增强现实(AR)近眼显示(NED设备的技术瓶颈之一是将虚拟世界和现实世界融合在一起的光学合成器(optical combiner)。
Among the numerous existing technological approaches, diffractive optical waveguides are favored by many leading companies due to their advantages such as wide eyebox, compact size, and good scalability. However, diffractive optical waveguides also face many technical bottlenecks, such as poor color uniformity and low energy efficiency.
1. The principle of diffractive optical waveguides
基于衍射光波导的近眼显示系统的原理如图1所示,该系统由微投影器、衍射光波导和人眼 3 部分组成。
Figure1: Diffractive Optical Waveguide Principle Diagram
The micro-projector is used to generate virtual images and consists of a micro-display chip and an imaging lens group. The micro-display chip is located at the back focal plane of the imaging lens group. Light emitted from any point on the micro-display chip is transformed into a collimated beam after passing through the imaging lens group, which then illuminates the diffractive optical waveguide. In other words, the micro-projector images the virtual image at infinity.
衍射光波导用于将虚拟图像“搬运”至人眼前,并同时起到复制出瞳的作用,它由一块平行度非常好的平板玻璃(即光波导)和在玻璃上制作的多块衍射光栅组成。图1中示例性地画出了两块光栅,分别称为耦入光栅和耦出光栅。当微投影器发出的平行光到达耦入光栅时会被光栅衍射,光栅的衍射级次T+1满足在平板玻璃中的全反射条件,因此会在玻璃中全反射传播,直至遇到耦出光栅。耦出光栅的透射级次T1被耦出玻璃,同时耦出光栅的反射级次R0会继续全反射传播,该过程不断重复,光束多次遇到耦出光栅,每次遇到耦出光栅时都会有一部分光被耦出,这样就起到了复制出瞳的作用。
The collimated light coupled out of the waveguide is focused by the crystalline lens onto the retina, where it forms an image that is perceived by the human eye. Meanwhile, images from the real world can directly enter the eye through the flat glass substrate. Thus, the human eye can simultaneously perceive both virtual images and the real world.
根据上述原理,衍射光波导的一个特点是可以复制入射光束,从而使得人眼可以在较大范围内看到完整的虚拟图像,因此衍射光波导具备眼动范围大的优势。
also,衍射光波导厚度仅由平板玻璃决定,因此足够轻薄,可以充分满足消费者的用户体验。
In terms of mass production, diffractive optical waveguides can ensure sufficient production capacity through nanoimprint technology and can well control costs.
These advantages make diffractive optical waveguides a mainstream augmented reality near-eye display solution.
2. The key parameters and measurement methods of diffractive optical waveguides
2.1field of view
根据衍射光波导的原理,虚拟图像上的某一点被微投影系统转换为某一方向的平行光,平行光经衍射光波导传播、扩展后又被晶状体会聚于视网膜上某一点。由上述过程可知,人眼观察到的虚拟图像上的点对应于空间中一个方向传播的平行光,而平行光的角度范围决定了人眼所见图像的大小,因此通常使用视场角(FOV)这一参数来衡量衍射光波导所显示虚拟图像的大小,视场角越大所显示的虚拟图像越大。
越大的视场角可以提供越丰富的信息,并且可以营造更充分的沉浸感,因此,提升视场角是研究者不懈追求的目标。

Figure2: Schematic diagram of the definition of field of view angle
衍射光波导所呈现的虚拟图像范围通常是矩形,因此至少需要两个参数来确定视场角的大小。常用的视场角表示方法有两种:一种方法是给出横向视场角H和纵向视场角V,如图2a)所示;另一种方法则是给出横纵视场角的比值aa=tanH/2tanV/2)和对角线视场角D,如图2b)所示。
According to simple geometric relationships, the conversion relationship between these two representation methods is

需要强调的是,必须通过两个参数才能完整地确定视场角范围,但很多研究者常常仅关注对角线视场角D,而忽略了横纵视场角的比值a。事实上对衍射光波导而言,即便对角线视场角都是 40°,不同的所导致的设计难度也是完全不一样的。
视场角的测量方法如图3所示,在人眼观察的位置放置校正好像差的测量镜组,不同视场角的平行光会聚在接收屏上不同的点,通过测量在接收屏上所成像的大小即可计算得到视场角的大小。

Figure3: Schematic diagram of measurement method of field of view angle
Figure3中测量镜组的焦距为 f,所成的像半宽为l,则对应的水平视场角大小为

