Method, computer-readable storage medium, device and system for selecting a photographic image

Abstract

A computer-implemented, in particular user-guided, method for selecting a photographic image (F) for use in production control, wherein the method comprises the following steps: a) Receiving from a multitude of input images (E1,...,En) specifying an object to be captured (30) in a multitude of lighting settings, wherein the lighting settings are specified by different configurations of an RGB-W lighting unit (20) and include a white image (E0); b) Selecting at least one candidate recording (K1,...,Km) from the initial recordings (E1,...,En) based on a first selection criterion; c) Defining a template image section (A0) within the white image (E0); d) Generating at least one target image section (A1,...,Am) in the at least one candidate image (K1,...,Km) that corresponds to the template image section (A0); e) Applying at least one image processing algorithm, in particular contour recognition, blob analysis, region analysis, code analysis or model matching, with a first set of parameters and / or parameter values ​​to the at least one target image section (A1,...,Am) to generate processed candidate images (K'1,...,K'm); f) Selecting at least one preselection recording (V1,...,Vk) from the set of processed candidate recordings (K'1,...,K'm) based on a second selection criterion; g) Check whether a pre-selection recording (V1,...,Vk) exists that meets a final screening criterion and, if so, select such a pre-selection recording (V1,...,Vk) as the final recording (F); otherwise h) Repeating steps e) to g), wherein the at least one preselection recording (V1,...,Vk) is used as the at least one candidate recording (K1,...,Km) and in step e) a second set of parameters and / or parameter values, in particular specified by a user, is used;wherein at least one of the selection criteria depends on an evaluation result of at least one image analysis algorithm, namely contour recognition, wherein a collection of reference images is used for the evaluation of the at least one selection criterion, wherein contours are extracted from object edges of the reference images during contour recognition and stored as reference contours, which are searched for in the multitude of input images, wherein the method further comprises calculating a value that indicates a match with a reference contour and wherein, if the value exceeds a threshold, the area under consideration is regarded as a match, wherein an image is rated as good if a predefined number of matches are found and wherein the image is selected as at least one candidate image if the image is rated as good.

Description

[0001] The invention relates to a method for selecting a photographic image for use in production control, a corresponding computer-readable storage medium, and a device and system for carrying out the method.

[0002] Cameras are used for various purposes in industrial manufacturing facilities, particularly in quality assurance. Here, cameras take photographic images (hereinafter also referred to as images) of the objects to be inspected, which are then analyzed using specialized image processing techniques. In this way, for example, the presence of plastic rivets and seals can be automatically detected, weld nuts identified, and the alignment of body parts checked before welding, all within the automotive industry.

[0003] Black and white cameras are frequently used in the aforementioned application areas because, depending on the intended use, they are less expensive to purchase. Furthermore, black and white cameras offer the advantage of producing images with a comparatively small amount of data.

[0004] In general, however, capturing a colored object in black and white, just like converting a color image to grayscale, involves loss of quality. For example, in such a process, a suitable pair of blue and red tones is mapped to the same gray value, making these colors indistinguishable after conversion to grayscale.

[0005] For this reason, color information of a photographed object cannot be recovered from a black and white image without additional information. Therefore, the success of reusing a black and white image depends crucially on whether relevant information (e.g., contours) is recognizable in the black and white image.

[0006] To highlight or better distinguish relevant areas or contours of the captured object in a black and white photograph, the object can be appropriately illuminated with colored light sources during the photographic process. For example, red areas of an object appear almost white when illuminated with a red light source in a black and white photograph, while blue areas appear very dark.

[0007] Besides suitable lighting, numerous other parameters influence image quality when taking a photograph, such as the camera's field of view, focus, and exposure time. The success of subsequent processing steps therefore depends crucially on the chosen parameter values. Often, a suitable set of parameter values ​​must be individually determined for each application, a process also known as camera training. Various solutions exist in the art to simplify this process, particularly when the training must be performed (partially) by a user.

