X-ray detector device and device for the x-ray inspection of products, especially foodstuffs

Description

[0001] The invention relates to a device for X-ray inspection of products, in particular foodstuffs.

[0002] For the X-ray inspection of moving products, line detectors are typically used, which are arranged transversely to the direction of movement of the products to be inspected. Instead of moving the product, the entire X-ray inspection device or at least the relevant X-ray detector device can be moved relative to the product to be inspected. The product to be inspected is scanned by means of the line detector, which detects the X-rays generated by one or more X-ray sources, and the image data generated line by line is converted into an image of the product to be inspected. To generate the image, the image data can be processed in a suitable manner. The image thus generated can be inspected using image processing. In particular, the image of the product to be inspected can be inspected to determine whether one or more predetermined characteristics are present or fulfilled.For example, a food item, such as a piece of meat, can be examined to determine whether it contains unwanted foreign bodies such as bone fragments, metal shavings from processing machines, glass fragments, plastics, stones or the like.

[0003] For this purpose, it is known to use spectrally integrating line detectors (non-spectral detectors), which detect practically the entire width of the X-ray spectrum of the X-ray radiation generated by the respective X-ray source. Such line detectors have a comparatively high spatial resolution in the range of, for example, 0.2 mm over a total detector width of, for example, 200 to 800 mm or more. The line detectors can be constructed in the form of modules that can be coupled with very small gaps of only a few pixels (for example, one to two pixels per module edge) until a desired scan width is achieved. With such non-spectral detectors, even very small foreign bodies or contamination can be detected. This type of detector also requires no cooling and is cost-effective to manufacture.

[0004] However, such non-spectral detectors only produce grayscale values ​​due to spectral integration. A grayscale value depends on the attenuation of the X-rays as they pass through the product being examined. The attenuation, in turn, depends on the thickness of the product and the material properties.

[0005] This type of sensor is particularly suitable for detecting the smallest, most absorbent foreign bodies, such as metal fragments.

[0006] The contrast of the image generated by line detectors can be improved using the dual-energy method. This method involves two line detectors whose scan images are superimposed. The line detectors capture different spectral ranges of the X-rays passing through the product. This is achieved by using at least one X-ray filter positioned in the beam path in front of one of the non-spectral detectors. However, such filters only function as high-pass filters and cannot be manufactured with sufficient flexibility with regard to the desired filter edge. In addition, they also attenuate the X-rays to be detected in the desired spectral range. By capturing different spectral ranges separately, the image signals from the respective line directors contain different information.Through weighted overlay (e.g., adding the weighted image data with the correct sign), an overall image can be created that offers better contrast for detecting specific foreign body materials than a single-energy image. However, with a fixed X-ray filter, contrast can only be improved for one or more specific materials. Therefore, dual-energy methods are not very flexible in their application, as the X-ray filter must be selected appropriately depending on the application.

[0007] In addition, the spectral ranges of the X-rays detected by the two line detectors usually overlap, so that each of the two image signals contains part of the same information. This prevents optimal contrast enhancement.

[0008] However, the dual-energy method allows for the masking of product areas consisting of only two materials using a suitable combination of the two image signals. For example, the contrast in the area of ​​a foreign body (i.e., a first material) within the product (i.e., a second material) can be optimized. However, this only applies to an essentially homogeneous product made of a single material (or a combination of materials with very similar X-ray attenuation properties) that contains foreign bodies made of another material (with different X-ray attenuation properties).

[0009] Furthermore, spectrally resolving line detectors have been developed in recent years, which can also be coupled in modules. However, such spectral line detectors, which enable sufficiently large scan widths of 200 to 800 mm or more, are currently only available with a relatively coarse spatial resolution, i.e. with a relatively large pixel pitch, for example of 0.8 mm. Such spectral line detectors are capable of capturing the entire spectral width of the X-ray radiation to be detected, for example in the range from 20 keV to 160 keV. These detectors provide a large number of energy channels, for example up to 256, for spectral resolution. Such a spectral line detector therefore makes it possible to generate a number of partial images corresponding to the number of energy channels. The information contained in each of the partial images can be used and processed in a variety of ways.In particular, this information enables the recognition of materials or material combinations as well as the detection of layer thicknesses.

[0010] In addition to the disadvantage of relatively coarse spatial resolution, such spectral line detectors are expensive and larger than simple non-spectral line detectors. Furthermore, operating spectral line detectors requires significantly more energy than operating non-spectral detectors, and the high energy consumption necessitates complex cooling.

[0011] Another significant disadvantage of spectral line detectors is the long start-up time, which can be in the range of 30 minutes. This long start-up time is due to the fact that the spectral line detector requires a certain amount of time to ensure sufficient charge separation.

[0012] Furthermore, US 2013 / 0079918 A1 discloses a sorting system for sorting materials, particularly metals. The system operates with two detectors, one for detecting the X-rays penetrating the materials to be sorted, and another for detecting fluorescence radiation, which is generated by the materials, primarily near the surface, and emitted in all directions. The fluorescence detector is arranged such that it detects the portion of the fluorescence radiation that is emitted obliquely upwards, essentially vertically from above, when the materials are irradiated. Both the fluorescence detector and the transmitted-light detector are designed as spectrally resolving sensors, although the respective spatial resolution and spectral resolution can vary.Furthermore, WO 2017 / 205914 A1 discloses a device for X-ray inspection of products that has a spectrally resolving line detector. The measured spectra are used both as energy data and, after integration, as gray values. The spatial resolution is therefore identical for the energy data and for the integrated gray values.

[0013] Based on this prior art, the object of the invention is to provide a device for X-ray inspection of products, in particular foodstuffs, which can be produced cost-effectively and enables improved detection of foreign bodies in a product to be examined.

[0014] The invention solves this problem with the features of patent claim 1. Further embodiments emerge from the subclaims.

[0015] The invention is based on the finding that at least one non-spectrally resolving line detector with higher spatial resolution and one spectrally resolving line detector with lower or the same spatial resolution can be advantageously combined in order to achieve high detection accuracy at the higher spatial resolution.

[0016] The device according to the invention for X-ray inspection comprises an X-ray detector device, an evaluation and control unit, and a radiation generation device with at least one X-ray source for generating X-rays with a predetermined spectral width, wherein the X-rays radiate through a product to be examined, which is moved in a direction of movement at a predetermined speed. The X-ray detector device comprises at least one first line detector with a first discrete spatial resolution (iea larger number of pixels per unit length), which is designed to detect the broadband X-radiation along one or more parallel first detection lines transverse to a direction of movement of a relative movement between a product to be examined and the X-radiation detector device and which detects the X-radiation passed through the product to be examined over a predetermined detection width in a non-spectral manner and generates first image data, wherein the first image data comprise a gray value for each discrete location.

[0017] It should be noted that the term "X-ray radiation" in the present description includes any type of high-energy electromagnetic radiation, in particular X-ray radiation in the range above 100 eV, in particular above 1 keV, as well as terahertz radiation, i.e. radiation in the range from approximately 0.3 THz to 30 THz.

[0018] Instead of a single such non-spectral line detector with higher spatial resolution, i.e., a small pixel pitch, two or more such line detectors can also be used, which are configured in a conventional manner to generate a dual-energy image, wherein at least one of these line detectors has a filter for filtering the X-rays to be detected. Instead of a filter, a corresponding device for X-ray inspection with such an X-ray detector device can also use a further X-ray source which has different spectral characteristics than a first X-ray source, in particular has the maximum of the spectrum at a different energy value.

