TWI861640B - Curved image positioning printing device - Google Patents

Abstract

The purpose of the present invention is to provide a curved surface image positioning printing device that can be directly printed on a non-flat surface product without manual adjustment and correction of the alignment, which includes: a printing platform, a depth camera, a plane camera, an image deformation pre-processing module, a curved surface precise alignment coating software and a 2.5D non-flat surface object model generation module; the image deformation pre-processing module can perform graphic pre-processing on a 2D image and convert it into a deformed image, The depth camera captures the depth information and color information of the product to be printed, and the PCL point group library is used to segment the 2.5D non-flat surface object model into multiple point groups. The surface precision alignment covering software uses the ICP algorithm to overlay the multiple point groups after segmentation of the 2.5D non-flat surface object model with the deformation image to form image data. The printing platform can print on the non-flat surface of the product to be printed based on the image data.

Description

Curved surface image positioning printing device

The present invention relates to the technical field of digital inkjet printing, and in particular to a curved image positioning inkjet printing device.

Printing or inkjet printing technology has been widely used in modern industry. If it is flat printing or inkjet printing, the process is quite simple, but if the object has a curved surface, it is difficult to print or inkjet print. Curved surface printing and flat printing have different technical difficulties in the deviation correction of printed content. The printed content of flat printing only has the position deviation of the printing plane defined by the X-Y axis, while the printed content of curved surface printing involves the position deviation caused by the relative tilt angle between the curved surface of the object and the Z axis. Therefore, special equipment is required to print on the curved surface of an object. A known curved surface printing device uses a transfer head with a curved surface to transfer the printed content preset on the flat surface of a flat mold to the curved surface of the object. Due to many factors, the process of curved surface transfer will inevitably cause the position or shape deviation of the printed content transferred to the curved surface of the object.

However, in the known curved surface transfer equipment, the technology for correcting the above deviation is all done manually. This method must be performed by experienced and skilled technicians. The execution process first requires trial printing on the curved surface of the object, and then manually judge and manually adjust the relative position between the transfer head and the stamp, and the relative position between the transfer head and the curved surface of the object. This method usually requires multiple manual adjustments to correct the above deviations. The disadvantage is that it requires more labor costs, and the operator may not be able to maintain a very accurate print position reading for a long time due to factors such as fatigue or vision loss. In addition, adding an operating platform will also take up space and increase costs.

In view of this, how to create a curved surface image positioning printing device that can improve the above-mentioned shortcomings and can be directly printed on non-flat surface products without manual adjustment and correction of the alignment, so as to make the application range wider and improve the added value and practicality of the printing device, will be what this creation wants to actively reveal.

In order to solve the above problems and achieve the purpose of the present invention, the technical means of the present invention is implemented as follows: a curved surface image positioning printing device, which comprises: a printing platform (1) capable of placing a product to be printed (10) and inputting a 2D image (20); a depth camera (2) arranged on the printing platform (1) for capturing depth information and color information of the product to be printed (10); a plane camera (3) arranged on the printing platform (1) and located on one side of the depth camera (2); an image deformation pre-processing module (4) arranged in the printing platform (1); 4); a curved surface precise alignment coating software (5) disposed in the printing platform (1); and a 2.5D non-planar surface object model generation module (6) disposed in the printing platform (1), which can cooperate with the depth camera (2) to output a 2.5D non-planar surface object model (30); the image deformation preprocessing module (4) can perform image preprocessing on the 2D image (20) to convert it into a deformed image (40), and can capture the depth information and color information of the product to be printed (10) through the depth camera (2), using PCL (Point Language Language (PLL)). Cloud Library) point group library to segment the 2.5D non-flat surface object model (30) into a plurality of point groups; the surface precise alignment coating software (5) adopts ICP (Iterative Closest Point) algorithm to overlap the plurality of point groups after segmentation of the 2.5D non-flat surface object model (30) with the deformation image (40) to form an image data (50); thereby, the printing platform (1) can print on the non-flat surface of the product to be printed (10) according to the aforementioned image data (50).

More preferably, the graphic preprocessing step of the image deformation preprocessing module (4) further includes a point cloud repairing step (A), which comprises: step 1 (A1): using the holistically-nested edge detection (HED) to perform boundary detection on the two-dimensional color image, capturing the accurate boundary of the image, capturing the image coordinate information of the surface to which the boundary belongs and defining the pixel to be repaired; step 2 (A2): applying the iterative low-pass dilation method to repair the pixel to be repaired; and step 3 (A3): reversely projecting the pixels with depth values on the repaired depth image to the point cloud, thus completing the repair.

More preferably, the graphic preprocessing step of the image deformation preprocessing module (4) further includes an image capture step (B): step 1 (B1): image to be processed; step 2 (B2): image grayscale; step 3 (B3): dynamic threshold value method binarization; step 4 (B4): median filtering; and step 5 (B5): calibration plate information capture.

More preferably, the 2D image (20) is converted into a deformed image (40) with a pixel shift parameter less than or equal to 5 pixels.

More preferably, the offset between the positioning points of the deformation image (40) and the plurality of point group positioning points after segmentation of the 2.5D non-flat surface object model (30) is less than or equal to 10%.

