Method, automation system and computer system for detecting optical codes

1. A method for detecting optical codes, the method comprising the following: providing an input image; detecting one or more optical codes in the input image by generating an output image on the basis of the input image; wherein within a framework of generating the output image, pixels of an area of the input image comprising a part of an optical code are assigned a first intensity in the output image; wherein within the framework of generating the output image, pixels of an area of the input image comprising no part of the optical code are assigned a second intensity in the output image; and providing regions of the output image with the first intensity for determining properties of the detected optical codes.

2. The method according to claim 1 , wherein generating the output image is carried out by a first neural network, or a deep, trained, convolutional first neural network.

3. The method according to claim 2 , wherein providing the input image comprises adjusting the input image; and wherein adjusting the input image is carried out in such way that the adjusted input image comprises a size compatible with a size of a first input layer of the first neural network or the deep, trained, convolutional first neural network.

4. The method according to claim 1 , wherein for determining the properties of the detected optical codes, a determination of positions of detected optical codes is carried out; and wherein determining the positions of the detected optical codes is carried out by assigning enclosing rectangles, outer contours and/or geometrical centers to the regions of the output image assigned with the first intensity.

5. The method according to claim 4 , wherein assigning the enclosing rectangles, outer contours and/or geometrical centers is carried out by blob detection, morphological operators and/or a binarization of the output image in combination with contour tracking.

6. The method according to claim 4 , wherein determining the positions of the detected optical codes comprises transforming the enclosing rectangles, the outer contours and/or the geometrical centers to correspond to the input image.

7. The method according to claim 4 , wherein for determining the properties of detected optical codes an extraction of code images is carried out on the basis of the determined positions of the detected optical codes and determining orientations and/or categories of the detected optical codes is carried out on the basis of code images.

8. The method according to claim 7 , wherein determining orientations and/or categories of the detected optical codes is carried out by a second neural network, or a deep, trained, convolutional second neural network.

9. The method according to claim 8 , wherein the extraction of the code images comprises adjusting the code images; and wherein adjusting the code images is carried out in such a way that the adjusted code images each comprise a size compatible with a size of a second input layer of the second neural network or the deep, trained, convolutional second neural network.

10. The method according to claim 7 , wherein determining orientations and/or categories of the detected optical codes comprises extracting general features by a third neural network, or by an auto-encoder, and a regression and/or classification.

11. The method according to claim 10 , wherein a characteristic of the extracted general features is examined for novelty by a novelty detector.

12. The method according to claim 1 , wherein for determining the properties of the detected optical codes information stored in the detected optical codes is read out.

13. The method according to claim 1 , wherein determining a property of one single optical code of the detected optical codes, determining a plurality of properties of one single optical code of the detected optical codes, determining a property of a plurality of the detected optical codes, and/or determining a plurality of properties of a plurality of the detected optical codes is carried out in parallel.

14. A non-transitory, computer readable data storage medium having computer program code stored thereon, the computer program code comprising commands that, when executing the computer program code on a computer, cause said computer to carry out a method for detecting optical codes, the method comprising: providing an input image; detecting one or more optical codes in the input image by generating an output image on the basis of the input image; wherein within a framework of generating the output image pixels of an area of the input image comprising a part of an optical code are assigned a first intensity in the output image; wherein within the framework of generating the output image pixels of an area of the input image comprising no part of the optical code are assigned a second intensity in the output image; and providing regions of the output image with the first intensity for determining properties of the detected optical codes.