† Corresponding author. E-mail:
Project supported by the National Basic Research Program of China (Grant No. 2013CB328803) and the National High Technology Research and Development Program of China (Grant Nos. 2013AA013904 and 2015AA016301).
A phase-only computer-generated holography (CGH) calculation method for stereoscopic holography is proposed in this paper. The two-dimensional (2D) perspective projection views of the three-dimensional (3D) object are generated by the computer graphics rendering techniques. Based on these views, a phase-only hologram is calculated by using the Gerchberg–Saxton (GS) iterative algorithm. Comparing with the non-iterative algorithm in the conventional stereoscopic holography, the proposed method improves the holographic image quality, especially for the phase-only hologram encoded from the complex distribution. Both simulation and optical experiment results demonstrate that our proposed method can give higher quality reconstruction comparing with the traditional method.
The concept of computer-generated holography (CGH) was first proposed by Kozman and Kelly.[1] With the development of computer technology, there are three categories of methods developed for calculating CGH, which are the point source algorithm, the polygon-based method, and the stereoscopic hologram algorithm.[2]
As we know, a three-dimensional (3D) scene can be precisely described by a number of point sources. Based on this characteristic, Slinger and Cameron proposed an algorithm which divided the 3D object into multiple point sources, and the light field generated by each point source was superimposed to form the whole light field of the 3D scene.[3–5] However, for a complex 3D object, the huge number of point sources increased the computational complexity, so the calculation efficiency was very low.
3D objects can also be divided into multiple planar segments[6,7] Detlef Leseberg proposed a method to calculate a rotary transform for the complex amplitude in a Fourier domain, and the complex distribution of hologram was calculated by superimposing the diffraction waves from the inclined planes.[8,9] But he did not give the complete rotation transform formula and results. Based on the angular spectrum theory and the coordinate rotation in Fourier domain, Kyoji Matsushima proposed a new method to calculate the transmission processes of a light field.[10] The method applied a twice Fourier transform and a spectrum interpolation for each segment to obtain a hologram with a complex distribution. However, similar to a point source algorithm, a 3D object often has a large number of planar segments, thus real-time calculation becomes a big problem. Lukas Ahrenberg proposed a solution for this problem.[11] Based on the angular spectrum theory, Ahrenberg directly calculated the light field distribution of the hologram by using only once fast Fourier transform (FFT) in the frequency domain, thus avoiding FFT computation for every segment. However, the shading and occlusion effect still remains a problem for this method. Hwi Kim solved the occlusion problem by the ray casting and ray tracing techniques.[12] Nonetheless, the heavy computational loads of the ray-object processes brought in the problem of calculation speed again.
In this paper, we focus on the stereoscopic hologram algorithm. King et al.[1] proposed to use a computer to calculate and store the perspective projection images of the 3D subject from several viewpoints and then recorded them on microfilms. The hologram can be generated through the interference of the reference beam and the light field of each projection. Based on this method, Toyohiko Yatagai proposed a new way to reconstruct 3D scenes.[13] After getting the perspective projection images of the 3D subject, the fast Fourier transform of each projection would be calculated to generate the corresponding sub-hologram by a computer. Then all sub-holograms would be tiled as a whole hologram according to the viewpoint order (i.e. stereoscopic hologram). According to the process of stereoscopic holography, the computational complexity is related to the number of viewpoints and the resolution of the sub-hologram, so the calculation speed could be promoted for real-time implementation. Computer graphics rendering techniques can be applied in stereogram computations to add multiple shading effects and this will eventually make the 3D scene more realistic, so the occlusion problem can be automatically solved during the rendering processes.[14]
As the diffraction efficiency of the phase-only spatial light modulator (SLM) is very high, which is 100% in theory, it is widely used for holographic display.[15–17] Since it can only modulate the phase of the incident light, if we merely take the phase distribution of the hologram, the loss of amplitude information would bring errors in the reconstruction.[18,19] Our goal in this paper is to realize a method that can obtain the phase-only stereoscopic hologram with accurate reconstruction of 3D scenes.
The Gerchberg–Saxton (GS) iterative algorithm is a phase retrieval algorithm, which was first proposed by Gerschberg and Saxton in 1972.[20] It has been widely used in CGH because of the high accuracy of the reconstructed phase information. But it can only be used for two-dimensional (or one-dimensional) signals, so it is commonly applied in the projection of 2D images or 3D scenes consisting of multiple 2D planes which introduces a propagation process between each plane and still leaves the shading and occlusion a big problem.[21–23] Our proposed method takes the GS iterative algorithm instead of the non-iterative algorithm in the stereoscopic holography. On one hand, it can achieve a phase-only hologram with high precision; on the other hand, it can solve the problems of computing speed and shading or occlusion effects, and directly record the light field information of the whole 3D scenes, avoiding the process between 2D planes.
There are several methods to describe the light field distribution of a 3D scene. One is to get the 2D intensity and depth image as a description of the 3D scene, and another is to use a light-field camera, or camera array or micro lens array and imaging sensor to acquire 2D parallax views of the 3D scene, which is always employed in stereoscopic holography. In our proposed method, a micro lens array is used to obtain the 2D perspective projection views (sub-images) of the 3D object, which is based on perspective geometry projection.[24] These views are then recorded on the imaging sensor, as shown in Fig.
To satisfy the perspective projection geometry, d1 is usually far greater than fm, which means d2≈fm.
After getting the sub-images of the 3D object, the hogel is calculated with respect to each sub-image by using the GS iterative algorithm, which is described in detail in the dotted box of Fig.
After all of the hogels are calculated for the corresponding sub-images, they are eventually synthesized to a whole phase-only hologram according to the viewpoint order.
In our experiment, the sub-images are generated by computer graphics rendering software, such as 3Dmax. The letters C and D are set up at the distances ZC = 200 mm and ZD = 300 mm from the micro lens array, respectively. The focal length of the micro lens array is fm = 3.3 mm and the diameter is d = 0.4 mm. The resolution of each sub-image is 120×120, and the number of the viewpoints is 16×16, which means that the final generated phase-only hologram contains 1920×1920 pixels.
As shown in Fig.
On the other hand, in the spatial domain, according to the theory of geometric similarity, we can figure out
Let N be the pixel number of the sub-image along the corresponding coordinate axis, the pixel interval Δp on the sub-image is
The simulation results of Fresnel reconstruction are shown in Fig.
The experiments of complex models, such as a teapot and a 3D scene which contains a teapot located at Zt = 200 mm and a background wall located at Zw = 300 mm, are also carried out and the results are shown in Figs.
The light path diagram is presented in Fig.
Figures
We propose a method to calculate phase-only CGH by using the GS iterative algorithm in stereoscopic holography. In contrast from the simple FFT processing of the traditional stereoscopic hologram calculation, the GS iterative algorithm is applied to the 2D perspective projection views of the 3D object (sub-images) to generate a phase-only hologram, which reduces the errors caused by missing the amplitude information so that the image information would be more complete. We also give the formula to calculate the sampling rate of the holographic plane. The simulated and optical experiments are conducted respectively. The results indicate that there is a significant improvement in the quality of the reconstructed image by our proposed method.
1 | |
2 | |
3 | |
4 | |
5 | |
6 | |
7 | |
8 | |
9 | |
10 | |
11 | |
12 | |
13 | |
14 | |
15 | |
16 | |
17 | |
18 | |
19 | |
20 | |
21 | |
22 | |
23 | |
24 | |
25 |