Image synthesis in stereo printing

From the three-dimensional printing process perspective, the synthesis of three-dimensional images is the most important part of the three-dimensional printing, its effect is directly related to the quality of printed matter. There are many ways to synthesize stereoscopic images. This article describes three typical stereoscopic image synthesis methods.

First, traditional stereo image synthesis

1. Methods for simultaneous photography and synthesis

The stereoscopic image capturing and synthesis are performed at the same time, and a stereoscopic camera with a large negative film is used. When the photographing is performed, the optical axis of the camera is centered on the subject and the grating superimposed on the photosensitive sheet is moved in the opposite direction.

2. Optical projection synthesis method

A multi-lens stereo camera or a single-mirror plane camera coupled with a positioning guide is used to take a set of flat images of the left and right scenes, and the film is subjected to a post-optical projection method.

1 grating and photosensitive film relative movement projection synthesis method

2 change projection angle projection synthesis method

Second, computer stereo image synthesis

1. Multi-view production of three-dimensional printing image processing method

It is the simplest stereo image processing method to make a three-dimensional printing image with left and right views, but it must be photographed by a professional stereo camera or simulated by simulating the observation angles of the left and right eyes. Since the photos taken in this way contain stereoscopic information, the image processing method is relatively simple, as long as the views are divided and combined, but each image must be the same size.

Generally, in order to obtain a better three-dimensional effect, generally two or more pictures are used, that is, shooting from multiple angles, so that the printed image can produce a better three-dimensional effect when viewed from different angles. According to the analysis of the imaging principle of the lenticular lens grating plate, as long as the respective views constituting the three-dimensional image are processed properly, the left and right views are respectively decomposed into very thin image bars and placed on both sides of the optical axis of the cylindrical lenses, respectively. Left and right eyes only see the corresponding view information.

The process of synthesizing a stereoscopic image from multiple views is as shown in FIG. 4 , the left and right visual images are longitudinally divided and numbered according to parameters of the grating plate, and then the same numbered bars are arranged in order according to the position and order of the images at the time of shooting to form a new one. The image includes stereoscopic information such that each image stripe (eg, A1, B1, C1 in FIG. 4) should occupy a full pitch. In this way, the corresponding position information of the set of scene objects is arranged in the same raster line number at equal intervals, and the key to the synthesis of raster fringe images lies in the mutual matching between the image scenes of the numbered scenes.

The left and right eyes can view the image of each point of the scene through different pitch units of the grating plate. The images formed by each unit viewed by the left and right eyes are different, but they can be synthesized in the brain to form a stereoscopic image. When the viewer makes any continuous movement on the same level, the combined images viewed by his left and right eyes also continuously change, and the three-dimensional image formed in the brain also changes continuously, as if the eyes were directly watching the real three-dimensional realm. This achieves the purpose of three-dimensional image display.

2. Method for processing stereoscopic printed images in plane view

Due to the development of computer image processing technology, very complicated image processing work can be realized. Through the image processing method, not only the left and right view images can be synthesized into a stereoscopic image that is the same as the observation of the actual scenery, but also arbitrary plane images that are not related to each other can be synthesized into a pseudo stereoscopic image, that is, a virtual stereoscopic image that does not actually exist is generated. . Using this method, not only can different images be combined into a pseudo-stereo image, but also different parts of a plane image can be made to have different stereoscopic effects. However, this method of image processing is more complicated than the multi-view production of three-dimensional printing image processing methods, and can only have a limited number of depth of field, and can not have any depth level like a real three-dimensional image, so called Pseudo-stereo view.

The biggest difference from the processing of multiview methods is that this method of processing the pseudo stereoscopic view does not arrange the divided image strips once in the left-right view order, but instead arranges them according to the stereoscopic imaging rules, and the left and right views are no longer arranged in the same column. Under the lens. As shown in Fig. 6, the A plane is a planar image closer to the eye, and the B plane is a planar image farther from the eye. To synthesize the two images into a stereoscopic image with a distant sense, they should be combined into On the C plane. For the near sights A, the left and right eyes are respectively located on the right and left sides of the C plane (a and a1); and the far sight B, the left and right eyes are respectively located on the right and left side of the C plane (b and b1). When observing the C plane image, the left visual image enters the left eye through the cylindrical lens, and the right visual image enters the right eye to form stereoscopic vision. Therefore, the purpose of image processing is to calculate each point in the C plane according to the distance from the eye to the eye. Projection points, and calculate the coordinate position of each depth image after synthesis. The principle for processing pseudo stereoscopic images with multiple depths of field is the same, except that the number of divided images is larger.

