Study On The Performance Optimization Of Multi-tap Electron Multiplying Camera

With the rapid development of low-light-level night vision technology,people pay more and more attention to the research of low-light-level imaging devices and imaging system integration technology.And because of this,people carry out research in related fields.In order to solve the problem that it is difficult to improve the signal-to-noise ratio of CCD imaging under low illumination,vacuum optoelectronic devices and all-solid-state low-light imaging devices have been produced successively.The all-solid-state low-light-level imaging device of EMCCD adopts a unique "on-chip gain" technology,which makes it have excellent characteristics such as high frame rate,high quantum efficiency,and low equivalent readout noise.What’s more,this EMCCD is favored by the low-light-level imaging field.The EMCCD with multi-tap output greatly improves the readout rate and reduces the equivalent readout noise.It is quickly applied to the adaptive optics system of large ground-based telescopes that require low illumination and high frame frequency.The domestic demand for multi-tap EMCCD cameras is increasing.But it is increasingly difficult to purchase multi-tap high-speed EMCCD cameras.As a result,it is necessary to conduct research on high-speed multi-tap EMCCD camera integration and other related technologies.This dissertation starts from the photoelectric conversion model of E2V’s multi-tap CCD220,and analyzes the signal-to-noise ratio that the two main factors that limit the performance of multi-tap EMCCD cameras are fixed image noise and dark current noise.Therefore,the full text focuses on how to optimize fixed image noise and current noise.In optimizing fixed image noise,non-uniformity correction is mainly used to reduce fixed image noise.The dissertation first analyzes the current non-uniformity correction algorithm,and finally selects the non-uniformity correction algorithm of two-point correction + gain correction.In order to meet the requirements of real-time correction of the CCD220 camera,the algorithm is implemented by FPGA hardware.In the hardware implementation process of the algorithm,the image data decoding module,the NOR Flash data reading and writing controller module,the DDR3 SDRAM data reading and writing controller module,the electronic gain address mapping module and the image non-uniformity correction module are respectively designed.The algorithm achieved a good correction effect after being implemented by FPGA hardware.What’s more,the non-uniformity of the image response was reduced from0.3277% before correction to 0.2366% after correction.It also has good robustness and was successfully applied to 2000 fps multi-tap CCD220 high-speed high-sensitivity cameraIn terms of optimizing the total dark current noise,the total dark current noise in MPP mode and non-MPP mode was calculated theoretically based on E2V’s CCD220.According to the working conditions of CCD220 camera at 2000 fps,-38℃,it was concluded that the CCD220 camera works lower dark current noise can be obtained in MPP mode.In order to further explore the relationship between the total dark current noise and the working mode,the localized CCD220 detector of CETC-44 was used as the test object.By changing the amplitude of the vertical drive voltage of the photosensitive area and the storage area,the working mode of the CCD220 detection and the low vertical drive voltage were obtained.The voltage amplitude is closely related.Based on theoretical calculations and experimental tests,a general method for finding the best operating point of the detector was obtained.