[0019] The X-ray detector device further comprises at least one second line detector with a second discrete spatial resolution that is smaller than the first discrete spatial resolution, wherein the at least one second line detector detects the X-ray radiation along one or more parallel second detection lines, which run parallel to the first detection lines, over the predetermined detection width in a discrete spectrally resolved manner and generates second image data, wherein the X-ray radiation quanta are assigned to a plurality of energy channels for spectral resolution, depending on their energy, and wherein the second image data for each discrete location comprise a spectral value for each energy channel.

[0020] Each of the line detectors can also have multiple lines. This allows multiple image lines of a product to be scanned to be recorded and read simultaneously. The multiple lines can also be used to implement the Time Delay Integration (TDI) method to improve the signal-to-noise ratio.

[0021] The image signals of all line detectors are fed to the evaluation and control unit, which is designed, among other things, to evaluate the first and second image data in such a way that at least one predetermined feature of the product to be examined is detected by a combination of the information contained in the first and second image data with the first discrete spatial resolution. The evaluation and control unit can be designed as a combination of hardware and software and, in particular, can comprise an image processing unit with a sufficiently fast image processing processor. The image processing unit generates one or more images from the image data of the line detectors and analyses the images with regard to at least one predetermined feature. In doing so, the evaluation and control unit or the image processing unit uses information which the second, ieThe spectrally resolving line detector provides the overall image to be analyzed and / or analyzes a generated overall image. In this way, the advantages of the spectrally resolving line detector can be utilized to generate an overall image and / or to analyze an overall image that also contains at least the information from the image data provided by the at least one first, non-spectral-resolving line detector.

[0022] This advantage can be utilized, in particular, even when the second discrete spatial resolution of the at least one spectrally resolving line detector is significantly lower than the first discrete spatial resolution of the at least one non-spectrally resolving line detector. The at least one predefined feature for which the evaluation and control unit examines a previously generated overall image can be the position of a foreign body within the product, the material of the foreign body, the thickness of the product, the material of the product, the thickness of material layers of the product, or the position of overlapping areas of multiple products.

[0023] According to a preferred embodiment of the invention, the line detectors are provided in a common housing. Instead of a housing, the line detectors can of course also be arranged on any open support in a fixed position relative to each other. The line detectors (with or without a housing; with or without a support) can also be provided in a common radiation protection housing. This allows a device for X-ray inspection of moving products to be implemented simply and cost-effectively, particularly with regard to radiation protection.According to one embodiment of the invention, the evaluation and control unit controls the at least one first and the second line detector in such a way that a synchronous acquisition of the image data at the first and second acquisition lines takes place depending on the speed of the relative movement between the product to be examined and the X-ray detector device in order to obtain line image data at the first and second acquisition lines which essentially correspond to the same beam path through the product to be examined. With such synchronized generation of the image data from the individual line detectors, an overall image can be generated with less effort, since at most an adaptation, e.g. interpolation of the image data (e.g. only from one line detector), in the direction of the acquisition lines is required. In contrast, the image data from the lines of the line detectors can be directly combined, e.g. added in a weighted manner.The speed of movement must be known with sufficient accuracy, as must the distance between the line detectors in the direction of movement.

[0024] According to another embodiment, the evaluation and control unit controls the at least one first and the second line detector such that asynchronous data acquisition occurs at the first and second acquisition lines, wherein line image data is acquired at the first and second acquisition lines at a line scanning rate that is preferably equal to or greater than the dimension of the pixels of the respective line detector in the direction of the relative movement divided by the speed of the relative movement. This allows the product to be examined to be scanned seamlessly (seen in the direction of movement). When generating an overall image from the image data of several line detectors, an adaptation of the image data, for example by interpolation, must be carried out if necessary, also viewed in the direction of movement.

[0025] It should be noted at this point that when generating an overall image from the image data of two or more line detectors, the evaluation and control unit or the image processing unit can apply any necessary operation to correct the distances between the line detectors, their different pixel pitches, angular errors between the line detectors, a different distance between the line detectors and the X-ray source, and the like. Required operations include, in particular, (linear or nonlinear) interpolation, a geometric transformation (to correct angular errors in non-parallel detection lines), and weighting of the image data (e.g., in the case of different detector sensitivities).

[0026] During the asynchronous acquisition of the image data from the at least one first, non-spectral line detector and the second spectral line detector, the evaluation and control unit can be designed such that the image data is fed to the image processing unit as a single data stream. This is particularly useful if the evaluation and control unit is decentralized and the image processing unit is located far away from a data acquisition unit to which the image data from the individual line detectors are fed as individual data streams. For this purpose, the data acquisition unit can provide the image data with time information containing the acquisition times in absolute form or relative to a reference time. For example, time information can be assigned to each image line in the shared image data stream.The image data from image lines of the various line detectors can also be appropriately identified in the data stream as originating from a specific line detector. Such data transmission can, of course, also be used with synchronous scanning.

[0027] As already briefly explained above, the evaluation and control unit can create an overall image from the image data of the at least one first line detector and the second line detector, in particular by interpolating and / or geometrically transforming the image data of the at least one second line detector and / or the image data of the at least one first line detector. The overall image thus contains information from the two different line detector types.

[0028] According to a further embodiment of the invention, the at least one first line detector and the second line detector comprise a plurality of similar modules arranged in a row at abutting points in the direction of the respective detection line, wherein the abutting points of the at least one first line detector and the abutting points of the second line detector are offset from one another. This allows the blind spots of one line detector to be at least partially compensated for by the image data of the other line detector. However, in many cases, the second line detector cannot provide the higher spatial resolution of the first line detector, nor can the first line detector provide the spectral resolution of the second line detector.However, this disadvantage is acceptable, particularly for the embodiments explained below, in which the image data of the second line detector are used to generate partial image data having a freely selectable spectrum, and these partial image data are combined with the image data of the first line detector to form a dual-energy overall image, or in which the image data of the second line detector are used to obtain information relating to overlapping areas or the product thickness, which information is then used to determine variable threshold values ​​for evaluating an overall image.

[0029] This embodiment allows for a compact design because the modules are arranged in a straight line, and the larger modules of the second, spectrally resolving line detector (due to the heat sinks) can be arranged close to the smaller modules of the at least one first, non-spectral-resolving line detector. To avoid such blind spots, only a staggered arrangement of the modules would be possible, which, however, would result in a significantly larger design.

[0030] According to a further variant of the invention, the evaluation and control unit can weight the spectral values ​​of the energy channels of the image data of the second line detector for each discrete location, in particular multiplying them by a factor in each case, and add the weighted spectral values ​​to a total spectral value. In this way, the evaluation and control unit can generate dual-energy image data or multiple-energy image data from the first image data of the at least one first line detector and the thus processed second image data of the second line detector.The image processing unit can freely select the spectrum of the processed second image data by applying any weighting (a weighting factor of zero corresponds to a sharp delimitation of a corresponding spectral range), thus generating a dual-energy image that is adapted to detect very specific features of the product being examined. For example, the spectrum of the processed second image data can be adapted to detect foreign bodies made of a specific material.

[0031] For example, weighting can be performed by weighting a specific spectral range, e.g., the energy channels from 20 to 40 keV, with a factor of 1, while all other energy channels are suppressed by weighting with zero. Such a sharply defined spectrum in a lower energy range could not be generated with an X-ray filter because these filters are only available as high-pass filters and, in addition, have a low slope.

