扩展角锥波前探测器动态范围的研究

发布时间：2018-01-20 01:54

本文关键词： 自适应光学液晶波前校正器大气湍流角锥波前探测器　出处：《中国科学院长春光学精密机械与物理研究所》2017年博士论文　论文类型：学位论文

【摘要】：众所周知,自适应光学系统可以对望远镜接收的动态畸变波前进行实时探测校正,恢复其自身的光学衍射极限成像分辨率。由于天文、军事等领域的发展趋势,要求望远镜有能力对暗弱目标进行高清晰度成像。传统的哈特曼波前探测器将入射光分为数十甚至数百个子光束的探测原理导致其信噪比低,无论望远镜口径多大,哈特曼探测器的极限探测星等最高为5.5星等,无法探测暗弱目标。近年开始尝试应用的角锥波前探测器对暗弱目标的探测能力要强于哈特曼波前探测器。角锥探测器是将入射光束会聚在角锥棱镜顶点,使光束分为4个子光束形成4个探测器入瞳孔径的像,波前畸变使角锥锥尖光斑亮度分布发生变化或整体偏移,从而使4个光瞳像能量分布不均,光瞳像内部亮度分布也不均,根据4个光瞳像对应像素上的亮度归一化差异可以计算得到与波前局部斜率正相关的响应信号,进而解算畸变波前。然而角锥探测器动态范围较小,当波前畸变较大,尤其是倾斜量较大时,光斑在角锥锥尖上的偏移量过大,造成只出现一个光瞳像而其他三个光瞳像亮度过低,或光瞳像局部亮度过低,导致无法解算波前。由于角锥波前探测器的动态范围小,目前只能应用于闭环控制系统,尚未有报道使用角锥波前探测器的工程化开环自适应光学系统。而开环控制模式是自适应光学的发展趋势,其中的波前探测器直接接收望远镜的入射波前,要求波前探测器动态范围要大、探测精度更高。针对角锥探测器动态范围小的问题,考虑将光斑在角锥锥尖上的偏移量不同转变为光斑在四锥面上曝光时间不同来定量减小四个光瞳像的亮度差异,扩展动态范围。设计了快反镜调制棱镜锥尖上的光斑沿圆轨迹运动使4个光瞳像依次出现;由于入射波前倾斜导致圆轨迹中心偏离棱镜锥尖,光斑在各锥面上的运动时间不同,各光瞳像的曝光时间也不同,从而调制其亮度,使原本亮度过低的光瞳像亮度提高至可被探测,且光瞳像亮度差异与畸变引起的圆轨迹中心偏移成正比,原来的波前重构方法仍然适用。上述调制型角锥探测器的动态范围理论上可从原来的1λ扩大至12λ(λ=780nm),满足实际大气湍流的探测需求。但调制型角锥探测器中又产生了如下问题:光斑经过角锥棱边时,由于光瞳像光场叠加引入四条亮带,造成重构波前的高阶误差;亮带亮度值是背底噪声的10倍,因此无法用阈值法直接扣除。另外,光瞳像因调制发生的位置偏移不能超过3像素,否则光瞳像边缘发生弥散,无法精确定位,而任一光瞳像定位不准确都会导致对应像素的错位误差。同时,由于棱镜引入的轴向色差以及光瞳孔径边缘衍射形成的光晕,也会使光瞳像边缘形成几个到十几个像素的弥散,产生光瞳像定位误差;另外,大像差输入使原响应矩阵计算过程中的小像差近似条件不再适用,同时调制过程导致光瞳像亮度差异与波前畸变对应关系发生变化,需要重新计算调制型角锥探测器的响应矩阵。根据zemax软件仿真设计结果,以光瞳像在对角方向内外边缘亮度值差异为判据,对角锥探测器进行了精密装调,获得了边缘最锐利、内外边缘点共线的光瞳像,光瞳像对角方向内外边缘亮度值差异为7.84%,调制过程引起的光瞳像振动幅度小于1像素。针对光场叠加引入的亮带,分析了光瞳间距不同时亮带对光瞳像内部亮度分布的影响,设计并调节光瞳像以3.8倍半径的中心距等距分布,此时可以忽略光瞳像间亮带对探测信号的影响。使用双角锥棱镜替代单角锥棱镜,消除了单棱镜色散效应以及色散导致的光瞳像边缘模糊,同时降低了棱镜加工难度。分析了光瞳像边缘亮度分布特征,发现对光瞳像边缘亮度极大值点定位并不能对光瞳像边缘精确定位。提出对光瞳像边缘亮度梯度极大值进行定位,定位精度达到像素级别;然后利用相关算法二次标定光瞳像中心坐标,令高像素密度的液晶校正器产生包含10个精细同心圆的靶图,此时光瞳像内部亮度分布也为同心圆,计算光瞳像中心坐标在粗定位坐标附近振荡时,对应的光瞳像同心靶图的互相关度,对光瞳像中心二次定位,精度提高至亚像素级别。搭建实验室静态像差探测校正平台,在系统光瞳与光瞳像严格共轭的前提下,光瞳像精确提取使开环探测误差相较光瞳像粗定位系统减小60%,闭环校正收敛速度加快,闭环校正残差减小15%。分析发现调制型角锥探测器响应矩阵相当于在无调制探测信号上叠加了由倾斜调制引入的调制多项式,其物理意义相当于调制导致的畸变响应信号幅度衰减。大像差输入情况下,响应矩阵中包含的三角函数项会发生高频振荡,导致无法精确计算;通过坐标归一化,使衰减项中三角函数自变量阈值变小,抑制了衰减项振荡。对响应信号本身包含的三角函数项,由于其自变量显含Zernike多项式,使用小系数Zernike基底抑制其在大像差输入情况下的振荡,并在重构过程中施加增益以保证响应矩阵幅值。计算了输入像差为10λ时,调制型角锥探测器对35项Zernike模式的响应矩阵。使用角锥探测器和Zygo干涉仪搭建了探测器精度验证光路,在精密装调和光瞳像精确定位的基础上,对不同幅度的包含35项Zernike模式的随机静态像差进行探测。探测结果显示,使用开环响应矩阵与传统响应矩阵相比,探测器探测误差RMS值减小50%,对不同幅度像差的探测误差RMS值均小于λ/20,基本满足开环系统探测需求。本论文获得的结论证明,通过扩展角锥波前探测器动态范围的研究,实现了探测动态范围满足实际大气湍流探测需求的角锥波前探测器。同时,该探测器具备在自适应光学系统中对大像差输入的高精度探测能力,探测误差RMS值小于λ/20。该研究将角锥波前探测器的应用范围从闭环系统扩展至开环系统,对提高自适应光学系统极限探测星等,实现对暗弱目标的高精度成像起到积极的推进作用。
[Abstract]:As everyone knows, the adaptive optics system on dynamic wavefront telescope received real-time detection correction, recovery of optical diffraction limit imaging resolution of its own. The development trend of astronomy, military and other fields, have the ability to request telescope high-resolution imaging of faint objects. Hartmann wavefront sensor divides the incident light into the traditional detection principle of dozens or even hundreds of sub beams due to the low signal-to-noise ratio, no matter how large aperture telescope limit Hartmann detector is the highest detectable magnitude 5.5 magnitude, to detect faint objects. In recent years, began to try to apply the pyramid wavefront detector faint target detection ability is stronger than Hartmann. Pyramid wavefront detector detector is an incoming light beam converged at the prism vertex so, the beam is divided into 4 sub beams form 4 detectors like the pupil aperture, the wavefront distortion Pyramid tip spot brightness distribution changes or the whole migration, thus making the 4 pupil like uneven distribution of energy within the pupil image brightness distribution is not uniform, according to the 4 pupil image brightness normalized difference corresponding pixel can be calculated positive correlation with wavefront local slope of the response signal, then calculates