[0008] German patent DE 10 2018 119 691 A1 discloses a camera system for capturing codes. According to the invention, the positions of contour points of the codes are determined and analyzed in relation to the edges of a search zone of the camera system. If, for example, a large proportion of the contour points fall within a boundary area of ​​the camera system's search zone, this indicates that the search zone should be selected differently. This can then be done manually or automatically. In this way, a higher level of readability can be ensured.

[0009] German patent DE 10 2019 113 647 A1 describes a camera system for capturing objects and / or object properties, which, in addition to the object being captured, also records further image features and information about false detections. For example, if it is determined that detections of images with an increased average gray value are frequently faulty, corresponding feedback can be provided to the user, who can then correct the settings.

[0010] In German patent application DE 10 2015 216 446 A1, a method for checking the completeness of a container with multiple objects using a 3D camera is improved by supporting the user in selecting appropriate parameters. These parameters are primarily threshold values ​​for height, which determine whether a container space is classified as occupied or unoccupied. A key aspect of the method is to provide the user with two displays simultaneously: the first shows a two-dimensional view of the container, and the second displays the recorded height values ​​with adjustable thresholds. This gives the user an improved overview of the impact of the set thresholds.

[0011] German patent application DE 10 2019 103 397 A1 discloses a method for operating a camera for capturing codes. A modular operating mode of the camera enables the learning phase, which is necessary for reliable recognition and usually has to be carried out manually, to be performed by the camera itself.

[0012] The publication EVANS, AN, Liu, XU: A morphological gradient approach to color edge detection. In: IEEE Transactions on Image Processing, 2006, Vol. 15, No. 6, pp. 1454-1463. doi: 10.1109 / TIP.2005.864164 describes an algorithm for edge detection in color images. It is also described that with only one channel, the presented algorithm reduces to a morphological gradient. However, morphological gradients cannot simply be transferred to color images because the RGB values ​​are vectors and there is no accepted ranking, meaning the values ​​cannot be ordered. This publication addresses how an order can be imposed on RGB vectors in order to process color images.

[0013] The publication WANG, H., et al.: A context-driven approach to image-based crack detection. Machine Vision and Applications, 2016, Vol. 27, No. 7, pp. 1103-1114. Doi:10.1007 / s00138-016-0779-1 describes an approach to crack detection in images, specifically developed for the automated inspection of aircraft. This approach utilizes additional contextual knowledge, such as geometric / structural or physical context. It is also described that polarized light and a polarizing filter should be used for illumination to prevent reflections on uncoated metal surfaces.

[0014] The publication KO, S.-S., Lin, Y.-C.: Optical Inspection system with tunable exposure unit for micro-crack detection in solar wafers. In: Optik, 2013, 124, Vol. 19, pp. 4030-4035. describes a system for detecting microcracks in solar wafers, which are the main cause of breakage and electrical failure. The system uses a near-infrared light source and a CCD camera whose exposure time is automatically adjusted by a tunable exposure unit to optimize image contrast. This enables reliable detection of microcracks in wafers of varying thicknesses in less than one second.

[0015] US Patent 2015 / 0269751A1 describes a method for the automatic digitization of notes. Visual information is extracted from notes by separating the color channels of an input image and performing edge detection for each color channel. Furthermore, contiguous polygons are identified for each color channel and then grouped so that a group of polygons represents a note. The patent also describes the use of a grayscale image, where brightness values ​​are divided into separate groups.

[0016] Another relevant state of the art is HAASDONK, B.: Digital Image Processing, Unit 1, Introduction. April 1, 2006. URL: https: / / lmb.informatik.unifreiburg.de / oeoole / haasdonk / DBV FHO / DBV FHO 2006 E01.pdf

[0017] The publication DOUGHERTY, J.: Using filters in color inspection, June 1, 2007. URL: https: / / www.vision-systems.com / camerasaccessories / article / 16738837 / usinq-filters-in-color-inspection describes how good results can be achieved in industrial inspection applications when using monochrome cameras with filters. The publication explains that filters can be used with monochrome cameras to reduce reflections, for example. The article further states that filters should be used to achieve optimal results with correct exposure and camera systems.

