Method and device for acquiring anonymous images in an industrial plant - Patents.com

Abstract

To provide a method and a device for acquiring an anonymous image for an industrial plant, which can remove a personal feature in the image by making the personal feature in the image unclear.SOLUTION: A method for acquiring an anonymous image for an industrial plant includes the steps of: acquiring an image 1000 including an opto-electronically readable code 1200 and a personal feature 1100; identifying the personal feature 1100 in an image 1000; and making the personal feature 1100 in the image 1000 unclear.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present invention relates to a method and apparatus for acquiring anonymous images in industrial plants, such as goods storage plants, goods transport plants and / or goods processing plants. [Background technology]

[0002] Image acquisition methods and devices are often used in industrial plants to monitor and control product or process streams.

[0003] However, in this process, personal characteristics are often also captured, which cannot meet the requirements of data protection. Moreover, personal characteristics require memory space and can make the subsequent processing steps of the captured images more complicated, for example when larger image files have to be processed. Summary of the Invention [Problem to be solved by the invention]

[0004] It is therefore an object of the present disclosure to provide an improved method and an improved device for obtaining anonymous images for industrial plants, in particular such that subsequent storage and processing of the images is facilitated. [Means for solving the problem]

[0005] This object is firstly met by a method for acquiring anonymous images in an industrial plant according to independent claim 1.

[0006] In a first step of the method, images are acquired using an image acquisition unit, for example a camera or an image sensor, in particular a line camera or a matrix camera. The image acquisition unit is in particular a so-called presentation camera. In the method, images are acquired regularly, continuously and / or in an event-related manner (for example by activating light barriers and / or actuating elements in the industrial plant).

[0007] In this method, the image comprises an optoelectronically readable code, such as a barcode, a QR code or a Semacode, which code is attached or affixed to an item (such as a product or a package) stored in, for example, an industrial plant.

[0008] The image further comprises at least one personal feature, in particular a human hand holding an object in front of a camera for the purpose of acquiring an image. For example, the personal feature may comprise a part of a fingerprint. In the present method, the personal features may also comprise clothing (in particular a sleeve or sleeves) and / or jewellery (such as a ring or a watch). Personal features are features that allow conclusions to be drawn about the person, in particular features that allow the person to be identified, more particularly features that allow the person to be unambiguously identified.

[0009] In a next step of the method, as will be explained in more detail below, this personal feature is identified in the image, for example using a processing unit.

[0010] The final step of the method is to blur the previously identified human features in the image, where "blur" means specifically blurred to the human eye, but can also include blurred to the processing unit, and in particular the blurring is irreversible, i.e. it cannot be undone by either the human or the processing unit.

[0011] According to the present anonymous image acquisition method, images acquired in industrial plants can be processed without taking into account data protection specifications, since personal features have been blurred and it is no longer possible to draw conclusions about persons at a later time, e.g. during further and / or subsequent processing for other purposes.

[0012] The images thus acquired can also be made available in particular to external cooperation partners and / or cooperation partners that are not bound by data protection specifications.

[0013] It is particularly advantageous that, since the features have been blurred, the image contains less information and therefore requires less memory space: in particular the original image is deleted and only the image in which the features have been removed or blurred is stored, thus requiring less power and memory for further processing of the image.

[0014] The image with blurred human features is preferably used to influence the control of an industrial plant, whereby a certain action of the industrial plant, for example further processing of a workpiece, can be triggered, for example by means of a readable code.

[0015] The method can be further developed such that identifying a human feature in an image comprises segmenting the human feature in the image, where "segmenting" means dividing an image into related parts and then classifying the parts or segments into individual classes, in particular predefined classes. In this embodiment, an algorithm for feature extraction can be used and the image can be pre-processed. Segmentation allows a statement of an identified area, in particular whether it is a human feature in this embodiment.

[0016] Segmentation can use known methods, such as pixel-oriented, edge-oriented, and / or region-oriented methods. Additionally or alternatively, model-based and / or texture-based methods can be used for segmentation.

[0017] This further development results in a particularly efficient solution, in particular in that it is achieved that only those areas of the acquired image which relate to human features are processed, i.e. blurred.

[0018] The method can be further developed to use neural networks to identify and / or segment human features in the image.

[0019] A neural network is a computer-based artificial intelligence, which may also be referred to as a neural network and / or an artificial neural network. Identification or segmentation may be performed using neural networks that have been trained with deep learning algorithms, in particular to be able to segment even newly acquired images that were not processed by the neural network during training.