需要强调的是,根据图3所测量得到的是实际视场角的大小,然而人眼所感知到的虚拟图像大小除与实际视场角大小相关外,还会受到图像清晰度、均匀性、模组的眼动范围等诸多因素的影响。例如对于两个真实视场角完全相同的近眼显示模组,眼动范围较大的模组会给使用者带来视场角更“大”的错觉,但目前这些因素对视场角的影响尚无定量研究。另外,模组对外界环境的遮挡程度也会影响使用者对视场角大小的判断。
Therefore, when designing a diffraction optical waveguide near-eye display module, we should not simply pursue the single parameter of field of view, but should ensure a balance between key parameters to avoid that the field of view perceived by the human eye does not reach Design expectations.
2.2 Eye movement range
The eye movement range is one of the important indicators of the near-eye display module. It refers to the three-dimensional area between the display module and the human eye, in which the human eye can observe virtual images that meet certain imaging standards.

Figure 4 Schematic diagram of the eye movement range of diffracted light waveguides
Figure 4展示了衍射光波导的眼动范围。微投影器发出的不同方向的平行光束构成了一个光棱台,这个光棱台的边界由微投影器的视场角和出瞳大小决定。当光束被衍射光波导耦出后,所有被耦出光束在空间中重叠的区域也会构成一个光棱台,人眼只有位于这个光棱台内才有可能接收到所有视场角的光线,否则就会损失某些角度的信息,因此这个光棱台内的区域就是衍射光波导的理论眼动范围。
需要强调的是,眼动范围是一个三维区域,但由于模组通常存在对称性,很多近眼显示模组仅使用适眼距离r和该适眼距离下人眼观察区域的大小Bx×By来描述它,如图4中虚线框区域所示。

Figure5: Schematic diagram of the eye movement range of a near-eye display module with a certain included angle
然而,出于人们对模组的舒适性提出的更高要求,衍射光波导常常具有更复杂的姿态。如图5所示,微投影器被组装在眼镜腿支架中,为了佩戴舒适,眼镜镜片与竖直方向存在一定的夹角θ,在这种情况下如果仅用适眼距离和人眼观察区域大小就不能完整描述眼动范围,还需要明确观察平面与衍射光波导的相对姿态,这一点常常被忽视。
In addition, the concept of eye movement range has a certain degree of ambiguity. There is currently no standard definition. The biggest controversy is how to determine the boundary of the eye movement range. In particular, the imaging uniformity of the diffractive optical waveguide is not good. In this case, how to determine the boundary of the eye movement range? Reasonably determining the imaging standards required for human eye observation is even more worthy of discussion.
通常确定眼动范围边界的方法是:确定某一参数A(如亮度或均匀性指标)作为评价指标,在确定的适眼距离r0下测量一定范围Bx0×By0内参数A的值,记该范围内参数A的最大测量值为Amax;设定某一阈值β,当虚拟图像的参数A变化为βAmax时判定此位置为适眼距离r0下的眼动范围边界;如果要确定三维区域的眼动范围边界,则需要在不同适眼距离下测量参数A的值,并使用同样的判定标准确定眼动范围的边界。
需要注意的是,评价参数A的选择会影响眼动范围的大小。
参数A为虚拟图像的亮度,这种判据简单直观,适用于对亮度要求比较高的使用场合,例如户外信息提示;
参数A为均匀性指标,这种方法计算相对复杂,更适用于对图像质量要求较高的场合,例如休闲娱乐;
参数A为调制传递函数(MTF),但这种参数更适合于基于折反射的几何光学近眼显示方案,因为衍射光波导的原理决定了不同位置的MTF不会有较大区别。
In fact, the various parameter indicators of the diffractive optical waveguide change slowly at different observation positions, and there is no clear boundary in the area where the human eye can see the complete virtual image. Therefore, it is necessary to clearly evaluate the eye movement range when measuring the size of the eye movement range. standards, and reasonable evaluation parameters must be determined based on the user's usage scenarios.
2.3 Brightness, brightness uniformity and energy utilization
The brightness of the augmented reality near-eye display module is a very important indicator. It determines the brightness and darkness of the virtual image seen by the human eye. Insufficient brightness will make the module unable to be used under outdoor strong light conditions, so the brightness of the diffraction light waveguide is increased. is very important.
also,衍射光波导的成像亮度并不均匀,而是一个随角度变化的分布,因此,成像亮度的均匀性也是一个重要的指标,它决定了人眼所观察图像的舒适度,并且影响了人眼所感知到的视场角、眼动范围等指标。
In addition to brightness, near-eye display modules also need to focus on energy utilization. Improving energy utilization can reduce module power consumption and extend battery life.
Brightness, brightness uniformity and energy utilization are interrelated.
根据光度学原理,一般使用光亮度L来评价所见物体的亮暗程度,即给定方向上单位面积、单位立体角内的光通量,单位是cd/m2,也称为尼特(nits)。
一般微投影系统更多使用光通量 Φ 来描述其提供的总光能,单位是lm