[0032] The invention thus enables a dual-energy method in which the line detector, which usually has a higher spatial resolution, detects the entire spectrum of the X-rays, while the line detector, which usually has a lower spatial resolution, provides a freely selectable spectrum. This spectrum can also be freely selected after the spectrally resolved detection of the X-rays.

[0033] This principle can also be implemented by having the spectrally resolving line detector perform preprocessing itself. For example, the spectral line detector can read only selected energy channels in the form of the image data signal, which corresponds to a weighting of the unread energy channels with a factor of zero. The weighting of the read energy channels and the determination of a total spectral value can then be performed by the evaluation and control unit according to a predetermined weighting profile.

[0034] The spectrally resolving line detector can also be designed so that not only specific energy channels (individual ones) are read out, but that the line detector simultaneously adds the spectral values ​​of the individual channels to a total spectral value. The total spectral value can then be weighted by the evaluation and control unit (with a single factor). The spectrally resolving line detector can also be designed so that the spectral width of the energy channels is adjustable. In this way, it is possible (by manual input or an input from the evaluation and control unit) to control the spectrally resolving sensor so that it delivers a spectral value for an adjustable spectral range, for example, a spectral range from 20 keV to 40 keV. This spectral value can then be weighted directly by the evaluation and control unit and used to generate a dual-energy image.

[0035] It should be noted at this point that the evaluation and control unit can perform any desired weighting (pixel by pixel) of the spectrum supplied by the spectrally resolving sensor by using a weighting profile that has a weighting factor for each energy channel. The weighting can simultaneously select an energy channel if the respective weighting factor is not equal to zero, and actually weight the spectral value of the respective energy channel with the respective weighting factor. If the weighting factors for neighboring energy channels used to generate partial image data for a dual-energy image are constant, the spectral values ​​of these energy channels can also be added to form a total spectral value before the total spectral value is used with the (constant) weighting factor to generate the dual-energy image.

[0036] According to one embodiment of the invention, information can be stored in the evaluation and control unit, or the evaluation and control unit can have access to information, regarding which weightings of energy channels are preferably suitable for detecting one or more specific, predetermined features of the product to be examined. In particular, this information can be present in the form of weighting profiles, wherein a weighting profile can assign a (basically freely selectable) weighting factor to each energy channel. Each weighting profile can be assigned a specific feature, which can be advantageously detected, in particular with good contrast, using the dual-energy overall image generated with this profile. For example, foreign bodies made of a specific plastic, e.g.made of polyethylene, can be detected with good contrast, while another weighting profile is particularly suitable for detecting bone fragments in the overall image in question.

[0037] The evaluation and control unit can also evaluate the image data from the at least one first line detector and the image data from the second line detector multiple times for a specific product to be examined, in particular for detecting various predetermined features. For each evaluation, dual-energy image data or multiple-energy image data, or a corresponding overall image, is generated using a different weighting or weighting profile for the spectral values ​​of the energy channels of the image data from the second line detector. This allows various evaluations to be performed with a single scan of the product.For example, an overall image generated using a first weighting profile may be examined for the presence of metal foreign bodies, and an overall image generated using a second weighting profile may be examined for the presence of plastic foreign bodies.

[0038] According to a further embodiment, the evaluation and control unit can evaluate the image data from the second line detector to detect product regions with different layer thicknesses and / or regions of overlapping products and identify such regions in the image data from the at least one first line detector or in combined image data. To inspect the regions thus identified, the evaluation and control unit can use a threshold value that is set depending on the gray values ​​of one or more of these regions in the relevant overall image. For example, the average gray value of such a local region in the overall image can be determined, and the threshold value above which the evaluation and control unit detects a foreign body can be determined depending on this gray value according to a predetermined rule.Such a dynamic threshold offers the advantage of accurately detecting foreign bodies or other features of a product under inspection, regardless of the thickness of the product or regardless of overlapping product areas.

[0039] By appropriately evaluating the image data from the second, spectrally resolving line detector, the thickness of a known material can be determined. Furthermore, by appropriately evaluating this data, it is even possible to determine which material or material combination is involved. This is because every material produces a characteristic energy spectrum when irradiated with an X-ray spectrum that has a sufficient spectral width of, for example, 20 to 160 keV. Thus, by comparing the energy spectrum detected without a product and the energy spectrum detected with a product, the characteristic X-ray attenuation curve as a function of energy for the respective (initially unknown) material can be determined pixel by pixel.By comparing this characteristic curve with known characteristic curves for the respective materials, the material can be identified. The evaluation and control unit can compare each detected energy spectrum with stored characteristic curves of products, for example, using a correlation analysis.

[0040] While the course of a detected energy spectrum allows conclusions to be drawn about the material, the thickness of the product and also differences in thickness depending on the location can be determined from the attenuation of the detected radiation (in one energy channel or several or all integrated energy channels) with knowledge of the respective material. In this way, overlap areas of two or more products can also be determined. For this purpose, the evaluation and control unit can access information containing the relationship between the attenuation and the thickness of the material in question. This can be values ​​for the energy-dependent absorption coefficient µ(E), where E denotes the energy of the X-ray radiation. This can be used to determine the thickness of a known material using the Lambert-Beer law.

[0041] According to a further embodiment of the invention, the evaluation and control unit for monitoring the state of a selected line detector determines image data of the line detector to be monitored and image data of a line detector selected as a reference line detector, preferably in each case without the presence of a product, wherein the evaluation and control unit sums the spectral values ​​of one, several, or all energy channels to determine the image data of the second line detector, and wherein the evaluation and control unit compares the image data of the line detector to be monitored with image data of the reference line detector and generates a "not ready signal" upon detection of inadmissible deviations between the image data of the line detector to be monitored and the image data of the reference line detector.

[0042] The energy channels whose spectral values ​​are added to form a total spectral value can be selected in such a way that the same spectral sensitivity curve results for both line detectors. For example, if it is known that a non-spectral-resolving line detector to be monitored has a sensitivity in the range from 20 keV to 80 keV and the spectral-resolving line detector (i.e. the reference line detector) has a sensitivity of 20 keV to 160 keV, only those energy channels of the spectral-resolving line detector that cover the sensitivity range of the non-spectral-resolving line detector will be used to determine the respective total spectral value, i.e. the energy channels in the range from 20 keV to 80 keV. This procedure is, of course, independent of whether the spectral-resolving line detector serves as the reference line detector or represents the line detector to be monitored.

[0043] Instead of the evaluation and control unit summing the spectral values ​​of the relevant energy channels, the spectrally resolving sensor itself can provide this functionality. In particular, it can provide only the selected energy channels in the image data signal or transmit the sum spectral value in the image data signal to the evaluation and control unit.

[0044] It is also possible to consider a known spectral sensitivity curve of a non-spectrally resolving line detector (within the spectral width of the active energy channels of the spectrally resolving line detector). For this purpose, the evaluation and control unit can appropriately weight the active spectral values ​​of the respective energy channels before adding them. In this way, the spectral sensitivity of the spectrally resolving line detector can be approximated to the spectral sensitivity of a non-spectrally resolving line detector.