[0018] Through the publications FERREIRA, T., Rasband, W.: ImageJ User Guide. Version ImageJ 1.44, February 9, 2011. BROEKE, J., PÉREZ, JMM, Pascau, J.: Image processing with ImageJ. Second edition. Packt Publishing Ltd., Birmingham-Mumbai. 2015. pp. 1-231

[0019] The image processing software ImageJ is described, which offers a wide range of functions for processing images.

[0020] Based on the prior art, the invention aims to improve the selection of photographic images. In particular, the method is intended to support the selection of a suitable RGB-W lighting setting as well as parameters and parameter values ​​of image processing algorithms.

[0021] This problem is solved by a computer-implemented method according to claim 1, a computer-readable storage medium according to claim 8, an image processing device according to claim 9 and a recording system according to claim 10.

[0022] In particular, the task is solved by a computer-implemented, especially user-guided, method for selecting a photographic image for use in production control, wherein the method comprises the following steps: a) Receiving from a multitude of input images specifying an object to be captured in a multitude of lighting settings, wherein the lighting settings are specified by different configurations of an RGB-W lighting unit and include a white image; b) Selecting at least one candidate from the initial recordings based on a first selection criterion; c) Defining a template image section in the white image; d) Creating at least one target image section in the at least one candidate image that corresponds to the template image section; e) Applying at least one image processing algorithm, in particular contour recognition, blob analysis, region analysis, code analysis or model matching, with a first set of parameters and / or parameter values ​​to the at least one target image section to generate processed candidate images; f) Selecting at least one pre-selection recording from the set of processed candidate recordings based on a second selection criterion; g) Check whether a pre-selection recording exists that meets a final screening criterion and, if so, select such a pre-selection recording as the final recording; otherwise h) Repeating steps e) to g), wherein the at least one preselection recording is used as the at least one candidate recording and in step e) a second set of parameters and / or parameter values, in particular specified by a user, is used. wherein at least one of the selection criteria depends on an evaluation result of at least one image analysis algorithm, namely contour recognition, wherein a collection of reference images is used for the evaluation of the at least one selection criterion, wherein contours are extracted from object edges of the reference images during contour recognition and stored as reference contours, which are searched for in the multitude of input images, wherein the method further comprises calculating a value that indicates a match with a reference contour and wherein, if the value exceeds a threshold, the area under consideration is regarded as a match, wherein an image is rated as good if a predefined number of matches are found and wherein the image is selected as at least one candidate image if the image is rated as good.

[0023] A key principle of the process is that at the beginning of the process, a large number of input images are available in different lighting settings, and not just one, but potentially several of these form the basis for further processing steps. This prevents an early commitment to a lighting setting that turns out to be suboptimal for later processing steps.

[0024] The input images include a white image, which in this context can be understood as an image of the object to be captured, in which the RGB-W illumination unit illuminates the object with white light. The white image serves to define a sample image section, whereby, in the context of this invention, a sample image section can be understood as the combination of any number of areas of a single image. In particular, a sample image section can comprise several non-contiguous areas of the image, for example, non-overlapping, elliptical areas, e.g., three.

[0025] Another aspect of the method is the selection of images based on a first and second selection criterion, as well as a final test criterion. The combination of these criteria ensures that a single final image is ultimately selected from the multitude of input images. According to the invention, several criteria are provided for this purpose, which do not necessarily have to be identical. This allows different aspects for the selection of images to be considered for different processing steps of the method.

[0026] A pre-selection of input images, referred to as the set of candidate images and selected according to the first selection criterion, is then used to apply at least one image processing algorithm for further optimization. Specifically, this algorithm detects contours within an image. For this purpose, image sections within the candidate images, called target image sections, are defined beforehand. This ensures that the algorithmic image processing is directed at the relevant areas of the image. Otherwise, there is a risk that features outside the relevant areas, particularly at the image edges, could influence the image processing.