Figure6: Schematic diagram of the measurement method of diffracted light waveguide brightness, chromaticity and microprojector flux
入眼光亮度和光通量的测量方法如图 6所示。
测量入眼亮度的光路如图 6a)所示,在适眼距离下放置亮度计,亮度计由仿人眼镜头、光电探测器和光谱仪组成。仿人眼镜头是一个光圈位于镜头表面的成像镜头,它的光学性能与人眼相近,可以模拟人眼的视觉特性。光电探测器可以接收镜头所成的像,并将中心一定视场的光输入到光谱仪中,根据光谱仪和光电探测器的成像结果即可得到全视场的亮度和色度信息。
测量微投影器的光路如图 6b)所示,微投影器所发出的光被积分球接收,经光电探测器处理即可得到光通量。

Figure7: Schematic diagram of the method for evaluating the eye brightness of diffracted optical waveguides
如图7所示,将人眼所观测到的虚拟图像平均划分为9个子区域(如果亮度变化很剧烈也可以划分为更多的子区域),测量每个区域的中心点亮度 L1 ~L9,则虚拟图像的平均亮度可定义为这9个点亮度的平均值。为了更全面地反映视场边缘的亮度情况,还可以再测量视场4个顶点的光亮度L10~L13,用 L1~L1313个点的亮度来衡量平均亮度。

Figure8: The impact of different brightness distributions on the human eye
然而对亮度的评估也应充分考虑人眼特性和使用场景,如图8所示,两幅纯绿色的图片9点平均亮度基本相同,区别在于图8a)视场中心亮度高、边缘亮度低,而图8b)则在视场左侧存在明显的暗区。人眼会感觉图8a)比图8b)要亮,这是因为人眼对中心视场的亮度信息比较敏感,更容易将中心视场偏亮的图片感知为平均亮度更高。因此在评估亮度时应充分考虑人眼特性,例如给不同点的亮度赋予不同的权重,中心点权重比边缘点高,从而得到符合人眼观感的亮度测量值。
The expression for the brightness measurement value is

关于亮度均匀性,可以用L1 ~L13的标准差除以平均值L——As a measure, that is

上式中之所以要除以平均值,是为了排除亮度的绝对值对均匀性的影响,U1越接近于0,代表亮度均匀性越好。虽然U1充分利用了各测量点信息,但并不直观,另一种更直接而简单的评价方式是使用L1 ~L13中的最小值比最大值作为评价指标,即

U2的取值范围是[01],越接近1代表均匀性越好。U2取值范围明确,相较于U1,物理含义更容易被理解。

Figure9Schematic diagram of light energy utilization evaluation index of diffraction optical waveguide
关于能量利用率,研究者通常使用入眼亮度L与微投影器入射通量Φ的比值ηD =L/Φ来评价衍射光波导的光能利用率,单位是cd·m−2 ·lm−1,图9给出了能量利用率的评价方式。使用该单位的好处是可以快速得到入眼亮度,从而让工程技术人员迅速判断能否在适合的场景下使用。例如衍射光波导的能量利用率是100 cd·m−2 ·lm−1,当搭配一个10lm的微投影器时,最高入眼亮度为1000 cd/m2 ,这个亮度可以满足大部分室内的使用场景,但在户外强光下使用则会稍显不足。
如果要评价整个模组的能量利用率,就需要测量微投影器的发光效率 γ,即每单位功耗所产生的光通量,单位是lm/W,则模组的能量利用率为ηM =γηD
需要说明的是,以入眼亮度评价能量利用率时忽略了眼动范围内其他位置的能量。事实上因为衍射光波导眼动范围大,能量被分散在眼动范围内,因此很多位置的光并没有进入人眼,如图9虚线以外的光线,由此导致上述测量方法所得到的衍射光波导能量利用率不高。因此在选择近眼显示装置时,需要权衡眼动范围和能量利用率;评价模组时,也应该给出在什么样的眼动范围条件下的能量利用率。
2.4 Chroma Uniformity
The brightness uniformity of diffraction light waveguides usually needs to be improved. When displaying a white image, the brightness distribution of different colors may be completely different, which will cause obvious chromaticity unevenness in the image, that is, the image observed by the human eye has More obvious color difference.
根据色度学原理,任意的颜色可以使用色空间中的一个点的坐标来描述,推荐使用 CIE1976 L* u * v * 均匀色空间来描述某一点的颜色,这种色空间的优点是色空间中两点的距离能够反映人眼对两个颜色的知觉差异大小,即两种颜色在色空间中的距离越远,人眼所感知到的颜色差异越大。
色度的测量方法与亮度相同,都如图6a)所示。
与评价亮度类似,可以将人眼所见虚拟图像分割为N个小区域,记每个小区域内中心点的色度值为(ui ' vi '),则色度均匀性的衡量指标可以定义为

