Two-dimensional code identification and positioning
The present specification provides a two-dimensional code identification method and device, and a two-dimensional code positioning and identification model establishment method and device. The two-dimensional code identification method includes: obtaining a to-be-identified two-dimensional code, and performing global feature positioning detection on the to-be-identified two-dimensional code by using a pre-established two-dimensional code positioning and identification model; performing focus adjustment, based on a predetermined image resolution, on the to-be-identified two-dimensional code on which positioning detection is performed; and decoding the to-be-identified two-dimensional code on which focus adjustment is performed. The present specification can improve the identification accuracy of two-dimensional codes shot in complex scenarios.
1. A computer-implemented method, comprising:
obtaining two or more images that represent one or more reference two-dimensional codes using a determined sampling mode, wherein the two or more images include at least a first image and a second image;
obtaining a sample two-dimensional code;
comparing the sample two-dimensional code to the first image and to the second image;
determining that the sample two-dimensional code best matches the first image based on comparing the sample two-dimensional code to the first image and to the second image;
positioning global features of the one or more reference two-dimensional codes based on the sample two-dimensional code and identifier information of the sample two-dimensional code; and
training a two-dimensional code positioning and identification model by providing the identifier information of the sample two-dimensional code as input data to the two-dimensional code positioning and identification model.
2. The computer-implemented method of claim 1 , wherein obtaining the two or more images that represent the one or more reference two-dimensional codes using the determined sampling mode comprises:
obtaining the first image of the two or more images from a collection device a first distance from a reference two-dimensional code of the one or more reference two-dimensional codes; and
obtaining the second image of the two or more images from the collection device a second distance from the reference two-dimensional code of the one or more reference two-dimensional codes.
3. The computer-implemented method of claim 2 , wherein comparing the sample two-dimensional code to the first image and to the second image comprises:
comparing a third resolution of the sample two-dimensional code to a first resolution of the first image and to a second resolution of the second image.
4. The computer-implemented method of claim 1 , wherein obtaining the two or more images that represent the one or more reference two-dimensional codes using the determined sampling mode comprises:
obtaining the first image of the two or more images at a first angle; and
obtaining the second image of the two or more images at a second angle.
5. The computer-implemented method of claim 1 , wherein obtaining the two or more images that represent the one or more reference two-dimensional codes using the determined sampling mode comprises:
obtaining the first image of the two or more images corresponding to a first environment condition; and
obtaining the second image of the two or more images corresponding to a second environment condition.
6. The computer-implemented method of claim 1 , wherein the global features comprise four corner points corresponding to an upper left corner, a lower left corner, an upper right corner, and a lower right corner.
7. The computer-implemented method of claim 1 , wherein the two-dimensional code positioning and identification model comprises a machine learning network, and wherein the machine learning network comprises a convolution neural network, deep learning network, deep convolutional neural network, regions with convolutional neural networks (R-CNN), Faster R-CNN, regression-based detection methods, you only look once (YOLO), or single-shot detector (SSD).
8. A non-transitory, computer-readable medium storing one or more instructions executable by a computer system to perform operations comprising:
obtaining two or more images that represent one or more reference two-dimensional codes using a determined sampling mode, wherein the two or more images include at least a first image and a second image;
obtaining a sample two-dimensional code;
comparing the sample two-dimensional code to the first image and to the second image;
determining that the sample two-dimensional code best matches the first image based on comparing the sample two-dimensional code to the first image and to the second image;
positioning global features of the one or more reference two-dimensional codes based on the sample two-dimensional code and identifier information of the sample two-dimensional code; and
training a two-dimensional code positioning and identification model by providing the identifier information of the sample two-dimensional code as input data to the two-dimensional code positioning and identification model.
9. The non-transitory, computer-readable medium of claim 8 , wherein obtaining the two or more images that represent the one or more reference two-dimensional codes using the determined sampling mode comprises:
obtaining the first image of the two or more images from a collection device a first distance from a reference two-dimensional code of the one or more reference two-dimensional codes; and
obtaining the second image of the two or more images from the collection device a second distance from the reference two-dimensional code of the one or more reference two-dimensional codes.
