Patent Application
20230118305
The present invention relates to a method and an apparatus for identifying an article. The invention also relates to the use of such a method or the use of such an apparatus in an industrial production.
In the manufacture of articles, in particular in the manufacture of sheet metal parts, an article, in particular a manufactured sheet metal part, is often found that cannot be associated with a specific order. Disposing of the article is a waste of resources and also harbors the risk of the article being missing from a particular order. However, the completely manual identification of the article is time-consuming and can often only be carried out successfully by experienced employees who are not always present or who have to be released from other tasks to identify the article. Completely automatic identification of an article, in contrast, often fails due to insufficient image capture of the article or inadequate identification programs.
In an embodiment, the present invention provides a method that identifies an article by associating an image of the article with stored article data. The method includes: A) creating the image of the article; B) extracting image features from the image; and C) extracting article features from the stored article data. The method step C) are configured to be carried out before, during, or after method steps A) and B). The following method steps are then carried out: D) comparing the image features to the article features; E) output of association information of the image with the stored article data.
Embodiments of the present invention will be described in even greater detail below based on the exemplary figures. The present invention is not limited to the exemplary embodiments. All features described and/or illustrated herein can be used alone or combined in different combinations in embodiments of the present invention. The features and advantages of various embodiments of the present invention will become apparent by reading the following detailed description with reference to the attached drawings which illustrate the following:
Embodimentsof the present invention provide a method and an apparatus that allows an effective association of a randomly found article. Embodiments of the present invention further provide a corresponding use of the method or the apparatus.
The present invention relates to a method for identifying an article. The article can be designed in the form of an essentially two-dimensional article, for example in the form of a flat sheet metal part, or in the form of a three-dimensional article, for example in the form of a sheet metal part having deformations. The identification takes place by associating an image of the article with stored article data. These article data can be stored in the form of CAD (computer assisted design) data. The method has the following method steps:
According to the present invention, it is thus provided to output at least one piece of information relating to the association of the article to be identified with stored article data. The information is preferably specified in the form of probability information, so that it can be seen with what probability the article corresponds to the stored article data. The output can take place, for example, on a screen, data glasses or as data transfer. Advantageously, a plurality of pieces of association information can be output with an indication of the probability.
According to the present invention, features are extracted from the image as well as from the stored article data. This allows a particularly reliable association of the image with the stored article data and thus a particularly high probability of identifying the article.
The method preferably has the following method step: F) reading out a user rating of the association information.
The method then provides for reading an evaluation of a user with regard to this association information. The evaluation by the user can take place, for example, by data transmission or by an input device such as a button, touchscreen, voice recording, speech recognition or similar input devices. If, for example, two or more identification hits are displayed to the user, sorted according to probability, then the user can select the one, which in his opinion really matches. He has thus given a user rating. This can make it possible to improve association information that will be output in the future.
An image extraction algorithm can read in data from the image, process it depending on predetermined image parameters and output image features in the form of the processed data. The predetermined image parameters can have so-called weighted variables. Their function and ascertainment are explained below.
The method can further preferably have the following method step:
The parameters can have so-called weighted variables. Their ascertainment and function are explained below.
The following method step is preferably carried out before, during and/or after method step H):
Saving the user rating allows a large number of user ratings to be collected, as a result of which the image extraction algorithm, the article extraction algorithm and/or the comparison algorithm can be significantly improved.
The user rating results memory can be cloud-based. Cloud-based here means an, in particular remotely located, preferably anonymized, storage device in which user ratings from more than one, advantageously from several hundred or several thousand different users, are stored. In this way, different users can contribute to the optimization of the method regardless of the manufacturing location. It has been recognized that the described method does not achieve resounding successes; that is to say, association information having the correct association with the highest probability is received only when several tens of thousands, in particular several hundred thousand, user ratings have been read out. Such a volume of data is often not achievable in a year for a single production facility. Accordingly, the method would probably have presumably remained uninteresting.
The image can be created or captured in the wavelength range that is visible to the human eye. As an alternative or in addition to this, the image can be made in the wavelength range that is not visible to the human eye, for example in the IR range, in the UV range and/or in the ultrasound range.
In method step A), a plurality of images of the article can be captured with the image capturing device. In method step B), the image features can be extracted from a plurality of images. In method step E), association information with the stored articles from a plurality of images can be output. It has been shown that the creation and processing of a plurality of images significantly improves the quality of the association information for a user, since in particular the influence of artifacts caused by the image position is reduced.
