Naked eye 3D display method and device based on multilayer transparent liquid crystal screen

Abstract

The invention discloses a naked eye 3D display method and device based on a multilayer transparent liquid crystal screen. The device comprises multiple layers of transparent liquid crystal screens. The method comprises the steps: monocular image depth map estimation and the depth-to-level conversion. The monocular image depth map estimation comprises the following steps: calculating depth information of part of video frames; performing association calculation to obtain a video frame with depth information, performing depth-to-level conversion according to the depth information in combination with consideration of distances among the multiple layers of liquid crystal screens and light attenuation among the screens, and correspondingly distributing object contents of different depths in theimage to each liquid crystal screen of the multiple layers of transparent liquid crystal screens for display, thereby realizing naked-eye 3D display. Based on the classical SFS algorithm principle, the calculation process of the method is optimized, the calculation speed of the method is increased, the calculated depth information is cut according to the characteristics of the liquid crystal multi-layer screen, depth calculation basically capable of calculating the image content in real time is achieved, and the requirements of engineering application are met.

Description

technical field [0001] The invention belongs to the technical field of liquid crystal screens, relates to naked-eye 3D display, and relates to a naked-eye 3D display method and device based on a multilayer transparent liquid crystal screen. Background technique [0002] At present, the mainstream naked-eye 3D technologies include: slit liquid crystal grating, lenticular lens, pointing light source, and active backlight. [0003] [1] Slit-type liquid crystal grating. The principle of this technology is that after adding a slit grating in front of the screen, when the image that should be seen by the left eye is displayed on the LCD screen, the opaque stripes will block the right eye; similarly, the image that should be seen by the right eye will be displayed When on an LCD screen, the opaque stripes shade the left eye, allowing viewers to see 3D images by separating the left and right eye views. [0004] [2] Lenticular lens, this technical principle is to project the pixels...

AI Technical Summary

Problems solved by technology

[0008] In the prior art, "3D Display System and Algorithm Design Based on Liquid Crystal Multilayer Screen" (Chen Xiaoxi, Li Guangyong, Wang Jimin, Jiang Shan, Huang Zhangying, Liu Chenxi. "Liquid Crystal and Display", 2017, 32(4): 302-307. ) proposed the concept of the liquid crystal multi-layer screen, but in the design of the display algorithm, only the attenuation of light after multi-layer transmission is studied and provided, how to display the distance and layering of the required content on the multi-layer LCD screen did not mention

The invention patent "A Virtual Reality Device and Corresponding Display Method" (CN201611264973.1) also proposed the concept of liquid crystal multi-layer screen, but in the algorithm design, it did not mention how to obtain the depth information of the image to be displayed, and mapped to each layer

In academia, although there are many research schemes for depth acquisition based on a single image, such as the SFS (Structure Restoration from Shadow) scheme proposed by Horn in 1970, this method uses the brightness and shadow of the image under different lighting conditions to extract Calculating the depth information of the surface of an object is one of the important means to obtain 3D data in computer vision. This type of algorithm is relatively mature and its content can already obtain more accurate depth information, but its calculation amount is relatively large, and it is difficult to match the normal LCD screen. The computing power of the band, and there is a certain amount of noise

With the development of machine learning technology, there have been new processing solutions recently, such as image depth acquisition based on deep learning, but the calculation amount of such solutions is unimaginable for the computing power of normal LCD screens.

Images