[0027] Another aspect of the invention is to repeat the described image processing if the results of the first execution are deemed insufficient, for example, because relevant contours have not yet been recognized in the algorithmic image processing. For this purpose, steps e) to g) can be repeated until a pre-selection image exists that meets the final test criterion.

[0028] In a preferred embodiment, a natural number i_max can be defined such that the number of repetitions of steps e) to g) according to step h) does not exceed i_max. For example, choosing i_max = 3 ensures that the process terminates after a maximum of three repetitions. Since algorithmic image processing is inherently resource-intensive, a high number of repetitions may be undesirable in practice. Early termination after a preset number of repetitions allows, for example, adjustment of the final test criterion and thus restarting the process.

[0029] Another embodiment may provide that if, in an implementation of step g), no pre-selection image exists that meets the final test criterion, the white image is selected as the final image. Depending on the application, it may be desirable for the method to always return an image, for which the white image, as a neutrally illuminated image of the object, is suitable.

[0030] In one embodiment of the method, a user can exist, so that - at least one of the selection criteria or the final test criterion is specified by the user in such a way that at least one recording is selected from the set of available recordings, and / or - the template image section is defined by the user, in particular by means of geometric selection tools and / or a brush tool.

[0031] This embodiment is particularly suitable when no suitable formal description exists for a selection criterion, the final test criterion, or the selection of the template image section that a computer can evaluate, but a user can actively make this selection. To define a selection criterion or final test criterion, the method preferably provides the user with an overview of all available images in lower resolution, from which they can select or deselect one or more images. Defining the template image section is user-friendly, using geometric selection tools, particularly in the form of rectangles and ellipses, and / or a brush tool.

[0032] According to another embodiment, at least one of the selection criteria or the final test criterion can consist of maximizing at least one characteristic of an image and / or exceeding a predefined threshold. These characteristics can include, in particular, the intensity difference between a brightest and a darkest point, dynamic range, contrast, and / or image sharpness. This can be used as an indicator that an image meets minimum requirements for further use. Furthermore, it is a criterion that is easy to verify algorithmically.

[0033] At least one of the selection criteria or the final evaluation criterion depends on the result of an evaluation from at least one image analysis algorithm. The image analysis algorithm in question is a contour recognition algorithm.

[0034] In contour recognition, contours are extracted from object edges, stored as reference contours, and then searched for in subsequent images. To assess how similar a section of the current image is to the reference contours, a value is calculated that can be interpreted as a percentage match with the reference. If the value exceeds a predefined threshold, the area in question is considered a match. An image is rated as good if a predefined number of matches are found.

[0035] In blob analysis, an image is segmented based on grayscale thresholds, meaning individual objects are separated from the background. Properties such as size are then determined for each individual object. If all properties of an object fall within predefined limits, the object is rated as "good." An image is rated as "good" if a specified number of good objects are found. Optionally, the total area of ​​all objects rated as good can also be considered as a decision criterion for the image.

[0036] Code analysis can be used to search for and read one- or two-dimensional codes of various types (e.g. EAN13, QR code) in an image.

[0037] Model matching allows an entire image section to be defined as a reference and then searched for in subsequent images. The similarity to the reference can be quantified via correlation.

[0038] In a preferred embodiment, the method can use a collection of reference images to evaluate at least one of the selection criteria or the final test criterion. Accordingly, an image is selected if at least one reference image exists such that the distance between the image and the reference image falls below a predefined threshold. The distance is measured using a suitable metric, such as the Earth Mover's Distance (EMD) or the Euclidean distance of the grayscale histograms.

[0039] A preferred embodiment is one in which the method in step d) can generate the at least one target image section in the at least one candidate image as follows: The at least one polygon corresponding to the template image section is transformed into at least one second polygon. Here, the at least one polygon corresponding to the template image section is defined in a first coordinate system of the white image, and the at least one second polygon is defined in a second coordinate system of the at least one candidate image. An advantage of this embodiment is that the at least one candidate image does not necessarily have to have the same dimensions as the white image.