the wavefront aberration. However pyramid detector dynamic range is small, when the wavefront distortion, especially tilt when a large amount of spots in the pyramid cone offset on the tip is too large, resulting in only one pupil like the other three pupil image brightness is too low, or pupil like local brightness is too low, leading to the solution due to the dynamic range of wavefront. Pyramid wavefront detector, at present can only be applied to the closed-loop control system, have not been reported using pyramid wavefront detector engineering open-loop adaptive optics system. The open-loop control mode is adaptive optics The development trend of the incident wavefront wavefront detector which receives the telescope, wavefront detector dynamic range is bigger, higher detection accuracy. The pyramid detector dynamic range small, considering the spot in the pyramid cone offset on the tip into a spot in the four different exposure time of different cone to quantitative reduced four light the pupil like brightness difference, extended dynamic range. The design of fast steering mirror modulation spot on the tip of the cone prism along the circular track of the 4 pupil like one; because the incident led to the circular wavefront tilt center deviation prism cone, the cone on the spot in the exercise time is different, each pupil as the exposure time is different, so as to modulate the brightness, the pupil brightness as low as brightness can be improved to detect, track and circle center offset pupil image brightness difference and distortion caused by proportional to the original Wavefront reconstruction method is still applicable. The dynamic range of the theory of modulation type detector on the pyramid can be expanded from the original 1 to 12 lambda lambda (lambda =780nm), to meet the demand of actual detection of atmospheric turbulence. But the modulation type detector and pyramid problems were as follows: after the spot edge pyramid, the pupil like light field the superposition of four bright bands, high order error caused by wavefront reconstruction; bright band brightness value is 10 times the background noise, and therefore can not be directly deducted by threshold method. In addition, the pupil like due to offset not modulation occurred more than 3 pixels, or pupil image edge diffusion, precise positioning, and any pupil like inaccurate positioning error will lead to the corresponding pixel. At the same time, due to the introduction of axial chromatic aberration and prism light aperture edge diffraction halo formation, also can make the pupil image edge to form a few dozen pixel diffusion, Have a pupil like positioning error; in addition, large aberration input makes small aberration original response matrix calculation of the approximation is no longer applicable, at the same time modulation process causes the pupil image brightness difference and wavefront distortion of corresponding change, need to re calculate the response matrix modulation type pyramid detector. According to the simulation results of ZEMAX software design, to light the pupil like in the diagonal direction of inner and outer edge brightness value difference criterion, diagonal cone detector precision alignment, obtained the most sharp edge, inner and outer edge of collinear pupil image, pupil image edge and diagonal brightness value difference is 7.84%, the pupil modulation process of image due to vibration amplitude is less than 1 pixels. The bright band optical field superposition is introduced, analyzed and pupil spacing is the bright band of pupil influence of image brightness distribution, design and adjust as 3.8 times the radius of the center distance and pupil distance distribution, This can ignore the pupil image