evaluating whether the attenuation curve (D(r)) or the intensity curve (I(r)) shows a characteristic anomaly (P1, P2) at edges (R1, R2) or in edge regions (RB1, RB2) of the accumulator or battery cell; and - if the attenuation curve (D(r)) or the intensity curve (I(r)) at the edges (R1, R2) or in the edge regions (RB1, RB2) shows the characteristic anomaly (P1, P2), (S40) triggering an alarm function.

2. Method according to claim 1, wherein (S10) determining the area containing the accumulator or battery cell in the X-ray data comprises: - determining that a region in the X-ray data comprises metal based on whether the region causes a minimum attenuation of X-rays and / or on what nuclear charge number or effective nuclear charge number is assigned to a material in the X-ray data.

3. Method according to claim 1 or 2, wherein (S20) providing the attenuation curve (D(r)) or intensity curve (I(r)) along the line (L) through the accumulator or battery cell from the X-ray data comprises: - determining a longitudinal direction (LR) of the battery cell; and - generating the attenuation curve (D(r)) or intensity curve (I(r)) for a location variable (r) along the line (L) through the accumulator or battery cell, wherein the line (L) is preferably configured such that the line (L) is orthogonal to the determined longitudinal direction (LR).

4. Method according to one of claims 1-3, wherein (S30) evaluating whether the attenuation curve (D(r)) or correspondingly the intensity curve (I(r)) shows a characteristic anomaly (P1, P2) at the edges (R1, R2) or in the edge regions (RB1, RB2) comprises at least one of the following steps: - determining whether the value of the attenuation increases along the attenuation curve (D(r)) from the edge (R1, R2) of the accumulator or battery cell towards the center (M) of the accumulator or battery cell; - determining whether the attenuation curve (D(r)) exhibits a jump (P1, P2) at the edges (R1, R2) or in the edge regions (RB1, RB2) of the accumulator or battery cell; - determining whether the attenuation curve (D(r)) has a maximum attenuation (Dmax1, Dmax2) in the edge regions (RB1, RB2) of the accumulator or battery cell and, starting from this, decreases towards the center (M) of the accumulator or battery cell and / or runs at a lower level compared to the edge regions (RB1, RB2); - determining a first derivative of the attenuation curve (D(r)) according to the location (r) and determining whether the derivative (D(r) / dr) shows a peak in the edge regions (RB1, RB2) of the accumulator or battery cell in each case.

5. Method according to any one of claims 1-4, wherein (S50) triggering an alarm function comprises at least one of the following steps: - superimposing the attenuation curve (D(r)) or the intensity curve (I(r)) as well as the line (L) along which the attenuation curve (D(r)) or the intensity curve (I(r)) is present into an X-ray image (B1, B2, B3, B4, B5) of the inspection object (O1, O2, 03); - triggering a manual check of the inspection object (O1, O2, O3); and - triggering a visual and / or audible alarm at an inspection apparatus (200) at which the method is applied.

6. Method according to any one of claims 1-5, wherein the tampered accumulator or battery cell (FZ) is an object with a metal sleeve (H2) or metal shell.

7. Method according to any one of claims 1-6, wherein the tampered accumulator or battery cell (FZ) is a tampered lithium battery cell (LZ).

8. Processing apparatus (110) for analyzing transmission images (B1, B2, B3, B23), wherein the processing apparatus (110) is configured to perform a method according to any one of claims 1 to 7.

9. Processing apparatus (100) according to claim 8, wherein the processing apparatus (110) is further connected to an output unit (130), wherein the output unit (130) is configured to implement an alarm function.

10. Processing apparatus (110) according to claim 8 or 9, wherein the processing device (110) has a communication interface (112) by means of which the processing device (110) can be networked with one or more inspection apparatuses (200) for the non-destructive inspection of inspection objects (O1, O2, O3) for data communication in order to receive fluoroscopic images (B1, B2, B3) from the one or more inspection apparatuses (200) for visual evaluation by means of the data communication.

11. Inspection apparatus (200) for transporting inspection objects (O1, O2, O3) through the inspection apparatus (200), for performing an imaging inspection method for non-destructive inspection of the inspection objects (O1, O2, 03) and for providing transmission images (B1, B2, B3) of the inspection objects (O1, O2, O3) to a processing device (110) according to any one of claims 8-10 and for this purpose is connected to the processing device (110) for data communication.

12. Inspection system (300) with at least one processing apparatus (110; 110.1, ..., 110.m) according to one of the claims 8-10, which is spatially separated with at least one inspection apparatus (200; 200.1, 200.2, ..., 200.n) according to claim 11 for data communication, wherein transmission images (B1, B2, B3) of inspection objects (O1, O2, O3) inspected at the at least one inspection system (200; 200.1, 200.2, ..., 200.n) are transmitted to the processing apparatus (110; 110.1, ..., 110.m) for visual evaluation.

13. Computer program product comprising a computer program having software means for implementing a method according to any one of claims 1 to 7 when the computer program is executed on a computer.

14. Data carrier comprising a computer program product according to claim 13.

15. Data stream comprising electronically readable control signals capable of interacting with a programmable computer such that when the computer executes the electronically readable control signals, the computer performs a method according to any one of claims 1 to 7.