[0040] The problem is further solved in particular by a computer-readable storage medium which contains instructions that cause at least one processor to implement a procedure as described above when the instructions are executed by the at least one processor.

[0041] Furthermore, the task is solved in particular by an image processing device comprising a computer-readable storage medium as described above and a computing unit. The computing unit is configured to execute instructions stored on the computer-readable storage medium.

[0042] In particular, the task is further solved by a recording system for generating a photographic image for use in production control, which has the following components: - an image processing device as described above, - a black and white camera and - an RGB-W lighting unit.

[0043] The black and white camera is designed to capture images of an object under a variety of lighting conditions, where the lighting conditions are defined by different configurations of the RGB-W lighting unit and include a white image. Furthermore, the black and white camera is designed to provide these images to the image processing device.

[0044] According to one embodiment, the components of a recording system as described above, i.e., the image processing device, the SW camera and the RGB-W lighting unit, can be installed in a single device.

[0045] In an alternative embodiment, only the image processing unit and the SW camera are built into a single device, while the RGB-W lighting unit is connected to the device as an external device.

[0046] In a further embodiment, the recording system can also include a transmitting unit and a receiving unit. The transmitting unit can be configured to transmit images to the receiving unit. The receiving unit can be configured to provide the received images to the image processing device. Thus, the image processing device does not need to be directly connected to the monochrome camera, but can be located elsewhere, for example, in a data center.

[0047] A preferred embodiment may provide that a recording system as described above further comprises a unit for interacting with a user. The interaction unit may be configured to operate via input and output devices. - to display a set of images, preferably in reduced image resolution on at least one overview page, and to capture the user's selection of at least one image from this set and / or - to capture a template image section, which the user defines in particular using geometric selection tools and / or a brush tool, and / or - To accept user input for parameters and / or parameter values.

[0048] A recording system according to this embodiment can, on the one hand, provide a user with an improved overview of the progress of the computer-implemented process, and on the other hand, accept user input which the computer-implemented process uses directly or indirectly to select a recording.

[0049] Displaying images, especially initial images, in reduced resolution on an overview page can give the user an overview of the available images. By selecting individual images, the user can then easily determine, for example, the first selection criterion.

[0050] Defining a sample image area can also be conveniently done by the user in a suitable graphical user interface. Preferably, geometric selection tools and / or a brush tool are provided for this purpose, allowing the selection of the sample image area within the white image display. A simple geometric selection tool can, for example, create a selection area in the form of a triangle, a square, or an ellipse. A polygon tool can also be provided for creating polygonal selection areas, allowing the segments of a polygon to be defined step by step. Furthermore, a brush tool can be provided that creates a selection from the total set of all pixels selected with the brush tool.This is particularly advantageous if the area to be marked cannot be easily selected using a geometric selection tool.

[0051] Further embodiments of the invention are set out in the dependent claims.

[0052] The invention is described below with reference to exemplary embodiments, which are explained in more detail with reference to the figures. These show: Fig. 1 an exemplary recording system; Fig. 2. A flowchart of a process; Fig. 3. A graphical user interface for determining pre-selection shots; Fig. 4. A graphical user interface for defining a template image section.

[0053] In the following description, the same reference numbers are used for identical and equivalent parts.

[0054] Fig. Figure 1 shows a recording system 70 according to the invention. In this embodiment, the recording system 70 serves to produce a photographic image of an object 30 for later production control.

[0055] In the illustration, object 30 is a cuboid component with five screws attached one below the other, the correct positioning of which is to be checked in production control.

[0056] For the purpose described, a black and white camera 10 is directed at the object 30, the surface of which is illuminated by an RGB-W lighting unit 20. To avoid unwanted reflections in the photographic image, the black and white camera 10 is additionally equipped with three polarizing filters 11.

[0057] The RGB-W lighting unit 20 allows for a variety of different lighting settings by switching the individual colored light sources on or off. Furthermore, the different settings of the polarization filters 11 result in further variations. In this way, the SW camera 10 can create a variety of input images E1,...,En as well as a white image E0, showing the object 30 under different lighting conditions.