More preferably, the printing platform (1) is an inkjet printer.

More preferably, the depth camera (2) is a CCD camera.

By implementing the above-mentioned invention, the functions and effects achieved by the invention are as follows: the invention can convert the image of the 2D image (20) into a deformed image (40) according to the depth information of the product (10) to be printed captured by the depth camera (2), and then automatically align and overlay the deformed image (40) and the 2.5D non-flat surface object model (30) through the surface precise alignment and overlay software (5) to improve the alignment between the two. The accuracy of the position reconstruction can avoid the deformation image (40) from being offset or deformed when it is printed on the curved surface of the product (10) to be printed, thereby effectively improving the accuracy of product printing and the production process. The curved surface product printing can be completed in one go, making the production process simpler and more convenient. Compared with the traditional curved surface printing equipment that needs to be manually adjusted and applied, the curved surface image positioning printing device of the present invention is more convenient and accurate, thereby improving the added value and practicality of the curved surface printing device.

1: Printing platform

2: Depth Camera

3: Flat camera

4: Image deformation preprocessing module

5: Curved surface precise alignment coating software

6: 2.5D non-flat surface object model

Generate module

10: Products to be printed

20:2D pictures

30:2.5D non-flat surface object model

40: Deformed image

50: Image data

A: Point cloud repair steps

A1: Step 1: Use complete boundary detection to detect the boundary of the two-dimensional color image, capture the accurate boundary of the image, and capture the image coordinate information of the surface to which the boundary belongs and define the pixels to be repaired.

A2: Step 2: Apply repeated low-pass dilation method to the pixels to be repaired

A3: Step 3: Reversely project the pixels with depth values on the repaired depth image to the point cloud to complete the repair.

B: Image capture steps

B1: Step 1: Image to be processed

B2: Step 2: Grayscale the image

B3: Step 3: Binarization using dynamic threshold method

B4: Step 4: Median filtering

B5: Step 5: Capture calibration plate information

[Figure 1] is a block diagram of the present invention.

[Figure 2] is a schematic diagram of the implementation of the present invention.

[Figure 3] is a schematic diagram of the process of the point cloud repair step of the present invention.

[Figure 4] is a schematic diagram of the image capture process of the present invention.

In order to further understand the features, technical means, and specific objectives and functions of the present invention, the present invention is described in further detail below in conjunction with the attached drawings and specific implementation methods.

As shown in FIG. 1 and FIG. 2, a curved surface image positioning printing device is disclosed, which comprises: a printing platform (1) on which a product to be printed (10) can be placed and a 2D image (20) can be input; a depth camera (2) disposed on the printing platform (1) for capturing depth information and color information of the product to be printed (10); a plane camera (3) disposed on the printing platform (1) and located on one side of the depth camera (2); an image deformation pre-processing module (4) disposed in the printing platform (1); and a depth camera (2) disposed on the printing platform (1). The invention relates to a surface accurate alignment coating software (5) in the printing platform (1); and a 2.5D non-flat surface object model generation module (6) arranged in the printing platform (1), which can cooperate with the depth camera (2) to output the 2.5D non-flat surface object model (30); the image deformation preprocessing module (4) can perform image preprocessing on the 2D image (20) and convert it into a deformed image (40), and can capture the depth information and color information of the product to be printed (10) through the depth camera (2), using PCL (Point Language Language (PLL)). Cloud Library) point group library to segment the 2.5D non-flat surface object model (30) into a plurality of point groups; the surface precise alignment coating software (5) uses ICP (Iterative Closest Point) algorithm to overlap and superimpose the plurality of point groups after segmentation of the 2.5D non-flat surface object model (30) with the deformation image (40) to form an image data (50), so that the printing platform (1) can print on the non-flat surface of the product to be printed (10) according to the aforementioned image data (50).

Thus, by using the image deformation preprocessing module (4), the depth information of the product to be printed (10) can be captured by the depth camera (2), and the image of the 2D image (20) can be converted into a deformed image (40). Then, the deformed image (40) and the 2.5D non-flat surface object model (30) can be automatically aligned and overlaid by the curved surface precise alignment and overlay software (5) to improve the accuracy of the alignment reconstruction between the two and avoid When the deformed image (40) is printed on the curved surface of the product (10) to be printed, it will be offset or deformed, thereby effectively improving the accuracy of product printing and the production process. The curved surface product printing can be completed at one time, making the production process simpler and more convenient. Compared with traditional curved surface printing equipment that requires manual adjustment and application, the curved surface image positioning printing device of the present invention is more convenient and accurate, thereby improving the added value and practicality of the curved surface printing device.

In the above, the 2D image (20) is converted into a deformed image (40), the pixel offset parameter is less than or equal to 5 pixels, and the offset of the positioning point of the deformed image (40) corresponding to the multiple point group positioning points after the segmentation of the 2.5D non-planar surface object model (30) is less than or equal to 10%, thereby making the positioning point of the deformed image (40) and the multiple point group positioning points after the segmentation of the 2.5D non-planar surface object model (30) more accurate when overlapping, and can be controlled within the spraying range of the print head during printing, without the need to adjust the angle of the product to be printed (10) when spraying the same area, thereby further saving production time.

