Eda Based Motion Deblurring for Sandroid Cubesats Imaging System

Motion blur, which increases the difficulty of locating and identifying target objects, has always been an ill-posed issue of space-based imaging system. In our paper, we point out the issue of deblurring degraded images caused by relative motion between space-based imaging system and observation targets. Such as motion of target objects to be observed or tremors of space-based imaging system. We propose a strategy to deblur degraded images that uses Estimation of Distribution Algorithm to obtain a better motion blur kernel in both linear and non-linear non-blind deconvolution situations. To produce a deblurred result, the strategy will be applied on a single image collected from the space-based imaging system on our Sandroid CubeSats, a member of NanoSats. Sandroid bases on Android operating system and our Sandroid CubeSats make full use of mobile device cluster as the main support of computing resources. Currently, the superior on-board computing capability of our Sandroid CubeSats makes sure that the proposed strategy is operative. In addition, experimental results prove that the proposed strategy is efficient and effective.