10. The non-transitory, computer-readable medium of claim 9 , wherein comparing the sample two-dimensional code to the first image and to the second image comprises:
comparing a third resolution of the sample two-dimensional code to a first resolution of the first image and to a second resolution of the second image.
11. The non-transitory, computer-readable medium of claim 8 , wherein obtaining the two or more images that represent the one or more reference two-dimensional codes using the determined sampling mode comprises:
obtaining the first image of the two or more images at a first angle; and
obtaining the second image of the two or more images at a second angle.
12. The non-transitory, computer-readable medium of claim 8 , wherein obtaining the two or more images that represent the one or more reference two-dimensional codes using the determined sampling mode comprises:
obtaining the first image of the two or more images corresponding to a first environment condition; and
obtaining the second image of the two or more images corresponding to a second environment condition.
13. The non-transitory, computer-readable medium of claim 8 , wherein the global features comprise four corner points corresponding to an upper left corner, a lower left corner, an upper right corner, and a lower right corner.
14. The non-transitory, computer-readable medium of claim 8 , wherein the two-dimensional code positioning and identification model comprises a machine learning network, and wherein the machine learning network comprises a convolution neural network, deep learning network, deep convolutional neural network, regions with convolutional neural networks (R-CNN), Faster R-CNN, regression-based detection methods, you only look once (YOLO), or single-shot detector (SSD).
15. A computer-implemented system, comprising:
one or more computers; and
one or more computer memory devices interoperably coupled with the one or more computers and having tangible, non-transitory, machine-readable media storing one or more instructions that, when executed by the one or more computers, perform one or more operations comprising:
obtaining two or more images that represent one or more reference two-dimensional codes using a determined sampling mode, wherein the two or more images include at least a first image and a second image;
obtaining a sample two-dimensional code;
comparing the sample two-dimensional code to the first image and to the second image;
determining that the sample two-dimensional code best matches the first image based on comparing the sample two-dimensional code to the first image and to the second image;
positioning global features of the one or more reference two-dimensional codes based on the sample two-dimensional code and identifier information of the sample two-dimensional code; and
training a two-dimensional code positioning and identification model by providing the identifier information of the sample two-dimensional code as input data to the two-dimensional code positioning and identification model.
16. The computer-implemented system of claim 15 , wherein obtaining the two or more images that represent the one or more reference two-dimensional codes using the determined sampling mode comprises:
obtaining the first image of the two or more images from a collection device a first distance from a reference two dimensional code of the one or more reference two-dimensional codes; and
obtaining the second image of the two or more images from the collection device a second distance from the reference two dimensional code of the one or more reference two-dimensional codes.
17. The computer-implemented system of claim 16 , wherein comparing the sample two-dimensional code to the first image and to the second image comprises:
comparing a third resolution of the sample two-dimensional code to a first resolution of the first image and to a second resolution of the second image.
18. The computer-implemented system of claim 15 , wherein obtaining the two or more images that represent the one or more reference two-dimensional codes using the determined sampling mode comprises:
obtaining the first image of the two or more images at a first angle; and
obtaining the second image of the two or more images at a second angle.
19. The computer-implemented system of claim 15 , wherein obtaining the two or more images that represent the one or more reference two-dimensional codes using the determined sampling mode comprises:
obtaining the first image of the two or more images corresponding to a first environment condition; and
obtaining the second image of the two or more images corresponding to a second environment condition.
20. The computer-implemented system of claim 15 , wherein the global features comprise four corner points corresponding to an upper left corner, a lower left corner, an upper right corner, and a lower right corner.
21. The computer-implemented system of claim 15 , wherein the two-dimensional code positioning and identification model comprises a machine learning network, and wherein the machine learning network comprises a convolution neural network, deep learning network, deep convolutional neural network, regions with convolutional neural networks (R-CNN), Faster R-CNN, regression-based detection methods, you only look once (YOLO), or single-shot detector (SSD).