[0045] Using this monitoring method, the readiness of the second, spectrally resolving line detector can be checked and determined when it is switched on. As explained above, it can take a long time, e.g., 30 minutes, for the spectrally resolving sensor to achieve sufficient charge separation. For this purpose, the gray value of a non-spectral-resolving sensor can be compared pixel by pixel with the corresponding gray value of the spectrally resolving sensor. The gray value of the spectrally resolving sensor is determined by summing the spectral values ​​of all energy channels (for each pixel).If the evaluation and control unit knows the ratio of the corresponding pixel gray values ​​of the two sensors when fully (correctly) functioning, it can generate the "not ready signal" if the gray value of a comparison pixel of the non-spectrally resolving line detector differs by more than one permissible value from the gray value of the corresponding comparison pixel of the spectrally resolving line detector.

[0046] The "not-ready" signal can be generated even if an unacceptable deviation is detected for a single pair of corresponding comparison pixels. It should be noted that a "not-ready" signal can be generated in all cases where the image data of the line detector being monitored deviates from the image data of the reference line detector by more than a specified value.

[0047] The reference image data evaluation and control unit can perform this pixel-by-pixel comparison using corresponding comparison pixels. A comparison pixel can consist of one or more pixels of a line detector to be considered in whole or in part. The comparison pixels are selected such that two corresponding comparison pixels cover the same distance on the respective detection lines, i.e., the width of each pair of corresponding comparison pixels is identical, with the distance covered by a comparison pixel on the respective detection line resulting from a parallel shift of the distance covered by the respective corresponding comparison pixel on the other respective detection line.To determine the gray value of a comparison pixel, the values ​​of all pixels that lie entirely or partially within the width of the comparison pixel are added together, whereby the gray value of a pixel that does not lie with its entire width within the width of the respective comparison pixel is weighted by the ratio with which the pixel in question lies with its width within the width of the comparison pixel.

[0048] If, for example, the wider pixels of the lower spatial resolution second line detector are each used as comparison pixels, and a narrower pixel of the higher spatial resolution first line detector lies entirely within the width of the comparison pixel of the second line detector and two pixels of the higher spatial resolution first line detector lie with half their width within the width of the comparison pixel of the second line detector, the gray value of the corresponding comparison pixel of the first line detector, which consists of the whole and the two half pixels of the first line detector, is determined by adding the gray value of the whole pixel and the gray values ​​of the half pixels weighted by a factor of 0.5.

[0049] In this way, a simple comparison of the image data signals of the line detector designated as the reference line detector and the line detector to be monitored can be carried out.

[0050] The device according to the invention for X-ray inspection of products, in particular foodstuffs, is designed to irradiate a product to be examined, which is moved in a direction of movement at a predetermined speed of movement, with the generated X-ray radiation.

[0051] According to one embodiment, the at least one first line detector and the second line detector are arranged such that the one or more second detection lines of the second line detector are not in the shadow of the at least one first line detector, and the one or more first detection lines of the first line detector are not in the shadow of the second line detector. The one or more first and second detection lines can be as close together as possible in the direction of movement of the product to be examined. This results in product regions being scanned by the at least two line detectors that are as close together as possible in the direction of movement.Fluctuations in the movement speed, which are unknown to the evaluation and control unit, therefore have little impact on the generation of an overall image, since the identical product area is irradiated and scanned at only a very short time interval. Furthermore, the beam path follows approximately the same path through the product, resulting in a small detector device.

[0052] The at least one first line detector and the second line detector can be arranged overlapping, viewed in the radiation direction, with the one or more first and second detection lines being spaced at different distances from the at least one X-ray source. This distance can be selected to be as small as possible, depending on the size of the at least one first and second line detector. This results in a very low overall height.

[0053] The invention is explained in more detail below with reference to the drawing. The drawing shows: Fig. 1 is a schematic representation of a first embodiment of a device for inspecting products with a first embodiment of an X-ray detector device with two non-spectrally resolving line detectors and one spectrally resolving line detector; Fig. 2 is a side view ( Fig. 2a ) one and a top view ( Fig. 2b ) to a further embodiment of an X-ray detector device, which comprises a combination of a non-spectrally resolving and a spectrally resolving line detector; Fig. 3 shows a section of a schematic plan view of the two lines of a further embodiment of an X-ray detector device similar to the embodiment in Fig. 2 ; and Fig. 4 a schematic section of a variant of the embodiment in Fig. 3 . Fig. 5 shows an exemplary energy spectrum generated by a pixel of a spectrally resolving line detector; Fig. 6 shows a weighting profile for weighting the energy channels of an energy spectrum according to. Fig. 5 ;

[0054] Fig. 1 shows a schematic representation of a first embodiment of a device 100 for X-ray inspection of products 102, in particular foodstuffs, with a radiation generating device 104 with at least one X-ray source 106 and with an X-ray detector device 108 with two non-spectrally resolving line detectors 110, 112 and one spectrally resolving line detector 114.

[0055] The X-ray source 106 generates a fan-shaped X-ray beam 116, which has a center plane E perpendicular to a direction of movement B in which the products 102 to be examined are moved by the X-ray beam 116. The X-ray beam 116 has an aperture angle in the plane of movement E, which is configured such that the entire width of the product 102 to be examined is penetrated by the X-ray beam 116. A conveying device (not shown), for example, a conveyor belt, can be provided to move the product 102.

[0056] The non-spectrally resolving line detectors 110, 112 each have a single row 118, 120 of pixels that detect the X-ray radiation of the X-ray beam 116 along a respective detection line. In this context, this is referred to as a detection line, even if the pixels of the respective row 118, 120 have a finite extension. The non-spectrally resolving line detectors 110, 112 have a higher discrete spatial resolution than the spectrally resolving line detector 114, which also comprises a row 122 of pixels that detect the X-ray radiation of the X-ray beam 116 along a detection line. The discrete spatial resolution of the line detectors 110, 112, ie the number of pixels per unit length or the pixel pitch, can be, for example, 0.2 mm and thus four times as large as the pixel pitch of the line detector 114, which can be, for example, 0.8 mm.

[0057] The non-spectrally resolving line detectors 110, 112 face each other with their detector rows 118, 120, which are each provided at an edge of a carrier 124, whereby, viewed in the direction of movement B, a distance remains between the detector rows 118, 120. This distance corresponds approximately to the width of the detector row 122 of the spectrally resolving line detector 114. The detector row 122 is provided approximately centrally on a carrier 126, which can also carry heat sinks and other components. The heat sinks can also form the carrier 126. The detector rows 118, 120 can, as in Fig. 1 shown, have the same distance from the X-ray source 106. The spectrally resolving line detector 114, which has a larger design than the non-spectral resolving line detectors 110, 112 due to the required heat sinks, is located below the line detectors 110, 112, as seen in the direction of the X-ray beam 116. Since the X-ray radiation detected by the line detectors 110, 112 and 114 should radiate through the product 102 as far as possible along the same beam path, it is advantageous to arrange the detector lines 118, 120 and 122 with the smallest possible distance, as seen in the direction of movement B, or within the smallest possible opening angle of the X-ray beam 116 in a plane that is not perpendicular to the center plane of the fan-shaped X-ray beam 116.

[0058] The line detectors 110, 112, 114 can, as in Fig. 1 shown, be provided in a common housing 128, which can be designed as a radiation protection housing. The housing 128 has an opening 130 on its upper side, ie, the side facing the X-ray source 106, which allows the X-ray beam 116 to penetrate into the housing in the direction of the pixel rows 118, 120, 122 of the line detectors 110, 112, 114.