It is worth noting that, in addition to considering the calculation of the projection of the close-range and far-away spots on the C-plane, it is also necessary to consider the shielding of the distant objects by close-range objects. If a different occlusion effect of the left and right eyes is calculated when the occlusion is processed, a better three-dimensional effect can be obtained.

Third, based on 3D4U stereo image synthesis

“3D4U PRINTING special visual effect design software” is a professional comprehensive software developed by Medizine, and is the only comprehensive software in the world that has a full set of professional three-dimensional and visual effects production and output. Now it has been widely used and affirmed by stereoscopic printers all over the world. Its image output format has gradually become the standard output format of the stereoscopic printing industry and digital output industry.

1. Produce a true stereo image from a flat image

A stereoscopic image sequence can generally be obtained from a stereoscopic camera system. But the images we usually have are not derived from such a special system, but rather are a single photo taken by a typical camera. Of course, we can divide some scenes by compositing the scenes, and then synthesize them with different aberrations to obtain a sequence diagram. In such a composite map, the depth change of each object itself is ignored, which is reasonable for small objects, but may not be ideal for larger objects. Since the aberration of each object is discontinuous, and it seems that the jump is relatively large, the sense of reality of such a synthesized three-dimensional image is always less perfect than that of a three-dimensional image obtained from a stereoscopic image. We know that a stereo image sequence is generally composed of a plurality of photographs taken in different parallel positions, and each single image does not differ from the image we normally take. If we know the shape of the object we are shooting, we can calculate a stereo sequence image from a simple image under certain conditions. 3D4U provides us with such a tool, a flat-to-stereo sequence image converter. We can generate a stereo sequence image from an image by using some known shapes or similar objects. In the generated stereoscopic sequence diagram, a continuously varying aberration is given, which can be directly stored in a stereo format image file MB3D or an MBF image sequence. The former format is used for raster image editing, and the latter is used for raster image synthesis.

2. Use Photoshop with 3D4U to make a stereogram

1 First make a good picture in Photoshop and save it in *.psd format. Each layer is cut and extracted, the background color of the layer is basically the same as the background color of the actual document, and is saved in the format of *.jpg. This step is to save the use of the mask tool in 3D4U quickly and easily. Hand masking time.

2 Next, call 3D4U to recall the extracted *.jpg image in 3D4U. After each *.jpg image is masked or stereoscopically produced, it must be stored in *.mbf or *.mb3d format. The final stereoscopic image can be rendered in a raster image editor.

3 For which picture uses the *.mbf format and which uses *.mb3d, we need to look at the shape of the picture content. For example, in order to reflect the three-dimensional effect of the beach, it is necessary to use a raster image → tool → stereo image formation adjustment. The stored file format is *.mbf format.

4You can look at the preview. Since the number of images 12 was used in the stereo image formation option, the maximum number of pictures in the rasterizer is adjusted to 12, which becomes a 12-frame animation. You can now feel the stereoscopic effect of the image. Plane characters and flowers can be stored directly in the *.mbf format after the mask, and shift can be used to easily watch the final mask effect.

5 After all the pictures are stored, you can start raster picture editing, open the Raster Image Editor, adjust the size, adjust the unit in File → Preferences → General, load all the pictures, and adjust the picture layers. After loading, aberrations are adjusted and saved after the preview.

6 Make a raster image. Since the maximum number of pictures is 12, we need to render 12 times. Open the file → parameter settings → custom, such as 24 lines imported 2mm grating, the output is calculated as 24 lines × 12 frames = 288 / inch, with 300dpi on it. Click on the vertical raster, make a raster image → adjust the output size, click OK, and so on.

7 Finally, you can generate a preview, and make a GIF animation to see the effect. After the grating board is restored to see the final effect, it can be printed.

Fourth, the conclusion

The above three image synthesis methods are widely used in stereoscopic printing, so they are the core and typical of stereoscopic image synthesis methods. With the development of computer image synthesis technology, stereo image synthesis methods will become more mature, and the three-dimensional printing will have greater development.