[0058] Further components of the recording system 70 are an image processing device comprising a computer-readable storage medium 71 containing instructions for carrying out a method according to the invention, and a computing unit 72 which executes the method according to the instructions stored on the storage medium 71.

[0059] It should be noted that the representation in Fig. Figure 1 schematically shows the components of a recording system according to the invention. Several components can be combined into a single device that fulfills the functions of all components. In particular, according to the embodiments described above, the SW camera 10, the storage medium 71, and the processing unit 72 can be installed in one device, while the RGB-W illumination unit 20 is optionally connected internally to said device or externally to it.

[0060] In Fig. Figure 2 shows a flowchart illustrating the process of a method 50 according to the invention.

[0061] In step 51, input images E1,...,En and a white screen E0 are received.

[0062] In step 52, a subset of the input images E1,...,En is determined based on a first selection criterion; these are referred to as candidate images K1,...,Km. The first selection criterion could, for example, consist of algorithmically checking whether the input images E1,...,En have a sufficiently high contrast.

[0063] In step 53, a template image section A0 is defined within the white image E0. This can be done automatically using area recognition algorithms or manually with suitable marking tools. In this way, for example, the object 30 to be captured can be separated from its background, or a relevant area of ​​object 30 can be selected.

[0064] The template image section A0 serves as the basis for the following step 54. Here, target image sections A1,...,Am are generated in the candidate images K1,...,Km, which correspond to the template image section A0. In a simple case, the candidate images K1,...,Km have the same dimensions as the white image E0, so the formal description of the template image section A0, for example as a polygon, can be directly adopted to define the target image sections A1,...,Am. However, it can also happen that at least one candidate image K1,...,Km has a different dimension than the white image E0. In such a case, a suitable transformation of the template image section A0, which is defined in a first coordinate system of the white image E0, is required into the at least one target image section A1,...,Am, which is defined in at least a second coordinate system of the at least one differently dimensioned candidate image K1,...,Km is defined.

[0065] In step 55, image processing algorithms are applied to the target image sections A1,...,Am. For example, a first algorithm can detect edges using the Sobel operator. A second algorithm can enhance the detected edges in the image with a line thickness that is a parameter value of the second algorithm. In the first execution of this step, the line thickness can be set to a preset default value, such as 3pt. Finally, the manipulated target image sections are combined with their respective complementary, unmodified target image sections to form processed candidate images K'1,...,K'm. The image processing algorithms result in enhanced contours within the target image sections.

[0066] From the set of processed candidate recordings K'1,...,K'm, pre-selection recordings V1,...,Vk are selected in step 56 based on a second selection criterion.

[0067] In step 57, the pre-selection images V1,...,Vk are evaluated based on a final selection criterion. In principle, the same or similar aspects as the first and second selection criteria are suitable for the final selection criterion, for example, sufficient contour definition. If at least one pre-selection image V1,...,Vk meets the final test criterion, the procedure selects this image as the final image F.

[0068] Otherwise, steps 55 to 57 are executed again. However, the most recently determined preselection images V1,...,Vk are used as candidate images K1,...,Km. Furthermore, a second set of parameters and / or parameter values ​​is used. If, for example, no preselection image K1,...,Km with sufficient contour definition according to the final inspection criterion exists in the first run, an obvious adjustment of the parameter values ​​would be to increase the intensity of the edge enhancement of the corresponding algorithm from step 55. This can cause the corresponding image to meet the final inspection criterion in the second execution of step 57 and be selected as the final image F.

[0069] In Fig. Figure 3 shows an embodiment of a graphical user interface 80 for selecting candidate images K1,...,Km from the input images E1,...,En. The user interface 80 displays the four input images E1, E2, E3, E4 on an overview page, each in a resolution that occupies slightly less than a quarter of the screen area. The user navigates to the previous or subsequent overview page, where the remaining input images are displayed, using the buttons 83. In the user interface 80, each input image E1, E2, E3, E4 contains a selection symbol 81, 82 at the upper left edge of the screen. The user can use this symbol to add or remove the respective input image from the set of candidate images K1,...,Km.