[0059] Instead of a single spectrally resolving line detector 114, two or more spectrally resolving line detectors can also be provided. This can be advantageous if the non-spectral-resolving line detectors are each designed to detect a different maximum spectral width. For example, one of the spectrally resolving line detectors can have a maximum spectral width of 20 keV to 160 keV with a spectral resolution of 256 energy channels, and another spectrally resolving line detector can have a maximum spectral width of 20 keV to 80 keV, also with a resolution of 256 energy channels. The additional spectrally resolving line detector thus has twice the spectral resolution of the first spectrally resolving line detector.

[0060] The line detectors 110, 112, 114 each generate an image data signal, which is fed to an evaluation and control unit 132. The evaluation and control unit 132 can have a data acquisition unit 134 and an image processing unit 136. The image data signals from the line detectors 110, 112, 114 are fed to the data acquisition unit 134. The image data signals can be combined by the data acquisition unit 134 into a single signal or image data stream, which is fed to the image processing unit 136. The image processing unit 136 is designed for further processing and analysis of the image data. The data acquisition unit 134 can also be designed to control the line detectors 110, 112, 114 in a suitable manner, in particular with regard to the sampling times.For this purpose, the data acquisition unit 134 can supply a clock signal to each of the line detectors, whereby the image data acquisition by the line detectors can then be synchronized with the clock signal.

[0061] In the context of synchronous image data acquisition, it is particularly possible for the non-spectrally resolving line detectors 110, 112 to detect X-rays, depending on the movement speed of the product 102 to be examined, that have penetrated the product to be examined essentially along the same path, albeit offset in time. By arranging the line detectors 110, 112 such that the detector lines 118, 120 are as close together as possible, speed changes unknown to the control and evaluation unit 132 do not adversely affect the synchronicity of the data acquisition. The same naturally also applies to a synchronized scanning of a product 102 using the spectrally resolving line detector 114.

[0062] The image processing unit 136 can process the image data acquired by the line detectors 110, 112, 114 in the following manner. Fig. 1 In the embodiment shown, the image data from the non-spectrally resolving line detectors 110, 112 can be evaluated according to a known dual-energy method. For this purpose, the line detector 110 can have an X-ray filter 138, which is provided in the beam path between the X-ray source 106 and the detector row 118. As a result, image data can be acquired by means of the non-spectrally resolving line detectors 110, 112, each of which has a gray value resulting from the integration of the entire detected X-ray spectrum, depending on the intensity of the detected X-ray radiation. As is known per se, with such an embodiment, a higher contrast can be achieved when detecting foreign bodies in a product 102 if the foreign bodies consist of a material that has a different absorption coefficient than the product itself.

[0063] However, the image data of the spectrally resolving line detector 114 provides additional spectral information with which the higher-resolution image that can be obtained from the image data of the non-spectral-resolving line detectors 110, 112 can be improved, or with which one or more higher-resolution images can be generated from the image data of the non-spectral-resolving line detectors 110, 112, from which specific properties of the product to be examined, for example, possibly present foreign bodies made of different materials, can be better recognized. In particular, the evaluation and control unit 132 or the image processing unit 136 in a device for X-ray inspection of products according to Fig. 1 Information about the material or material composition of the product being irradiated can be obtained from the spectral image data of the spectrally resolving line detector 114. Information regarding the thickness of the material or the total thickness of multiple layers of the same material can also be obtained in this way. How such an evaluation of the image data can be performed is explained further below.

[0064] However, the provision of a spectrally resolving line detector not only has advantages in conjunction with two or more non-spectral resolving sensors, as described above with reference to the embodiment of an X-ray inspection device according to Fig. 1 Rather, the appropriate evaluation of the image data supplied by a spectrally resolving line detector also offers advantages when these are combined with image data from a single non-spectral-resolving line detector or when the image data from the single non-spectral-resolving line detector are processed for the purpose of simple analysis using the information obtained from the image data of the spectrally resolving line detector.

[0065] Such embodiments with a single non-spectral resolving line detector may have a similar structure as in Fig. 1 shown, whereby only one of the two line detectors 110 or 112 is omitted.

[0066] Fig. 2 shows schematically the arrangement of a non-spectrally resolving line detector 110 together with a spectrally resolving line detector 114 within such an X-ray detector device 200. The illustration of the remaining components, in particular the housing and the evaluation and control unit, has been omitted. The arrangement of the line detectors 110 and 114 takes place, as in the embodiment in Fig. 1 , so that the detector lines 118 and 122 have a small distance in the horizontal direction and also a small distance in the vertical direction (see the side view according to Fig. 2a ).

[0067] Out of Fig. 2b , which shows the two line detectors 110 and 114 in a partial plan view, it can be seen that each of the line detectors 110, 114 is constructed in the form of modules, each of which is coupled in the direction of the respective detector row 118, 122. The spectrally resolving line detector 114 consists of modules 114 i (the numerator i denotes any natural number greater than or equal to 1), which have a greater width than the modules 110 k (the numerator k denotes any natural number greater than or equal to 1) of the spectrally resolving line detector 110. Of the modules of the spectrally resolving line detector 114, only two modules are shown, namely modules 114 1 and 114 2. These can, for example, have a width (seen in the direction of the detector row 122) of 10 cm. Of the modules of the non-spectral resolving line detector 110, modules 110 1 , 110 2 , 110 3 and 110 4 are shown.These modules have half the width (seen in the direction of the detector row 118) of the modules of the line detector 114. However, this leads to the disadvantage that at every second joint 140 of the modules 110 1 , 110 2 , 110 3 and 110 4 of the line detector 110 there is also a joint 142 of the modules 114 1 and 114 2 of the line detector 114. Since each detector has an edge region in the immediate vicinity of its side edge that is not covered by pixels, the spatial resolution of the respective line detector is no longer given at a joint. Typically, the edge region of each module in which no pixels are provided has a width of approximately one to three pixels. This means that for each line detector 110, 114, information covering a width of two to six pixels is missing in the respective image data at a joint.

[0068] It is therefore advisable to offset the modules of the line detectors 110, 114 from one another in such a way that the joints of the modules of one of the line detectors are not located at the joints of the modules of the other of the line detectors.

[0069] This measure is even clearer from Fig. 3 which schematically illustrates a section of the detector row 118 of the spectrally resolving line detector 114 and the detector row 122 of the non-spectral-resolving line detector 114 in a plan view. Three modules of each line detector 110, 114 are shown (partially). Since in the embodiment shown, the modules of the non-spectral-resolving line detector 110 are also half as wide as the modules of the spectrally resolving line detector 114, an offset of the modules from one another never results in a position at which a joint between both line detectors is present at the same time. In the embodiments of the line detectors according to Fig. 3 the pixel pitch of the spectrally resolving line detector 114 is twice as large as the pixel pitch of the non-spectral resolving line detector 110. The edge dimensions of the (square) pixels of both sensors thus also differ by approximately a factor of two.

[0070] To prevent the blind spots of a detector line from coinciding between two adjacent pixels, the line detectors or the respective modules can also be arranged offset in such a way that, in particular, at the position of a blind spot between two adjacent larger pixels of the non-spectral-resolving line detector 114, there is not simultaneously a blind spot between adjacent pixels of the non-spectral-resolving line detector 110. This procedure is described in Fig. 4 , which shows a section of the detector lines 118, 122 based on the schematic, enlarged representation in Fig. 3 shows, once again more clearly.