[0020] The identification or segmentation can in particular be carried out by means of processing units specially provided for this purpose, such as microprocessing units with a high degree of parallelism or programmable logic components.

[0021] This further development leads to particularly powerful solutions: in particular, neural networks provide an especially reliable way of identifying or segmenting personal features.

[0022] The method may be further developed such that blurring human features in the image comprises pixelating human features in the image.

[0023] By "pixellization" we mean replacing the original fine pixel resolution of the human features with a much coarser resolution, so that the individual partial features of the human features are no longer visible, but only the coarse structure of unrecognizable pixels is used, which can be done by suitable pixellization algorithms.

[0024] The method may be further developed such that blurring human features in the image comprises removing human features in the image.

[0025] By "remove" here we mean to remove pixels relating to human features and / or to replace these pixels with pixels of a single color, for example white or black.

[0026] The method may be further developed to further comprise reading said optoelectronically readable code.

[0027] The reading of the optoelectronically readable code is performed using a suitable reading algorithm. For example, information present or encoded in the optoelectronically readable code can be read and / or evaluated. The result of the read code, for example a number or serial number (such as a tracking number) or other content of the code, can then be stored together with the location and / or time of reading.

[0028] In this embodiment, the optoelectronically readable code may specifically include information and / or instructions regarding the object to which the optoelectronically readable code can be attached. The optoelectronically readable code may specifically include future and / or past storage, delivery and / or output locations. Alternatively or additionally, the optoelectronically readable code may include a reference to the contents and / or attributes of the object.

[0029] The information read out from the optoelectronically readable code can then lead to actions or no actions in the industrial plant. This information can in particular be taken into account during future actions in the industrial plant, such as for example further processing, further transport or further storage of the object. For example, the next destination of the object, in particular within the industrial plant, can be determined based on the read out information.

[0030] The method may be further developed to further include storing the image including the blurred human features.

[0031] The storage may in particular consist in storing the code that was read, together with the location and time of reading.

[0032] This further development allows recording the passing of an object through a predefined waypoint or the current storage location of the object, thereby allowing goods processed or produced by the industrial plant to be tracked within the industrial plant.

[0033] The method may be further developed such that storing the image comprises transmitting it to a remote storage unit.

[0034] In this embodiment, the remote storage unit is in particular a storage unit located far away from the image acquisition unit and the industrial plant. The remote storage unit may in particular be a storage unit in which the image with blurred human features is stored for verification. In this embodiment, the remote storage unit may in particular be in a different country than the image acquisition unit or the industrial plant.

[0035] The method may be further developed to further comprise cropping the image to a sub-area which includes said optoelectronically readable code.

[0036] As used herein, "cropping" refers to removing image content that does not contain an optoelectronic readable code, and may also be referred to as "masking." In this embodiment, cropping may also be performed as cutting out portions of human features that have been pre- or post-blurred.

[0037] Cropping can be done in a routine manner, ie always with a constant crop factor, and / or dynamically, for example based on identification of personal characteristics and / or optoelectronic readable codes.

[0038] If cropping is performed before identification or segmentation and / or before blurring, less processing power is required for identification or segmentation and / or blurring because less image content needs to be processed.

[0039] In this embodiment, cropping can occur both before and after transmission to the remote storage unit, or cropping can occur without transmitting the image to the remote storage unit, thereby reducing memory space in both local and remote storage units.

[0040] This further development requires particularly less memory space and / or processing power.

[0041] The method may be further developed to include encrypting the captured image, transmitting the captured image to a remote processing unit, and decrypting the captured image at the remote processing unit.

[0042] In this embodiment, encryption can be performed by any common encryption algorithm. For example, the acquired images can be sent to a remote processing unit by SFTP. Alternatively or additionally, the acquired images can be encrypted and decrypted using, for example, symmetric or asymmetric encryption protocols.

[0043] Encryption may in particular be performed immediately after image capture, in particular prior to identification and / or blurring of personal features, and such raw images may be transmitted to a remote processing unit where they can be decrypted.

[0044] The remote processing unit may be a processing unit directly connected to the remote storage unit or a processing unit distinct from or remote from the remote storage unit.

[0045] This further development prevents someone from stealing an acquired image containing personal characteristics and being able to draw conclusions about a person in an unauthorized manner.

[0046] The method may be further developed such that identification and / or segmentation of human features in an image is performed in the remote processing unit.