The image extraction algorithm, the article extraction algorithm and/or the comparison algorithm preferably has/have an algorithm with a plurality of data aggregation routines.
A data aggregation routine can be designed to aggregate a plurality of “ascertained data” into a new data packet. The new data packet can have one or more numbers or vectors. The new data packet can be made available in full or in part to further data aggregation routines as “ascertained data.” “Ascertained data” can be, for example, image data, article data or data packets made available by one of the data aggregation routines.
Particularly preferably, the image extraction algorithm, the article extraction algorithm and/or the comparison algorithm is/are designed in the form of an algorithm having a plurality of connected data aggregation routines. In particular, several hundred, in particular several thousand, such data aggregation routines can be interconnected. This significantly improves the quality and speed of the algorithm(s).
The image extraction algorithm, the article extraction algorithm and/or the comparison algorithm can have a function with weighted variables. One, in particular a plurality of, particularly preferably all, data aggregation routines can be designed to combine a plurality of “ascertained data” in each case with a weighted variable, in particular to multiply and, thus, convert the “ascertained data” into “combined data” then to aggregate, in particular add, the “combined data” into a new data packet. The change in weighting, in particular the change in the weighted variables, is particularly preferably carried out on the basis of the user rating. In order to ascertain suitable weighted variables, the image extraction algorithm, the article extraction algorithm and/or the comparison algorithm can be run through with data, in particular image data and/or article data whose association is known in each case. The weighted variables can be determined, preferably in a first phase, separately for the image extraction algorithm, the article extraction algorithm and the comparison algorithm.
The image features and article features can themselves be data packets, in particular a plurality of structured data, in particular data vectors or data arrays, which themselves may again represent “ascertained data,” e.g. for the comparison algorithm, in particular for the data aggregation routines of the comparison algorithm. The exact structure of these image features and article features can change, in particular improve, preferably be optimized, via the automated evaluation of the user rating.
Since the weighted variables, which are changed, in particular improved, preferably optimized by the user rating of a first user or a first group of users, are managed in a cloud-based manner, other users can also use them in their algorithms and benefit from the method.
The aforementioned algorithms or a further secondary or superordinate algorithm can be designed to monitor and recognize when any or all of the algorithms output association information with a predefined accumulation that is rated as poor by the user and then output a negative indication. The output can be visual, for example on a screen, or in another suitable form, for example as data output. The monitoring algorithm can also be designed to react to the output of such a negative indication with an improvement routine that changes further properties or the interaction of one or more of the aforementioned algorithms.
In a further preferred embodiment of the method, the article features are available in the form of design data, material data, surface quality data and/or thermal conductivity data.
The comparison algorithm may comprise the formation of a scalar product and/or the formation of a difference between the image features and the article features. The aforementioned measures have proven to be particularly effective in associating the image features with the article features.
Furthermore, the outputting of the association information of the image preferably includes the output of a plurality of probabilities of association with different stored articles. A user can thus choose from various association options. The identification of the article can therefore be carried out with a very high probability of success.
An embodiment of the present invention provides an apparatus for identifying an article by associating the article with stored article data, the apparatus having the following:
The apparatus according to the invention is preferably designed to carry out a method described here.
The image capturing device can be designed in the form of a camera, in particular for visible light.
The apparatus can furthermore have the following:
Furthermore, the apparatus can have the following:
The image extraction algorithm, the article extraction algorithm and/or the comparison algorithm can be designed in the form of an algorithm having a plurality of connected data aggregation routines.
The method and/or the apparatus can be used particularly advantageously in industrial production with a computer-based production control for processing reflective articles, in particular sheet metal parts. In the past, the images of sheet metal parts captured with an image capturing device were often of insufficient quality for identification, because contours can only be distinguished from the background with great difficulty and light reflections simulate false contours. For this reason, methods and apparatuses that compare image data to article data without the described algorithms have so far not led to any success in industrial production. The term “reflective articles” is intended to mean articles having a smooth surface which reflect the light in such a way that when the image is captured, in addition to the contours, undesired light reflections from other articles can also occur. Examples of such reflective articles are metals, glass, plastics having a smooth surface, coated materials such as coated plates made of plastics, wood, metal, glass, etc.
It is particularly advantageous if the production control is at least partially cloud-based. Parameters, in particular the weighted variables, can then be used for changing, in particular improving, in particular optimizing, the algorithms from a first production facility in other production facilities and vice versa. There is a much larger database available and the identification for each individual production facility can be significantly improved.
