噪声环境下无波前探测自适应光学系统扩展目标成像校正

Extended object imaging based on adaptive optics without a wavefront sensor under noise

无波前探测自适应光学扩展目标成像应用中,成像噪声的污染严重影响着自适应光学系统的校正能力,尤其是像差较大的情况下。选取对噪声不敏感的图像质量标准作为控制算法优化的目标函数是一种可能途径。文中把图像的能量谱熵作为控制算法的目标函数,使用61单元变形镜、随机并行梯度下降算法建立无波前探测自适应光学系统扩展目标成像校正模型。结果表明能量谱熵值的大小能准确地反映出有、无噪声时的图像质量变化情况。把能量谱熵作为自适应光学系统的性能指标函数时,尽管校正效果相比无噪声时有所降低,但情况最坏时斯特列而比SR值仅降低了0.1左右(SNR=5dB),自适应光学系统校正能力受噪声的影响显著减小。

Imaging noise can weaken correction capability of adaptive optics （AO） system without a wavefront sen- sor in application of extended object, especially for big aberrations. It is a possible solution to choose certain image quality standard, which is insensitive to the noise of image, as the performance metric of AO system. An AO sys- tem with Stochastic Parallel Gradient Descent （SPGD） algorithm and a 61 - element deformable mirror was simula- ted to restore the image of a turbulence - degraded extended object and the energy spectrum entropy （ESE） was used as the optimized object by control algorithm. The results show that ESE can give expression to the quality of image whether or not the image includes noise. Although the correct capability of AO system under noise was a lit- tle weaker than that of the image without noise, Strehl Ratio only decrease of by 0.1 at worst （ Signal Noise Ratio is 5dB）, which proves the effect of noise on correct capability of AO system can be mitigated obviously if the ESE is choused as the performance metric of AO system.