[0071] As explained above, the evaluation and control unit 132 can use the image data signals from the line detectors to monitor one of the line detectors (selected line detector) by comparing it with another of the line detectors (reference line detector). This is particularly useful for monitoring the readiness of the spectrally resolving line detector 114, since such line detectors require a longer start-up time (e.g., 30 minutes) before usable image data is output. The start-up time can be determined, in particular, by the necessary creation of sufficient charge separation.

[0072] For such a comparison, the image data of the reference line detector can be compared, particularly pixel by pixel, with the image data of the selected line detector. The comparison can be performed, for example, in such a way that the evaluation and control unit 136 knows a target ratio of the signal values ​​supplied by corresponding comparison pixels (for example, corresponding target ratios for one or more comparison pixel pairs are stored in the evaluation and control unit) and generates a "not ready" signal if one or more comparison pixel pairs do not meet the target ratio.

[0073] The term "comparison pixel" is introduced in this context because the selected line detector and the reference line detector may have different pixel sizes or a different pixel pitch (see the Fig. 3 und 4 ). In particular, the pixels of the one of these two line detectors that has the larger geometric dimensions can be used as the comparison pixel. The respective pixel values ​​can then be used directly (without further processing) as the value of the comparison pixel of this line detector. If the other line detector has a different pixel size, the evaluation and control unit 136 can use all (smaller) pixels of this line detector to determine a corresponding value for this comparison pixel, which - viewed in the direction of the detector line - lie within the width of the comparison pixel of the line detector with the larger pixel dimensions.

[0074] This should be emphasized again Fig. 4 for monitoring the spectrally resolving line detector 114 with the larger pixel dimensions using the non-spectral resolving line detector 110 as a reference detector for an embodiment according to Fig. 2 explained in more detail. A pixel with its actual dimensions, for example, 0.8 mm wide, is used as a comparison pixel for the spectrally resolving line detector 114. This width covers the hatched areas of three pixels of the non-spectral-resolving line detector 110, which together form the comparison pixel of the line detector 110. To form the target ratio, the evaluation and control unit 136 forms the ratio of the pixel values ​​supplied by the line detector 114 for the respective pixel (or comparison pixel) and the proportional pixel values ​​supplied by the line detector 110 for the three actual pixels that form the corresponding comparison pixel of the line detector 110.In this case, the fully covered actual pixel (shown completely hatched) is taken into account with its unchanged pixel value and the two outer actual pixels of the line detector 110, which are only partially covered by the comparison pixel of the line detector 114, are taken into account with the proportional pixel value with which the pixel in question (seen in the direction of the detector line) lies within the width of the comparison pixel of the line detector 114.

[0075] This procedure is performed pixel by pixel for each comparison pixel of the line detector 114 to be monitored and the corresponding comparison pixel of the reference line detector 110.

[0076] However, instead of selecting the actual pixels with the larger geometric dimension as comparison pixels, any other width of a detector line or the actual pixels covered by this width can also be selected as comparison pixels.

[0077] If, during such monitoring, it is determined that the line detector to be monitored does not provide correct results, i.e., the device 100 for X-ray inspection of products or the X-ray detector device 108 generates a "not ready signal," a higher-level control unit (not shown), to which the "not ready signal" is supplied and which is designed to control an entire processing system comprising the device 100, can shut down the processing system and / or generate an alarm signal.

[0078] However, the higher-level control unit can also be configured so that the system continues to operate despite the presence of a "not ready" signal. The evaluation and control unit 136 of the X-ray inspection device 100 continues to operate in such an emergency mode, with only the correctly functioning line detector(s) being controlled and / or the relevant image data signals being used for evaluation. Such an emergency mode can be maintained, for example, until a specific product batch has been processed.

[0079] In this emergency operating mode, the evaluation and control unit 136 can also provide further measures, for example, a special control of the at least one X-ray source 106. For example, to increase the detector accuracy of the remaining line detector(s), in particular to improve the signal-to-noise ratio, the power of the X-ray source 106 can be increased. Thus, the partially existing redundancy of spectrally resolving line detectors and non-spectrally resolving line detectors can be utilized in an emergency operating mode to continue operating the X-ray inspection device 100, albeit with reduced capabilities. However, if these capabilities or properties are sufficient to ensure adequate inspection of the products to be examined, depending on the application, the emergency operating mode represents a considerable advantage.

[0080] The following explains which information the spectrally resolving sensor 114 can provide, which leads to improved detection accuracy when processing the image data of the higher spatially resolving non-spectral-resolving sensors 110, 112.

[0081] Fig. 5 shows, by way of example, the image data supplied by the spectrally resolving line detector 114 for a specific pixel. Here, the entire detected energy range, for example, from 20 keV to 80 keV, is divided into a specific number of energy channels, with each energy channel having a specific (usually constant) spectral width, which results from the width of the entire detected spectral range divided by the number of energy channels, for example, 128 or 256 energy channels. The line detector 114 supplies a spectral value for each energy channel, which is Fig. 5 is labeled "Counts" because such a spectrally resolving line detector usually counts individual photons and assigns the registered photons to a specific energy channel depending on their energy. Fig. 5 The curve shown corresponds to a typical bright image generated by the line detector 114 without the presence of a product in the beam path of the X-ray beam 116.

[0082] The spectral values ​​are transmitted as image data with the image data signal to the evaluation and control unit 136. The evaluation and control unit 136 can evaluate this image data in a variety of ways to obtain information, either to use this image data for an improved evaluation of the image data from the non-spectral-resolution line detectors 110, 112 or to combine the image data from the spectral-resolution line detector 114 with the image data from the non-spectral-resolution line detectors 110, 112 to form an overall image, which, optionally using additional information obtained from the image data from the non-spectral-resolution line detector 114, enables improved image evaluation.

[0083] The spectral resolution capability of the line detector 114 can be exploited to acquire partial image data for generating a dual-energy image. For this purpose, the evaluation and control unit 136 can perform any desired weighting of the spectral values ​​pixel by pixel. Such weighting can be achieved by assigning a factor to each individual energy channel, which is multiplied by the respective spectral value. This can also achieve a sharp limitation of the spectrum by assigning a factor of zero to selected energy channels.

[0084] Fig. 6 shows a weighting profile which has a separate weighting factor for each energy channel. This profile provides a constant weighting, for example, with a factor of one, for a number of energy channels at the lower range of the entire spectrum. All other energy channels are weighted with a factor of zero. The spectral values ​​weighted in this way can be added to generate a sum spectral value. The sum spectral values ​​can then be used as partial image data to generate a dual-energy image, wherein a non-spectrally resolving line detector, for example the line detector 110 in the embodiment according to Fig. 2, can supply further partial image data for generating the dual-energy image. The evaluation and control unit 136 combines the grayscale image data supplied by the non-spectrally resolving line detector 110 with the partial image data from the spectrally resolving line detector 114 to create the desired dual-energy image. For this purpose, the evaluation and control unit 136 weights the respective partial image data pixel by pixel. This weighting can already be performed for the image data supplied by the spectrally resolving line detector 114 as part of the weighting of the energy channels.

[0085] This approach offers the advantage that, unlike in the case of a non-spectral resolving line detector, the energy channels selected to form the sub-image data for a dual-energy image can be weighted not only with a constant factor but with a variable factor.

[0086] As already explained above, the spectrally resolving line detector 114 can also be configured to preselect which energy channels are transmitted to the evaluation and control unit 136 as part of an image data signal. For example, the line detector 114 can be set manually or by the evaluation and control unit 136 to output only certain energy channels as an image data signal. The line detector 114 can also be configured to output the selected energy channels in an integrated form, i.e., to add up the spectral values ​​of the selected energy channels.

