Decorative firing and gilding detection method and device

Abstract

The invention discloses a decoration firing and gilding detection method and device in the technical field of non-contact optical detection. The method comprises building a detection system; the method comprises the following steps: placing N glass bottle samples on a workpiece table, collecting multi-view images, performing morphological processing to obtain a sample set, and manually marking the images in the sample set; building a deep learning network, and inputting a sample set for training; performing multi-view imaging on the to-be-detected glass bottle, performing morphological processing on an imaging result, and inputting the imaging result into the deep learning network for initial detection; the similarity of decoration firing and gilding marking results is used as feedback input, the upper computer controls the workbench to adjust the glass bottle, and secondary imaging is carried out on the marking position; and performing fine detection on the secondarily imaged image after morphological processing, and outputting a detection result. In the application, the method has good automation capability, high-precision detection of decoration firing and gilding can be realized through secondary imaging, adverse effects of light intensity and spectrum can also be inhibited through morphological processing, and the detection precision is improved.

Description

Technical Field

[0001] The invention relates to the technical field of non-contact optical detection, and in particular to a method and device for detecting baking and gilding. Background Art

[0002] Decorative baking and gilding on the surface of glass bottles are important representatives of brand image. Therefore, for the related industries using glass bottles, high-precision detection of baking and gilding on the surface of glass bottles is extremely important for their product quality image and market influence. In current practice, baking and gilding detection are mostly manual detection, which has the following problems: high missed detection rate, low quantification degree and low detection efficiency. Summary of the invention

[0003] In order to overcome the problems of high missed detection rate, low quantification degree and low detection efficiency in the existing glass bottle baking and gilding detection, the present invention provides a baking and gilding detection method and device.

[0004] The technical solution adopted by the present invention to solve its technical problem is:

[0005] The morphological-based method for detecting baked and gilded glass bottles includes the following steps:

[0006] S1. Build a detection system, which consists of a spherical light source, a workbench, a multi-view imaging system and a host computer. The multi-view imaging system and the spherical light source are used to image the glass bottles on the workbench, and the multi-view imaging system is electrically connected to the host computer, and the host computer has feedback adjustment capabilities for the workbench;

[0007] S2, placing N glass bottle samples on the workpiece table in turn, collecting multi-view images and performing morphological processing to obtain a set of sample sets, and manually marking each image in the sample set, with the marking contents being fire-roasting and gilding contents;

[0008] S3, build a deep learning network and input sample sets for training;

[0009] S4. After the training is completed, the glass bottle to be tested is imaged from multiple perspectives. The imaging results are morphologically processed and then input into the deep learning network for preliminary inspection. The baked and gilded parts are marked in the preliminary inspection results.

[0010] S5, using the similarity of the results of baking and gilding marking as feedback input, the upper computer controls the workbench to adjust the glass bottle and perform secondary imaging of the marking position;

[0011] S6. The secondary image is subjected to precise detection after morphological processing, and the detection result is output.

[0012] In this application, a multi-view imaging system is used to perform imaging analysis on glass bottles to reduce the missed detection rate and improve detection efficiency. The host computer controls the adjustment of the workbench to further improve the degree of automation. Secondary imaging can achieve high-precision detection of baked patterns and gilding. Morphological processing can also suppress the adverse effects of light intensity and spectrum, thereby improving detection accuracy.

[0013] In some embodiments, the workbench in S1 has two-dimensional translation capability and rotation adjustment capability;

[0014] The feedback input of the host computer in S5 is the similarity of the baking and gilding detection, and the feedback output is the two-dimensional translation and rotation angle of the workbench.

[0015] In some embodiments, the multi-view imaging system in S1 includes at least two imaging devices, each of which includes a set of dual telecentric imaging lenses and a camera.

[0016] In some embodiments, the morphological processing in S2 and S6 includes binarization, image erosion or dilation.

[0017] The present invention also provides a device for the baking and gilding detection method described in any of the above embodiments, comprising a spherical light source, a workbench, a multi-view imaging system and a host computer. The multi-view imaging system and the spherical light source are distributed on both sides of the workbench and are used for imaging glass bottles on the workbench. The multi-view imaging system and the host computer are electrically connected, and the host computer and the workbench are electrically connected to enable the host computer to adjust the position of the glass bottle on the workbench.

[0018] In this application, a multi-view imaging system is used to perform imaging analysis on glass bottles to reduce missed detection rate and improve detection efficiency; the upper computer controls the workbench adjustment to further improve the degree of automation.

[0019] In some embodiments, the multi-view imaging system includes at least two imaging devices, each of which includes a set of dual telecentric imaging lenses and a camera.

