数字光处理
光学
镜头(地质)
计算机科学
发光二极管
空间光调制器
光传递函数
人工智能
投影(关系代数)
物理
投影机
算法
作者
Haonan Jiang,Zibin Lin,Yao Li,Yinguo Yan,Ziping Zhou,Enguo Chen,Qun Yan,Tao Guo
出处
期刊:Applied Optics
[The Optical Society]
日期:2021-08-09
卷期号:60 (23): 6971-6971
被引量:8
摘要
Digital light processing (DLP) is currently a cutting-edge technology for desktop projection optical engines. Due to the passive luminescence characteristics, the DLP projection engine needs a few specific illumination optical components for light collimation, homogenization, and color combination, together with a projection lens matching the DLP chip and magnifying the image. In this paper, we propose a design approach that first splits the DLP projection optical engine into individual components for separate design, and then integrates them into a whole system for further verification. For the first step, the collimating lens group is designed for light collection, and the dichroic mirrors are used to fold the light path based on tri-color LED light sources. For the second step, a fly-eye lens and the corresponding relay lens group are designed to achieve uniform illumination on the DMD chip. The third step is to optimize the projection lens group for high-resolution projection display. Based on the design and simulation, the optical efficiency is 63.4% and the uniformity reaches 94.9% on the projection screen. The modulation transfer function (MTF) of the projection lens is higher than 0.4 at 66 lines for the distance of 500 ∼ 1500 m m , and the distortion is lower than 1%. Simulation results show that the total luminous flux is estimated to reach 224.15 lm when the powers of tri-color LEDs are 21 W, 15.5 W, and 25 W, respectively. A projector prototype is built and tested for further verification, which provides a luminous flux of 220.43 lm and uniformity of 90.22%, respectively. The proposed design, demonstrated by both simulation and experiment, exhibits high feasibility and application potential in state-of-the-art commercial projector design.
科研通智能强力驱动
Strongly Powered by AbleSci AI