[0047] In particular, the remote processing unit may be specifically configured for this purpose to perform the identification or segmentation and may also be adapted to use suitable neural networks specifically trained for this application. For this purpose, the remote processing unit may further specifically be adapted to meet specific data protection specifications, for example by the location and / or authentication of the remote processing unit.

[0048] According to this embodiment, the obscuring can be performed specifically in accordance with data protection requirements and is not performed within the industrial plant.

[0049] The initially mentioned object is likewise achieved by a device for acquiring anonymous images in an industrial plant according to independent claim 12.

[0050] For further embodiments and their advantages, reference is made to the above description of the method for obtaining anonymous images in an industrial plant.

[0051] The device for acquiring anonymous images in an industrial plant is in particular adapted to partially or completely carry out the embodiments of the method for acquiring anonymous images in an industrial plant described above.

[0052] BRIEF DESCRIPTION OF THE DRAWINGS An embodiment of a method and apparatus for obtaining anonymous images in an industrial plant will now be described in detail in relation to the following drawings. [Brief description of the drawings]

[0053] [Figure 1] 1 is a schematic diagram of an embodiment of an apparatus for acquiring anonymous images in an industrial plant; [Diagram 2] 1 is a flow chart of an embodiment of a method for acquiring anonymous images in an industrial plant. [Diagram 3] (a)-(d) Individual steps of the application of the method for obtaining anonymous images in an industrial plant. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0054] In the present embodiments, the same reference numbers indicate the same or similar features.

[0055] FIG. 1 shows a schematic diagram of an embodiment of an apparatus 1 for acquiring anonymous images in an industrial plant.

[0056] In this embodiment, the device 1 first comprises a local unit 10 and a unit 20 remote from the local unit 10. The remote unit 20 is shown as a cloud service capable of implementing parts of the device 1 or steps of the method 100, as will be described in particular and by way of example in connection with FIG.

[0057] The local unit 10 comprises a local storage unit 11, a local processing unit 12 and an image acquisition unit 13. The local unit 10 also comprises a wireless and / or wired communication unit, which is not shown for clarity, where "local" means located in or part of the industrial plant.

[0058] The remote unit 20 comprises a remote storage unit 21 and a remote processing unit 22. The remote unit 20 also comprises a wireless and / or wired communication unit, which is not shown for the sake of clarity, where "remote" means spatially separate from the local unit 10 and specifically not located within the industrial plant.

[0059] The remote unit is specifically one or more servers set up in a remote location to constitute a cloud service. In this embodiment, for the sake of clarity, the remote storage unit 21 and the remote processing unit 22 are shown as belonging to the same cloud service.

[0060] Alternatively, the storage unit 21 is provided by a first cloud service and the processing unit is provided by a different second cloud service, and the first cloud service and the second cloud service are connected via a wireless and / or wired communication unit.

[0061] More specifically, the remote unit 20 does not include at least an image acquisition unit, or the local image acquisition unit 13 may be the only local unit located in the industrial plant besides the communication unit.

[0062] The device 1 is in particular adapted to carry out the method described below in connection with FIG.

[0063] FIG. 2 shows a flow chart of an embodiment of a method 100 for acquiring anonymous images in an industrial plant.

[0064] In a first step 101 of the method 100, an image containing an optoelectronically readable code and a personal characteristic is acquired.

[0065] In a following step 102, the acquired image is encrypted.

[0066] The next step 103 is to transmit the encrypted image to a remote processing unit.

[0067] In a following step 104, the encrypted image is decrypted in the remote processing unit.

[0068] In the next step 105, human features in the decoded image are identified by a remote processing unit through segmentation using a neural network.

[0069] In a following step 106, human features in the image are blurred by pixelating the human features in the image and / or by removing the human features from the image.

[0070] In the next step 107, the image is cropped to the sub-region containing the optoelectronically readable code.

[0071] In the next step 108, the optoelectronically readable code is read.

[0072] In a subsequent step 109, the image including the blurred human features is transmitted to a remote storage unit where it is stored in a final step 110. The method 100 can then begin again at step 101.

[0073] Figures 3(a)-(d) show the sequence of application of the method for acquiring anonymous images in an industrial plant. In this embodiment, in a first step, an image 1000 is acquired, which contains both human features 1100 and an optoelectronically readable code 1200, as shown in Figure 3(a).

[0074] In the illustrated example, the personality feature 1100 consists of both hands, with a portion of clothing (ie, a sleeve) and a ring being shown as examples of other parts of the personality feature 1100.