In the above, the printing platform (1) is preferably an inkjet printer, but in actual application, it can be implemented in a different way according to the actual application. The inkjet printer is not the focus of this invention, so it is only briefly described here for your understanding.

In the above, the depth camera (2) is a CCD camera, which can capture the color image (Color Frame), depth image (Depth Frame), the corresponding relationship matrix between the color image and the depth image (Depth Frame To Color Frame Mapping), the frame number (Frame Number) and the time stamp (Time Stamp) of the product to be printed (10), so as to obtain the complete measurement information of the product to be printed (10) and improve the measurement accuracy.

As shown in FIG. 1 and FIG. 3 , the image preprocessing step of the image deformation preprocessing module (4) further includes a point cloud repair step (A), and the point cloud repair step (A) is as follows: Step 1 (A1): using a complete boundary detection (holistically-nested edge Detection, HED) performs boundary detection on the two-dimensional color image, captures the accurate boundary of the image, and captures the image coordinate information of the surface to which the boundary belongs and defines the pixels to be repaired; Step 2 (A2): Apply iterative low-pass dilation method to the pixels to be repaired; and Step 3 (A3): Inversely project the pixels with depth values on the repaired depth image to the point cloud to complete the repair; thereby, the accuracy of hole repair in the pixels to be repaired can be improved, thereby ensuring that the shape of the repaired pixels will not change.

As shown in FIG1 and FIG4, the image preprocessing step of the image deformation preprocessing module (4) further includes an image capture step (B): step 1 (B1): image to be processed; step 2 (B2): image grayscale; step 3 (B3): dynamic threshold value method binarization; step 4 (B4): median filtering; and step 5 (B5): calibration plate information capture.

The above details the method, function and effect of the present invention; however, the above is a preferred embodiment of the present invention and is not intended to limit the scope. Therefore, all modifications consistent with the intent of the present invention should be included in the patent scope of the present invention within the scope of equal effect.

1: Printing platform

2: Depth Camera

3: Flat camera

4: Image deformation preprocessing module

5: Curved surface precise alignment coating software

6: 2.5D non-flat surface object model generation module

Claims (7)

A curved surface image positioning printing device, comprising: A printing platform (1) capable of placing a product to be printed (10) and inputting a 2D image (20); A depth camera (2) disposed on the printing platform (1) for capturing depth information and color information of the product to be printed (10); A plane camera (3) disposed on the printing platform (1) and located on one side of the depth camera (2); An image deformation pre-processing module (4) disposed in the printing platform (1); A curved surface precise alignment coating software (5) disposed in the printing platform (1); and A 2.5D non-planar surface object model generation module (6) disposed in the printing platform (1) can cooperate with the depth camera (2) to output a 2.5D non-planar surface object model (30); The image deformation preprocessing module (4) can perform graphic preprocessing on the 2D image (20) to convert it into a deformed image (40), and can capture the depth information and color information of the product to be printed (10) through the depth camera (2), and use the PCL (Point Cloud Library) point group library to segment the 2.5D non-flat surface object model (30) into a plurality of point groups; The curved surface precise alignment coating software (5) adopts the ICP (Iterative Closest Point) algorithm to overlap and superimpose the multiple point groups after segmentation of the 2.5D non-flat surface object model (30) with the deformation image (40) to form an image data (50); thereby, the printing platform (1) can print on the non-flat surface of the product to be printed (10) according to the aforementioned image data (50). As described in claim 1, the curved image positioning printing device, wherein the graphic preprocessing step of the image deformation preprocessing module (4) further includes a point cloud repairing step (A), and the point cloud repairing step (A) is: Step 1 (A1): Use holistically-nested edge detection (HED) to detect the boundary of the two-dimensional color image, capture the accurate boundary of the image, and capture the image coordinate information of the surface to which the boundary belongs and define the pixels to be repaired; Step 2 (A2): Applying a repeated low-pass dilation method to the pixels to be repaired; and Step 3 (A3): Reversely project the pixels with depth values on the repaired depth image to the point cloud to complete the repair. As described in claim 1, the curved image positioning printing device, wherein the graphic preprocessing step of the image deformation preprocessing module (4) further includes an image capture step (B) as follows: Step 1 (B1): Image to be processed; Step 2 (B2): grayscale the image; Step 3 (B3): Binarization using dynamic threshold method; Step 4 (B4): Median filtering: and Step 5 (B5): Capture calibration plate information. A curved image positioning printing device as described in claim 1, wherein the 2D image (20) is converted into a deformed image (40), and the pixel offset parameter is less than or equal to 5 pixels. A curved image positioning printing device as described in claim 1, wherein the offset of the positioning point of the deformation image (40) corresponding to the plurality of point group positioning points after segmentation of the 2.5D non-flat surface object model (30) is less than or equal to 10%. The curved image positioning printing device as described in claim 1, wherein the printing platform (1) is an ink direct inkjet printer. The curved surface image positioning printing device as described in claim 1, wherein the depth camera (2) is a CCD camera.

Images