[0020] In some embodiments, a morphological processing module is further included for processing images output by the multi-view imaging system.

[0021] In some embodiments, the worktable has two-dimensional translation capability and rotation adjustment capability.

[0022] In some embodiments, a deep learning module is deployed on the host computer, and the deep learning module is used to decide the adjustment amount of the workbench by the host computer.

[0023] The beneficial effects of the present invention are:

[0024] The multi-view imaging system is used to analyze the glass bottles to reduce the missed detection rate and improve the detection efficiency. The host computer controls the adjustment of the workbench to further improve the degree of automation. The secondary imaging can achieve high-precision detection of baked patterns and gilding. Morphological processing can also suppress the adverse effects of light intensity and spectrum, improving detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 A schematic flow chart of the method for detecting the baking and gilding provided by the present invention;

[0026] Figure 2 The present invention is a schematic diagram of the structure of a device for the method of detecting baking and gilding provided by the present invention.

[0027] Marked in the figure: 1-spherical light source, 2-workbench, 3-multi-view imaging system, 4-host computer. DETAILED DESCRIPTION

[0028] The present invention is further described below in conjunction with the accompanying drawings.

[0029] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0030] Combination Figure 1 and Figure 2 As shown, the present invention provides a method and device for detecting baking and gilding.

[0031] refer to Figure 2 As shown, the baking and gilding detection device comprises a spherical light source 1, a workbench 2, a multi-view imaging system 3 and a host computer 4. The multi-view imaging system 3 and the spherical light source 1 are distributed on both sides of the workbench 2 and are used for imaging the glass bottles on the workbench 2. The multi-view imaging system 3 and the host computer 4 are electrically connected, and the host computer 4 and the workbench 2 are electrically connected. The host computer 4 has feedback adjustment capability for the workbench 2, so as to realize the adjustment of the position of the glass bottle on the workbench 2 by the host computer 4, thereby facilitating the detection of the baking and gilding positions.

[0032] In the present application, the glass bottles are imaged and analyzed by a multi-view imaging system 3 to reduce the missed detection rate and improve the detection efficiency; the upper computer 4 controls the adjustment of the workbench 2 to further improve the degree of automation.

[0033] In this embodiment, a morphological processing module is also included to process the image output by the multi-view imaging system 3, thereby suppressing the adverse effects of light intensity and spectrum and improving detection accuracy.

[0034] In this embodiment, the workbench 2 has two-dimensional translation capability and rotation adjustment capability. For example, the workbench 2 has a base, the base is arranged on a two-dimensional linear guide rail, and the workbench 2 can rotate around a central axis relative to the base.

[0035] In this embodiment, a deep learning module is deployed on the host computer, and the deep learning module is used to decide the adjustment amount of the host computer to the workbench. Through the deep learning module, the glass bottle samples are learned, and the baking and gilding positions are first marked in the subsequent glass bottle inspection, and then the host computer 4 controls the action of the workbench 2, so as to perform secondary imaging of the baking and gilding parts, further improving the automation degree and detection accuracy of the device.

[0036] refer to Figure 1 As shown in the figure, the morphology-based glass bottle baking and gilding detection method includes the following steps:

[0037] S1. Build a detection system, which consists of a spherical light source 1, a workbench 2, a multi-view imaging system 3 and a host computer 4. The multi-view imaging system 3 and the spherical light source 1 are used to image the glass bottles on the workbench 2, and the multi-view imaging system 3 and the host computer 4 are electrically connected. The host computer 4 has feedback adjustment capabilities for the workbench 2;

[0038] In this embodiment, the multi-view imaging system 3 includes two sets of imaging devices, each of which includes a set of dual telecentric imaging lenses and a camera. Two sets are provided, and the implementation is simpler and more convenient. Using the dual telecentric imaging lenses, proportional image transmission can be performed, and after appropriate calibration, it can be used for length measurement. The present invention uses a dual telecentric lens group to image the baked and gilded parts to be measured, and the spatial size information of the baked and gilded parts can be obtained.

[0039] S2, 30 glass bottle samples are placed on the workpiece table in turn, and 10 binocular multi-view images are collected respectively and then morphologically processed to obtain a set of sample sets, and each image in the sample set is manually marked, and the marked content is fire roasting and gilding content;

[0040] S3. Build a deep learning network based on convolutional neural network and input sample sets for training;

[0041] S4. After the training is completed, the glass bottle to be tested is imaged from multiple perspectives. The imaging results are morphologically processed and then input into the deep learning network for preliminary inspection. The baked and gilded parts are marked in the preliminary inspection results.