[0070] In the illustrated embodiment, the input images E1, E2, E3, and E4 show object 30, with the contours of object 30 being of varying intensity. The contours in E1 are only faintly defined, while those in E2 and E3 are moderately defined. Input image E4, however, shows such strong contours that a detailed representation in intermediate areas is no longer possible. Of the in Fig. In this case, the user has selected input recordings E2 and E3 from the three input recordings E1, E2, E3, and E4 shown, while E1 and E4 have been deselected. The set of candidate recordings therefore includes at least E2 and E3, but not E1 and E4. The selected input recordings E2 and E3 are each marked by a filled selection symbol 82, while the remaining input recordings E1 and E4 are each marked by an unfilled selection symbol 81.

[0071] Fig. Figure 4 shows an embodiment of a graphical user interface 90 for marking the template image section A0 in the white image E0. At the upper right edge of the user interface 90 there is a toolbar 91 which provides the user with various selection tools: a triangle tool 92, a rectangle tool 93, an ellipse tool 94, a polygon tool 95 and a brush tool 96.

[0072] Furthermore, Fig. Figure 4 shows that the user uses the polygon tool 95 to define an edge that is to enclose the upper screw head of object 30 (where the edge is shown in the representation of Fig. 4 is not yet closed).

[0073] It should be noted that all the parts described above, individually—even without additional features described in the respective context, even if these have not been explicitly identified as optional features in the respective context, e.g., by using: in particular, preferably, for example, e.g., parentheses, etc.—and in combination or any sub-combination, are to be considered independent embodiments or further developments of the invention as defined in particular in the introduction and the claims. Deviations from this are possible. Specifically, it should be noted that the word "in particular" or parentheses do not denote features that are mandatory in the respective context. Reference symbol list 10 SW camera 11 Polarization filters 20 RGB-W lighting unit 30 objects 50 procedures 51 Receiving incoming recordings 52 Selecting Candidate Recordings 53 Defining a template image section 54 Defining target image sections 55 Applying image processing algorithms 56 Selecting pre-selection shots 57 Test step 70 recording system 71 Storage medium 72 computing units 80 User interface for selecting candidate recordings 81 Selection symbol (not filled in) 82 Selection symbol (filled in) 83 Page Change Buttons 90 User interface for marking the template image section 91 Toolbar 92 Triangle tool 93 Rectangle Tool 94 Ellipse tool 95 Polygon tool 96 Brush tool A0 template image section A1,...,At the target image section E0 White screen E1,...,En Entrance recordings F final recording K1,...,Km candidate recordings K'1,...,K'm processed candidate recordings V1,...,Vk Preselection recordings