Further advantages of the invention result from the description and the drawings. Likewise, according to the invention, the aforementioned features and those which are to be explained below can each be used individually for themselves or for a plurality of combinations of any kind. The embodiment shown and described is not to be understood as exhaustive but rather has an exemplary character for the description of the invention.
The article data 12 are preferably stored in the form of CAD data, in particular in an article database 16 in the form of a CAD database.
The image 14 is created with an image capturing device 18. The image capturing device 18 can be designed in the form of a camera.
The image 14 shown in
An image extraction algorithm 20 is applied to image 14 in order to extract image features 22. This is indicated schematically in
An article extraction algorithm 26 is applied to the article data 12 in order to extract article features 28. This is indicated schematically in
The image features 22 and the article features 28 are fed to a comparison algorithm 32. The comparison algorithm 32 is stored in a comparison unit 34. The comparison algorithm 32 can have data aggregation routines that are interconnected, in particular weighted relative to one another. The comparison algorithm 32 is preferably designed to form a scalar product or a difference between the image features 22 and the article features 28.
As a result of the comparison algorithm 32, association information 36 is output. The association information is output in an output unit 38. As shown in
In order to improve future associations, i.e. to increase the quality of future association information, the apparatus 10 has an input unit 40. The input unit 40 is designed to read out a user rating 42. The user rating 42 is then used to optimize the image extraction algorithm 20, the article extraction algorithm 26 and/or the comparison algorithm 32 or the parameters thereof.
The user rating 42 can be stored in a user rating results memory 44, so that the method or the apparatus 10 can be optimized with a large number of user ratings 42. Particularly preferably, the user rating results memory 44 is cloud-based. As a result, user ratings 42 across devices can flow into the optimization of the method or the apparatus 10.
The invention relates to a method and an apparatus 10 for recognizing an article. At least one image 14, in particular in the form of a photograph, of the article is created. Based on the image 14, image features 22 are ascertained by means of an image extraction algorithm 20. On the basis of stored article data 12, article features 28 are ascertained and compared to the image features 22 in order to output association information 36. According to the invention, provision is made in particular to provide a user rating 42 both to improve the image extraction algorithm 20 and the article extraction algorithm 26. As an alternative or in addition to this, it is provided according to the invention in particular that both the image extraction algorithm 20 and the article extraction algorithm 26 are based on interconnected, preferably weighted, data aggregation routines.
While embodiments of the invention have been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary and not restrictive. It will be understood that changes and modifications may be made by those of ordinary skill within the scope of the following claims. In particular, the present invention covers further embodiments with any combination of features from different embodiments described above and below. Additionally, statements made herein characterizing the invention refer to an embodiment of the invention and not necessarily all embodiments.
The terms used in the claims should be construed to have the broadest reasonable interpretation consistent with the foregoing description. For example, the use of the article “a” or “the” in introducing an element should not be interpreted as being exclusive of a plurality of elements. Likewise, the recitation of “or” should be interpreted as being inclusive, such that the recitation of “A or B” is not exclusive of “A and B,” unless it is clear from the context or the foregoing description that only one of A and B is intended. Further, the recitation of “at least one of A, B and C” should be interpreted as one or more of a group of elements consisting of A, B and C, and should not be interpreted as requiring at least one of each of the listed elements A, B and C, regardless of whether A, B and C are related as categories or otherwise. Moreover, the recitation of “A, B and/or C” or “at least one of A, B or C” should be interpreted as including any singular entity from the listed elements, e.g., A, any subset from the listed elements, e.g., A and B, or the entire list of elements A, B and C.
10 apparatus
12 article data
14 image
16 article database
18 image capturing device
20 image extraction algorithm
22 image features
24 image extraction unit
26 article extraction algorithm
28 article features
30 article extraction unit
32 comparison algorithm
34 comparison unit
36 association information
38 output unit
40 input unit
42 user rating
44 user rating results memory
| Number | Date | Country | Kind |
|---|---|---|---|
| 10 2018 215 538.0 | Sep 2018 | DE | national |
This application is a continuation of International Patent Application No. PCT/EP2019/073383, filed on Sep. 03, 2019, which claims priority to German Patent Application No. DE 10 2018 215 538.0, filed on Sep. 12, 2018. The entire disclosure of both applications is hereby incorporated by reference herein.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2019/073383 | 9/3/2019 | WO |