[0087] In this case, the evaluation and control unit 136 has less effort to process the image data of the line detector 114.

[0088] Thus, the non-spectrally resolving line detector 114 enables the generation of a dual-energy image using a flexible spectrum. This can be determined by simply evaluating the image data signal from the line detector 114, or the line detector 114 can be controlled in such a way that it already provides corresponding spectrally limited image data or even sum spectral values ​​(see above).

[0089] In this way, a variable dual-energy image can be generated using a single non-spectrally resolving line detector 110, which nevertheless has a high spatial resolution, and a spectrally resolving line detector 114, which, however, has only a relatively low spatial resolution. The spectrum of the spectrally resolving line detector 114 can be varied such that certain features of a product to be examined are more clearly visible in the dual-energy image, for example, with higher contrast.

[0090] The evaluation of the image data from the two line detectors 110, 114 obtained by means of a single scan can also be carried out by performing multiple partial evaluations. For example, different dual-energy images can be generated using differently weighted image data from the spectrally resolving line detector 114. For example, the spectrum of the image data from the line detector 114 can be selected in one evaluation (for example, by appropriate weighting) so that foreign bodies made of a specific material, for example, steel, can be detected with high contrast. In a further evaluation, the spectrum of the image data from the line detector 114 can be selected differently, for example, to generate a dual-energy image in which foreign bodies made of a different material, for example, polyethylene, can be detected with high contrast.

[0091] In addition to generating partial image data for a dual-energy image, the one or more spectrally resolved line detectors offer further advantages. In particular, the evaluation of a dual-energy image generated in this way, or the evaluation of a dual-energy image generated exclusively using image data from sensors with non-spectral resolution, can be improved. These improvements take effect at the image evaluation level, i.e., during the analysis of an already generated overall image (such an overall image is always a grayscale image).

[0092] The spectral resolution provided by such a line detector makes it possible, using suitable evaluation methods, to obtain information about the material or materials of which the product under investigation is made. However, this information can only be determined pixel by pixel with the lower spatial resolution of the spectrally resolving line detector. Furthermore, it is possible to obtain information about the thickness of one or more material layers of which the product under investigation is made from the image data of at least one spectrally resolving line detector.

[0093] Information about the material or materials the product to be examined is made of can be obtained by various methods of evaluating the spectral data supplied by the spectrally resolving line detector. For example, by shining a light through a product made of a known material or a known combination of materials, it is possible to determine the spectral profile of the attenuation caused by the product. To do this, a bright image can first be detected, i.e. the spectral profile of the X-ray radiation I 0 (E k ,xi ) (I 0 denotes the radiation intensity detected without product; E k denotes the energy of the X-ray quanta in energy channel k; xi denotes the location in the line detector, in particular pixel i), when no product is being shined through.The measurement result I(E k ,xi ) (I denotes the radiation intensity detected with the product) when irradiating a product together with the bright image thus provides the attenuation D(E k , xi ) caused by the absorption in the product, whereby the attenuation results from the Lambert-Beer law: . D E k x i = ln I E k x i / I 0 E k x i = μ E k · d i where di denotes the thickness of the product in the beam path of the radiation component detected by pixel xi, and µ(E k ) denotes the absorption coefficient of the material in the beam path. In the case of a material combination with an at least approximately constant composition within the beam path, an average absorption coefficient can be used for the material combination in question. Since the thickness di represents only a constant factor (per pixel), a characteristic absorption behavior curve In(I / I 0 ) results that is independent of the thickness and can be determined using the detected spectrum (i.e., the spectral values ​​per pixel) and a previously detected spectrum of a bright image. This spectral curve can be known to the evaluation and control unit for a product to be examined without foreign bodies (hereinafter also referred to as a "good product"), i.e.corresponding information can be stored in the evaluation and control unit or it can have access to corresponding externally available information.

[0094] Thus, by comparing the measured spectral intensities (measurement of the bright image and measurement of the product) with one or more known spectral intensity curves for each pixel, the evaluation and control unit can determine whether a known material is contained in the relevant beam path of the product. For example, if the product to be examined is yogurt or cheese, the evaluation and control unit can confirm for each pixel of the spectrally resolving line detector whether the spectral curve matches the expected product. If the measured spectral curve deviates from the known curve by more than a permissible deviation, it can be concluded from the image data of the spectrally resolving sensor that a foreign body is contained in the partial volume of the product corresponding to the respective pixel.

[0095] However, if it can be confirmed that there are no unacceptable deviations from the expected characteristic spectral profile at the respective pixel, a good product can be assumed in the respective beam path. Since the absorption coefficient µ(E k ) of the respective material is also known and can be stored in the evaluation and control unit or at least made accessible to it, the respective thickness of the product or the length of the beam path through the product (relative to the respective pixel xi ) can be calculated.

[0096] With this information, the evaluation and control unit can evaluate a gray image, which was generated, for example, using the dual-energy method explained above, using a dynamic threshold value that is used to detect foreign bodies.

[0097] In X-ray inspection of products to detect foreign bodies, i.e., any type of undesirable component or element of a product to be examined, it is common practice to use a static threshold. For example, if a foreign body exhibits higher absorption than the surrounding material of the product to be examined, the static threshold will be set within a range that, on the one hand, generates the lowest possible probability of a false alarm while, on the other hand, allowing sufficiently reliable detection of foreign bodies. For example, an average gray value can be determined from the overall image of the product, which is then multiplied by a predetermined factor to determine the static threshold.

[0098] Since the grayscale image to be analyzed can exhibit large areas of significantly different brightness or gray values ​​in the case of a product with unknown thickness variations or in a case where multiple products, such as meat cuts, may overlap (the absorption of X-rays by the product depends on the material thickness), the use of such a static threshold value leads to problems. However, if it is known from the information provided by the spectrally resolving line detector that only "permissible" material is present even in the darker areas, the evaluation control unit can assume that at least the larger darker areas do not represent foreign bodies, but rather that they are variations in the product's thickness or overlapping product areas.This allows the evaluation and control unit to set a separate threshold for analyzing each of these areas. This dynamic threshold can vary from area to area of ​​the gray image being analyzed. This can reduce the likelihood of a false alarm.

[0099] The combination of at least one non-spectrally resolving line detector with a higher spatial resolution and at least one spectrally resolving line detector with a lower spatial resolution can thus be used to achieve an advantageous generation of an overall image, in particular a dual-energy image or multiple-energy image, from the respective image data and / or to enable an advantageous evaluation of a dual-energy image or multiple-energy image generated in this way or of a dual-energy image or multiple-energy image generated in a manner known per se by using image data from non-spectrally resolving line detectors. List of reference symbols

[0100] 100 X-ray inspection device 102 Product 104 Radiation generation device 106 X-ray source 108 X-ray detector device 110 Non-spectral line detector 110 k module (index k greater than or equal to 1) 112 Non-spectral line detector 114 Spectral line detector 114 i module (index i greater than or equal to 1) 116 Fan-shaped X-ray beam 118 Pixel row 120 Pixel row 122 Pixel row 124 Carrier 126 Carrier 128 Housing 130 Opening 132 Evaluation and control unit 134 Data acquisition unit 136 Image processing unit 138 X-ray filter 140 Joint (modules 110a, 110b, 114c) 142Joint point (modules 114a, 114b, 114c 200X-ray detector device

