Real-time Wave-front Processor For Solar Adaptive Optical System Based On Cross Correlation Coefficient Algorithm

To realize the high resolution observation of the sun, large aperture solar telescopes, which have operated or under development, are installed with Adaptive Optical systems. Solar Adaptive Optical(AO) system, which works on low-contrast, extended objects such as the solar granulation and sunspot, is providing correction to the real time wavefront distortion introduced by the Earth’s atmosphere to obtain solar high resolution images. Compared with the night-time AO systems, daytime solar adaptive optical systems, which use correlation tracking algorithm, have a higher requirement in system bandwidth and frame rate and at the same time have serious challenges in the design of real time processor. This paper mainly did research on the high speed and real time wavefront processing technology for soalr adaptive optical system with cross correlation coefficient algorithm, aiming at meeting the real time requirement in the condition of large computation and high frame rate.The main research and innovations of this paper are as follows.We designed a real time wavefront processor for low-order 37 elements solar AO system based on FPGA+DSP, obtaining accurate image shifts results with cross correlation coefficient algorithm. Firstly, the system read image data out from the image sensor, then obtained image shifts of integer pixels by processing these data with correlation algorithm in FPGA, and finally obtained subpixel accuracy by parabolic interpolation in DSP. Taking advantages of the parallel computation and rich resource of FPGA, the system design used 3 kinds of parallel acceleration technology to improve the system’s computation speed and performance. The parallel acceleration technology of cross correlation pipeline, 2-dimension systolic array and multi-channel parallel architecture are keys to the design.We tested the validity of the design by adding sin voltage disturbance to the system through tilt mirror with both point source and extended source objects in a laboratory. The experiment simulated a low-order 37 element solar AO system and the subaperture arrangement of Shack-Hartmann wavefront sensor is 7×7. Among them, 30 subaperture are valid. The subaperture image is 24×24 pixels, the reference image is 12×10 pixels, and the frame rate of the detection camera is 2100 hz. Thelatency to compute a subaperture row is about 19 us, and to finish the calculation of total 7 subaperture rows will take 163 us. The experimental results show that the system can obtain correct image shift and satisfy the time latency requirement of the AO system.The research we did has important significance both on research and practical engineering for solar adaptive optical system.