а) Experimental setup
Hologram recording was done on an optical setup schematically shown in Fig. 4, which includes galvo scanner, a spatial light modulator, and a half-wave plate.
The light beam from the Pharos laser system (1) passes through the dynamic beam expander (3) to cover the whole area of the SLM matrix (5). By adjusting the polarization of the incident light with a half-wave plate (2) we achieve the amplitude modulation regime of the spatial light modulator. The expanded beam propagates to the light-splitting cube (4) and reflects from the SLM matrix. The reflected beam travels through the system of mirrors (6) and enters the galvo scanner (7), which uses an F-Theta lens and (8) to scan the sample (9). With a six-axis positioner (10) we provide focusing during sample processing. The control of all devices performed using a PC (11).
The setup allows implementing two approaches for recording quasi-holograms on the different substrates. The first method requires static polarization of a laser beam. Separate parts of the holographic image are recorded sequentially, gradually changing the polarization with the half-wave plate after scan. The second approach involves a dynamic change of the laser beam polarization using SLM during a single scan. The software developed by the authors provides control and synchronization of all devices. The described means allow recording quasi-holograms using both bitmap and vector images. A detailed description of all recording approaches is discussed below.
b) Method of recording holograms of vector images
The easiest method to implement and the least time-consuming one is that with static beam polarisation, which is suitable for segmented vector images. The hologram is recorded by scanning a substrate with different light polarization acquiring a different spatial orientation of LIPSS for each segment of the image. An example of a vector image used to record a hologram in this work is shown in Fig.5.
By modifying the LIPSS orientation for each segment of the hexagon one can achieve the pseudo volume effect. More complex images require reducing the size of individual segments. Hence, when processing at high speed and pulse repetition rate one cannot avoid parasitic transition lines between segments. Recording a high-quality hologram in a relatively short period of time requires a high-speed shutter, such as an acousto-optic modulator, with a switching rate of 2-10 μs [31].
c) Method of recording holograms of bitmap images
Another approach was used to record the hologram using a bitmap image. The main difference is that the processing is carried out continuously. The polarization of the incident laser beam is switched dynamically by the spatial light modulator during the scanning. Thus, one can obtain different spatial orientations of LIPSS by a single scan. This approach does not require high-speed shutters to prevent parasitic lines. The main drawback is the slow processing speed, due to the low frame rate of the SLM. For recording a single image row, the signal in the SLM should switch M times for the image with dimension M×N. When recording a hologram with a size of 10×10 mm and dimension 500 × 500 pixels, the maximum possible processing time can be 10 mm/s at a frame rate 500 Hz.
Since the spatial orientation of LIPSS is controlled by SLM, it is necessary to determine the dependence of the polarization angle variation on the grayscale level. In this work, the SLM was the LCoS matrix, which we extracted from the budget projector LG PH 150G with a frame rate of 59 Hz. In order to study the appropriate dependence, we used a simple optical scheme with a half-wave plate and an analyzer (see Fig. 6).
The total transmission of this scheme is equal to:

where θ denotes the angle between the laser polarization after the half-wave plate and the analyzer, φ represents the angle between the polarization and the optical axis of the liquid crystal, δ stands for polarization retardation related to voltage applied.
Obviously, when θ = 90º and φ = 45º the transmission is equal to:

Having obtained the experimental dependence of the intensity variation at each grayscale level and knowing the input intensity, we calculated the dependence of the polarization angle rotation for our LcOS matrix.
As can be seen from Fig.7. the obtained dependence is quite nonlinear. Within the range from 0 to 100 variation of the polarization angle is not being observed. For our purpose, optimal is a linear area in the range of 130-255. Obviously, the image greyscale level range must be adjusted before recording according to the operating range. The whole recording process is described in Fig.8.
First, the input image (1) is converted to a grayscale color model with adjusting graystone range according to the linear range of the SLM. During laser scanning, the voltage of the whole SLM matrix is dependent on the pixel grey level (2) and switches with a certain period of time, providing control of the polarization of the incident beam, and the spatial orientation of LIPSS (3). By scanning row by row, we obtain a quasi-hologram of the bitmap image (4).