Claims

[1] Computer-implemented method, in particular user-guided, for selecting a photographic image (F), for using the photographic image in production control, the method comprising the following steps: a) Receiving from a multitude of input images (E1,...,En) specifying an object to be captured (30) in a multitude of lighting settings, wherein the lighting settings are specified by different configurations of an RGB-W lighting unit (20) and include a white image (E0); b) Selecting at least one candidate recording (K1,...,Km) from the initial recordings (E1,...,En) based on a first selection criterion; c) Defining a template image section (A0) within the white image (E0); d) Generating at least one target image section (A1,...,Am) in the at least one candidate image (K1,...,Km) that corresponds to the template image section (A0); e) Applying at least one image processing algorithm, in particular contour recognition, blob analysis, region analysis, code analysis or model matching, with a first set of parameters and / or parameter values ​​to the at least one target image section (A1,...,Am) to generate processed candidate images (K'1,...,K'm); f) Selecting at least one preselection recording (V1,...,Vk) from the set of processed candidate recordings (K'1,...,K'm) based on a second selection criterion; g) Check whether a pre-selection recording (V1,...,Vk) exists that meets a final screening criterion and, if so, select such a pre-selection recording (V1,...,Vk) as the final recording (F); otherwise h) Repeating steps e) to g), wherein the at least one preselection recording (V1,...,Vk) is used as the at least one candidate recording (K1,...,Km) and in step e) a second set of parameters and / or parameter values, in particular specified by a user, is used;wherein at least one of the selection criteria depends on an evaluation result of at least one image analysis algorithm, namely contour recognition, wherein a collection of reference images is used for the evaluation of the at least one selection criterion, wherein contours are extracted from object edges of the reference images during contour recognition and stored as reference contours, which are searched for in the multitude of input images, wherein the method further comprises calculating a value that indicates a match with a reference contour and wherein, if the value exceeds a threshold, the area under consideration is regarded as a match, wherein an image is rated as good if a predefined number of matches are found and wherein the image is selected as at least one candidate image if the image is rated as good. [2] Method according to claim 1, wherein the number of repetitions of steps e) to g) according to step h) does not exceed a predetermined value. [3] Method according to one of the preceding claims, wherein in the case that in an embodiment of step g) no preselection image (V1,...,Vk) exists which satisfies the final test criterion, the white image (E0) is selected as the final image (F). [4] Method according to any one of the preceding claims, wherein - at least one of the selection criteria or the final test criterion is specified by a user in such a way that at least one recording is selected from the set of available recordings, and / or - the template image area (A0) is defined by a user, in particular by means of geometric selection tools and / or a brush tool. [5] Method according to any of the preceding claims, wherein at least one of the selection criteria or the final test criterion consists in maximizing at least one characteristic of a recording, in particular an intensity difference between a brightest and a darkest point, dynamic range, contrast, and / or image sharpness, and / or exceeding a predetermined threshold. [6] Method according to one of the preceding claims, wherein a recording is selected according to at least one of the selection criteria or the final test criterion if at least one reference image exists in the collection of reference images such that a distance value between the recording and the reference image in a metric falls below a predetermined threshold. [7] Method according to one of the preceding claims, wherein in step d) the at least one target image section (A1,...,Am) in the at least one candidate image (K1,...,Km) is generated by transforming the at least one polygon, which corresponds to the template image section (A0) and is defined in a first coordinate system of the white image (E0), into at least one second polygon, which is defined in a second coordinate system of the at least one candidate image (K1,...,Km). [8] Computer-readable storage medium containing instructions that cause at least one processor to implement the method according to any of the preceding claims when the instructions are executed by the at least one processor. [9] Image processing device comprising the following: - a computer-readable storage medium (71) according to claim 8 and - a computing unit (72) designed to execute stored instructions from the computer-readable storage medium. [10] Recording system (70) for producing a photographic image (F) for use in production control, comprising the following: - an image processing device according to claim 9, - a black and white camera (10) and - an RGB-W illumination unit (20), wherein the SW camera (10) is configured to produce images of an object (30) to be detected in a variety of illumination settings, wherein the illumination settings are specified by different configurations of the RGB-W illumination unit (20) and include a white image (E0), and is further configured to provide the images to the image processing device. [11] System according to claim 10, wherein the image processing device, the SW camera (10) and the RGB-W illumination unit (20) are installed in a single device. [12] System according to claim 10, wherein the image processing device and the SW camera (10) are installed in a single device and the RGB-W illumination unit (20) is an external device connected to the device. [13] System according to any one of claims 10 to 12, further comprising: - a transmitter unit and - a receiving unit, wherein - the transmitting unit is designed to transmit the created recordings to the receiving unit and - the receiving unit is designed to provide the received recordings to the processing unit. [14] System according to any one of claims 10 to 13, further comprising an interaction unit configured to operate via input and output devices - to display a set of images (E1,...,En), preferably in reduced image resolution on at least one overview page, and to capture a user's selection of at least one image from this set and / or - to capture a template image section (A0) that a user defines, in particular, using geometric selection tools and / or a brush tool, and / or - To accept user input for parameters and / or parameter values.

Images