[0042] S5, taking the similarity of the results of baking and gilding marking as feedback input, the upper computer 4 controls the workbench 2 to adjust the glass bottle and perform secondary imaging on the marking position;

[0043] S6. The secondary image is subjected to precise detection after morphological processing, and the detection result is output.

[0044] In the present application, the method has good automation capability, and the secondary imaging can realize high-precision detection of baking and gilding. The morphological processing can also suppress the adverse effects of light intensity and spectrum, and improve the detection accuracy.

[0045] In this embodiment, the workbench 2 in S1 has two-dimensional translation capability and rotation adjustment capability; the feedback input of the host computer 4 in S5 is the similarity of the baking and gilding detection, and the feedback output is the two-dimensional translation amount and rotation angle of the workbench 2.

[0046] The purpose of the secondary imaging in S5 is to review the initial inspection results of the deep learning network, so as to achieve the key inspection of the initial inspection mark position image obtained in S4 in S6, and the inspection results are the inspection indicators required for baking and gilding inspection.

[0047] In this embodiment, the morphological processing in S2 and S6 includes binarization, image erosion or dilation. In practice, one of the three can be selected for use, or they can be used in combination.

[0048] In summary, the method and device for detecting baked patterns and gilding provided in the present application perform imaging analysis on glass bottles through a multi-view imaging system 3 to reduce the missed detection rate and improve the detection efficiency; the host computer 4 controls the adjustment of the workbench 2 to further improve the degree of automation; and the secondary imaging can achieve high-precision detection of baked patterns and gilding; morphological processing can also suppress the adverse effects of light intensity and spectrum, thereby improving detection accuracy.

[0049] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for detecting baking and gilding, characterized in that: The following steps are involved: S1. Build a detection system, which consists of a spherical light source (1), a workbench (2), a multi-view imaging system (3) and a host computer (4). The multi-view imaging system (3) and the spherical light source (1) are used to image the glass bottle on the workbench (2). The multi-view imaging system (3) and the host computer (4) are electrically connected, and the host computer (4) has feedback adjustment capability for the workbench (2); S2, placing N glass bottle samples on the workpiece table in turn, collecting multi-view images and performing morphological processing to obtain a set of sample sets, and manually marking each image in the sample set, with the marking contents being fire-roasting and gilding contents; S3, build a deep learning network and input sample sets for training; S4. After the training is completed, the glass bottle to be tested is imaged from multiple perspectives. The imaging results are morphologically processed and then input into the deep learning network for preliminary inspection. The baked and gilded parts are marked in the preliminary inspection results. S5, using the similarity of the results of baking and gilding marking as feedback input, the upper computer (4) controls the workbench (2) to adjust the glass bottle and perform secondary imaging of the marking position; S6. The secondary image is subjected to precise detection after morphological processing, and the detection result is output.

2. The method for detecting baking and gilding according to claim 1, characterized in that: The workbench (2) in S1 has two-dimensional translation capability and rotation adjustment capability; The feedback input of the upper computer (4) in S5 is the similarity of the baking and gilding detection, and the feedback output is the two-dimensional translation amount and rotation angle of the workbench (2).

3. The method for detecting baking and gilding according to claim 2, characterized in that: The multi-view imaging system (3) in S1 comprises at least two imaging devices, each of which comprises a set of dual telecentric imaging lenses and a camera.

4. The method for detecting decoration and gilding according to any one of claims 1 to 3, characterized in that: The morphological processing in S2 and S6 includes binarization, image erosion or dilation.

5. A device for the method for detecting decoration and gilding according to any one of claims 1 to 4, characterized in that: The invention comprises a spherical light source (1), a workbench (2), a multi-view imaging system (3) and a host computer (4); the multi-view imaging system (3) and the spherical light source (1) are distributed on both sides of the workbench (2) and are used for imaging glass bottles on the workbench (2); the multi-view imaging system (3) and the host computer (4) are electrically connected, and the host computer (4) and the workbench (2) are electrically connected, so that the host computer (4) can adjust the position of the glass bottle on the workbench (2).

6. The device for the method for detecting baking and gilding according to claim 5, characterized in that: The multi-view imaging system (3) comprises at least two imaging devices, each of which comprises a set of dual telecentric imaging lenses and a camera.

7. The device for detecting the baking and gilding process according to claim 6, characterized in that: It also comprises a morphological processing module, which is used to process the image output by the multi-view imaging system (3).

8. The device for detecting decoration and gilding according to any one of claims 5 to 7, characterized in that: The workbench (2) has two-dimensional translation capability and rotation adjustment capability.

9. The device for the method for detecting baking and gilding according to claim 8, characterized in that: A deep learning module is deployed on the host computer (4), and the deep learning module is used to decide the adjustment amount of the workbench (2) by the host computer (4).

Images