Claims

1. Device for the X-ray inspection of products, in particular foodstuffs, comprising an X-ray detector device (108), comprising an evaluation and control unit (132), and comprising a radiation generation device (104) which has at least one X-ray source (106) for generating X-rays of a predetermined spectral width, wherein the device is designed such that the X-rays radiate through a product to be examined (102), the product being moved in a direction of movement at a predetermined speed of movement, (a) wherein the X-ray detector device (108) has at least one first line detector (110, 112) which has a first discrete spatial resolution, is designed to capture the X-rays along one or more parallel first capture lines transverse to a direction of movement of a relative movement between the product to be examined (102) and the X-ray detector device (108), and is designed such that it detects, in a non-spectrally resolved manner, the X-rays that have passed through the product to be examined (102) over a predetermined capture width and generates first image data, wherein the first image data comprise a gray value for each discrete location, (b) wherein the X-ray detector device (108) has at least one second line detector (114) which has a second discrete spatial resolution which is smaller than the first discrete spatial resolution, wherein the at least one second line detector (114) is designed such that it captures the X-ray radiation that has passed through the product to be examined (102) along one or more parallel second capture lines which run parallel to the first capture lines, and detects said radiation in a discrete spectrally resolved manner over the predetermined capture width and generates second image data, wherein the X-ray quanta, for spectral resolution, are assigned, depending on their energy, to a predetermined number of energy channels, and wherein the second image data for each discrete location comprise a spectral value for each energy channel, and (c) wherein the evaluation and control unit (132), to which the first and second image data are supplied, is designed to evaluate the first and second image data such that at least one predetermined feature of the product to be examined (102) is detected via a combination of the information contained in the first and second image data and at least the first discrete spatial resolution.

2. Device according to claim 1, characterized in that the predetermined feature of the product to be examined (102) is at least one feature from the group: position of a foreign body within the product, size of the foreign body, material of the foreign body, thickness of the product, material of the product, thickness of material layers of the product, position of overlapping regions of products (102).

3. Device according to any of the preceding claims, characterized in that the line detectors (110, 112; 114) are provided on a common carrier (124, 126) and / or in a common housing (128) and / or in a common radiation protection housing.

4. Device according to any of claims 1 to 3, characterized in that the evaluation and control unit (132) is designed such that it controls the at least one first and the at least one second line detector (110, 112; 114) such that a synchronous capturing of the image data at the first and second capture lines takes place depending on the speed of the relative movement between the product to be examined (102) and the X-ray detector device (108), in order to obtain, at each of the first and second capture lines, line image data which substantially correspond to the same radiation path through the product to be examined (102).

5. Device according to any of claims 1 to 3, characterized in that the evaluation and control unit (132) is designed such that it controls the at least one first and the at least one second line detector (110, 112; 114) such that asynchronous data capturing takes place on the first and second capture lines, wherein line image data is preferably captured on each of the first and second capture lines at a line sample rate which is greater than a pixel pitch of the relevant line detector divided by the speed of movement of the product to be examined (102).

6. Device according to any of the preceding claims, characterized in that the evaluation and control unit (132) is designed such that it creates an overall image from the image data of the at least one first line detector (110, 112) and the at least one second line detector (114), in particular by interpolating and / or geometrically transforming the image data of the at least one second line detector (114) and / or the image data of the at least one first line detector (110, 112).

7. Device according to any of the preceding claims, characterized in that the at least one first line detector (110, 112) and the at least one second line detector (114) each comprise, in the direction of the corresponding capture line, a plurality of similar modules (110k; 114i) arranged in a row next to one another at joints,and in that the joints (140) of the at least one first line detector (110, 112) and the joints (142) of the at least one second line detector (114) are offset from one another.

8. Device according to any of the preceding claims, characterized in that (a) the evaluation and control unit (132) is designed such that it weights the spectral values of all of the energy channels of the image data of the at least one second line detector (114) for each discrete location and adds the weighted spectral values to form a total spectral value or (b) in that the at least one second line detector (110, 112) is designed such that it outputs only image data which have spectral values of predetermined energy channels for each discrete location, and in that the evaluation and control unit (132) is designed such that it weights them and adds the weighted spectral values to form a total spectral value, or (c) in that the at least one second line detector (114) is designed such that for each discrete location it only adds image data which have spectral values of predetermined energy channels to form a total spectral value, and the evaluation and control unit (132) is designed to weight this total spectral value, and in that the evaluation and control unit (132) is designed such that it generates dual-energy image data or multiple-energy image data from the first image data of the at least one first line detector (110, 112) and the thus processed second image data of the second line detector (114).

9. Device according to claim 8, characterized in that information is stored in the evaluation and control unit (132) or the evaluation and control unit (132) has access to information about which weightings of energy channels are preferably suitable for detecting one or more specific predetermined features of the product to be examined (102).

10. Device according to claim 8 or claim 9, characterized in that the evaluation and control unit (132) is designed such that it evaluates several times the image data of the at least one first line detector (110, 112) and the image data of the at least one second line detector (114) for a specific product to be examined (102), in particular for detecting various predetermined features, dual-energy image data or multiple-energy image data being generated for each evaluation in each case by using a different weighting of the spectral values of the energy channels of the image data of the second line detector.

11. Device according to any of the preceding claims, characterized in that the evaluation and control unit (132) is designed such that it evaluates the image data of the at least one second line detector (114) to identify product regions with different layer thicknesses and / or regions of overlapping products and identifies such regions in the image data of the at least one first line detector (110, 112) or in combined image data, and in that the evaluation and control unit (13) is designed such that it uses a threshold value for inspecting the regions thus identified, which threshold value is set depending on the gray values of one or more of these regions.

12. Device according to any of the preceding claims, characterized in that the evaluation and control unit (132) is designed such that, in order to monitor the state of a selected line detector (110, 112, 114), it determines image data of the line detector (110, 112, 114) to be monitored and image data of a line detector (110, 112, 114) selected as a reference line detector, preferably in each case without the presence of a product (102), the evaluation and control unit (132) preferably being designed such that, in order to determine the image data of the at least one second line detector (114), it totals the spectral values of one, a plurality of or all of the energy channels, and in that the evaluation and control unit (132) is designed such that it compares the image data of the line detector (110, 112, 114) to be monitored with image data of the reference line detector (110, 112, 114) and generates a "not ready signal" if inadmissible deviations are identified.

13. Device according to claim 12, characterized in that the evaluation and control unit (132) is designed such that it compares, pixel by pixel, by using mutually corresponding comparison pixels, the reference image data with the image data of the line detector (110, 112, 114) to be monitored, the comparison pixels being selected such that two mutually corresponding comparison pixels cover the same width on the respective capture lines, and the gray value of a pixel which does not lie with its entire width within the width of the relevant comparison pixel being weighted, when determining the gray value of the corresponding comparison pixel, using the ratio with which the pixel lies with its width within the width of the comparison pixel.

14. Device according to any of the preceding claims, characterized in that the at least one first line detector (110, 112) and the second line detector (114) are arranged such that the one or more second capture lines of the at least one second line detector (114) are not in the shadow of the at least one first line detector and the one or more first capture lines of the first line detector (110, 112) are not in the shadow of the second line detector (114).

15. Device according to claim 14, characterized in that the at least one first line detector (110, 112) and the at least one second line detector (114), viewed in the direction of radiation, are arranged in an overlapping manner, the one or more first and second capture lines being at different distances from the at least one X-ray source (106).