[0075] The optoelectronically readable code 1200, in this example configured as a QR code, is applied to an object 1300, which is shown here by way of example as a rectangular piece of paper.

[0076] The image 1000 shown in FIG. 3(a) is the image described in relation to FIG. 2 and was acquired in a first method step 101 by a human using both hands to hold an object 1300 in front of the presentation camera.

[0077] In the next step, personal features 1100 are identified in the image 1000 using a segmentation algorithm, actually implemented by a neural network, as described in connection with method step 105 of FIG.

[0078] The result of this step 105 is shown in Figure 3(b), where only the personal features 1100 are included in the image 1000. However, it is depicted this way merely for the purposes of understanding the identification or segmentation, as in reality the object 1300 containing the optoelectronic readable code 1200 is still included in the image 1000, i.e. the image shown in Figure 3(b) does not represent all the current image information, but rather only shows the result of the identification or segmentation.

[0079] In the next step, the personal feature 1100 contained in the image 1000 is blurred. This is shown in Figure 3(c) where the personal feature 1300 has been pixelated. Alternatively, the removal or deletion of the personal feature 1100 is illustrated in Figure 3(d). This blurring corresponds to method step 106 described in connection with Figure 2.

[0080] Furthermore, in the image 1000 shown in Fig. 3(d), a portion of the image has been cropped to remove other information not necessary for reading the optoelectronic code 1200 (bottom layers, portions of human features, etc.), so that the image 1000 only contains the partial area of ​​the image 1000 in which the optoelectronic readable code 1200 was found, as shown in Fig. 3(d). This image 1000 shown in Fig. 3(d) corresponds to the image generated in connection with method step 107 described in Fig. 2.

[0081] The above-described embodiments enable an efficient solution resulting in reduced memory load and reduced processing load, which at the same time contributes to complying with the requirements of data protection of human objects in industrial plants such as goods storage plants, goods transport plants and / or goods processing plants. [Explanation of symbols]

[0082] 1...device 10…Local unit 11...Local storage unit 12…Local processing unit 13...Image acquisition unit 20…Remote unit 21...Remote storage unit 22...Remote processing unit 100…method 101...Method Steps 102...Method step 103...Method step 104...Method step 105...Method Step 106...Method Step 107...Method Step 108...Method Step 109...Method Step 110...Method step 1000…Acquired images 1100…Personality traits 1200...optoelectronically readable code 1300…object

Claims

1. A method (100) for obtaining anonymous images in an industrial plant, comprising: Acquiring (101) an image (1000), said image including an optoelectronically readable code (1200) and a personality feature (1100); Identifying (105) the human features (1100) in the image (1000); and Blurring (106) the human features (1100) in the image (1000); The procedure is as follows: The method (100), further comprising the step of cropping the image (1000) to a sub-area that includes the optoelectronically readable code (1200).

2. The method (100) of claim 1 , wherein the identifying (105) comprises segmenting the human features (1100) in the image (1000).

3. The method (100) of claim 2, wherein the identification (105) and / or the segmentation is performed using a neural network.

4. The method (100) of claim 1 , wherein the blurring (106) comprises pixelating the human features (1100) in the image (1000).

5. The method (100) of claim 1 , wherein the blurring (106) comprises removing the human features (1100) in the image (1000).

6. The method (100) of claim 1, further comprising the step of reading (108) the optoelectronically readable code (1200).

7. The method (100) of claim 1, further comprising the step of saving (110) the image including the blurred human features (1100).

8. The method (100) of claim 7, wherein storing (110) the image comprises transmitting (109) to a remote storage unit (21).

9. Encryption (102) of the acquired image (1000); transmitting (103) said acquired image (1000) to a remote processing unit (22); and Decoding (104) the acquired image (1000) in the remote processing unit (22); The method (100) of claim 2, further comprising the steps of:

10. The method (100) of claim 9, wherein the identifying (105) and / or the segmenting is performed in the remote processing unit (22).

11. A device (10) for acquiring anonymous images in an industrial plant, comprising: an image acquisition unit (13) configured to acquire an image (1000), said image comprising an optoelectronically readable code (1200) and a personality feature (1100); a storage unit (11, 12) configured to store the acquired images (1000); and a processing unit (12, 22) configured to identify the human feature (1100) in the image (1000) and to blur the human feature (1100) in the image (1000); Equipped with The apparatus (10), wherein the processing unit (12, 22) is further configured to crop the image (1000) to a sub-area that includes the optoelectronically readable code (1200).

Images