between bright band effects on signal detection. Instead of a single prism using double prism, eliminating the pupil single prism dispersion and dispersion causes like edge blur, while reducing the processing difficulty. Analysis of the prism pupil image edge brightness distribution characteristics, found on the pupil image edge the maximum brightness is not on the point positioning pupil image edge accurately. The positioning of the pupil image edge brightness gradient maximum, the positioning accuracy of the pixel level; and then use the algorithm two calibration pupil image center coordinates, so that the liquid crystal corrector high pixel density which contains 10 fine concentric target map at this time, the pupil like internal radiance distribution for concentric circles, calculate pupil image coordinates of the center of the coarse positioning coordinates of oscillation, the pupil image corresponding to the target map concentric degree of cross-correlation, the pupil image center two To improve the positioning precision, the sub-pixel level was set up. The laboratory static correction of aberration detection platform, in the premise of pupil and pupil like strictly conjugate, pupil image accurate extraction of the open-loop detection error compared with the pupil image coarse positioning system is reduced by 60%, with fast convergence speed closed-loop correction, reduce the residual 15%. analysis found that modulation pyramid type detector response matrix is equivalent to no modulation detection signal superimposed on the modulation polynomial introduced by the tilt modulation closed-loop, the distortion of its physical meaning is equivalent to the modulation induced response signal amplitude attenuation. Large aberration input, in response to the trigonometric terms contained in the matrix will lead to high frequency oscillation, through precise calculation; coordinate normalization, the attenuation of trigonometric function in variable threshold decreases, inhibited the decay oscillation. The trigonometric function for the response signal itself contains, due to the self change The amount of explicit Zernike polynomial in the reduction of large aberration input on the oscillation using small coefficient Zernike substrate, and in the process of reconstruction is applied to ensure the gain of the response matrix and compute the input amplitude. Aberration is about 10. When the response matrix modulation pyramid detector for 35 Zernike modes. The interferometer is built to verify the accuracy of the detector the light path using pyramid detector and Zygo, and the pupil as the basis of accurate positioning in precision, static random aberration on different amplitude contains 35 Zernike mode detection. The detection results show that using the open-loop response matrix and traditional response matrix compared to RMS detector error detection error decreases 50%, RMS for different amplitude aberration values were less than lambda /20, open loop system basically meet the detection requirements. The article concludes that, through the research to extend the dynamic range of pyramid wavefront sensor, real The dynamic detection range to meet the actual needs of the atmospheric turbulence detection pyramid wavefront sensor. At the same time, the detector has a high precision detection ability of input aberration in an adaptive optical system, detection error RMS value is less than lambda /20. the study will be used to pyramid wavefront sensor range from the closed-loop system to the open-loop system, to improve the magnitude of adaptive optical system, realize high precision imaging of faint objects play a positive role in promoting.

【学位授予单位】：中国科学院长春光学精密机械与物理研究所
【学位级别】：博士
【学位授予年份】：2017
【分类号】：TH74

【参考文献】