即用所有测量点在色度图上最远的距离来衡量色度不均匀性,如图 10 所示。

Figure10Schematic diagram of the chromaticity non-uniformity measurement method of diffracted light waveguides
Figure10中给出了虚拟图像上13个点在色空间中的位置(这13个点在虚拟图像上的位置如图7所示),其中相距最远的两个点(已标红)之间的距离即为色度均匀性的衡量指标Δu'v'
Due to the dispersion characteristics of gratings, diffracted optical waveguides are prone to chromaticity unevenness. But once the grating is made, the chromaticity non-uniformity is certain and can therefore be compensated for through calibration and software correction.
2.5 Grating stray light and contrast
For diffractive optical waveguides, the stray light level of the grating is an important but often overlooked parameter.
The so-called stray light of grating refers to the stray light that exists in directions other than the grating diffraction direction (referring to the direction determined by the grating equation). It mainly comes from high-frequency errors of the grating such as surface roughness and grating line bending.

Figure11Schematic diagram of diffraction grating line bending and stray light
Figure11对比了不同的光栅微观形貌及所造成的杂散光。图11a)和图11c)是较为理想的光栅槽形和对应的衍射光点,图11a)所展示的光栅栅线笔直,表面粗糙度低,因此衍射光点能量集中,非衍射级次方向上不存在杂光。而图11b)所示的光栅存在明显的栅线弯曲,因此对应的衍射光点能量分散,非衍射级次方向上存在大量杂光。
The stray light of the grating will make the originally dark areas brighter, that is, there will be a larger background noise, causing the image contrast (the ratio of the brightest to the darkest brightness) to decrease.

Figure12Schematic diagram of imaging effects under different contrast ratios
Figure12a)和图12c)、图12b)和图12d)分别是低对比度和高对比度的成像效果,可以看出对比度降低会使观察者感觉到图像细节不清晰、颜色不够艳丽,进而降低用户的使用体验。
Therefore, attention should be paid to the stray light level of the grating, which is of great significance to improving the process level to enhance the contrast of the diffractive optical waveguide.

Figure13Measurement method of grating stray light
光栅杂散光的测量方法如图13所示,一束激光入射到待测光栅上,经过一个特殊的透光片到达白纸屏上。该透光片中心区域镀有不透光的金属膜,该区域大小恰巧可以将透射级次遮挡而不影响其余杂散光透过。使用CCD记录白纸屏上一定范围内的图像灰度值,当CCD的参数固定时,灰度值即可反映光栅的杂散光水平。
衍射光波导对比度的测量方法是使用一张4×4的黑白棋方格,如图12a)所示,用所有白色区域点的亮度平均值Lw与所有黑色区域点的亮度平均值Lb的比值C=Lw/Lb来衡量对比度。
 
Reference: "Key Parameters of Diffracted Light Waveguide Augmented Reality Near-Eye Display"_Progress in Lasers and Optoelectronics》
Source:小小光08 

艾邦建有AR/VR产业链微信群,目前有HTC、PICO、OPPO、亮亮视野、光粒科技、影创、创维、佳视、歌尔、立讯精密、多哚(纳立多)、欣旺达、耐德佳,联创电子、至格科技、灵犀微光、舜宇光学、广景视睿、珑璟光电、京东方、海信视像、科煦智能、阿科玛、金发科技、思立可、新安天玉、四方超轻、大族激光、发那科、承熹机电等加入,也欢迎大家长按下方图片识别二维码加入微信群:
Document Download
Welcome to join by clicking hereAR/VR DirectoryCurrently, there are more than 3000 members, including companies such as GoerTek, HTC, OPPO, Skyworth, PICO, ByteDance, Black Shark, Lenovo, Nreal, Lynx, Luxon, Lingxi MicroLight, Luxshare Precision, Leading Ideal, OFILM, Huaqin, Wentai, Luxshare, Lumus, Sisvel, and Shunyu, among others. Click on the keywords below for filtering.

By d

en_USEnglish