Clear Reconstruction Of Moving Scene Based On Coded Exposure Imaging

Computational imaging is a new imaging technology to solve the disadvantages of traditional photoelectric imaging.This imaging technology introduces calculations into traditional imaging to break through the limits of traditional visual information recording.Computational imaging technology aims to analyze targets from multiple dimensions,and achieve better imaging effects by collecting multiple dimensions of information and realizing information conversion between different dimensions,which include space,time,angle,spectrum and phase.The coded exposure imaging technology,as an important branch of computational imaging,mainly solves the problem of image motion blur.This technology sacrifices time dimension information in exchange for image spatial information,which can preserve image detail information to the greatest extent.This paper mainly studies the clear reconstruction technology of motion scene based on coded exposure.Aiming at the restoration from a single frame image,a blurred scene restoration method based on the optimal coding sequence selection and motion prior information of the Memetic algorithm is proposed.By analyzing the principle of encoding exposure imaging,the optimal encoding sequence search criterion is established;in order to solve the limitations of traditional search algorithms,the Memetic algorithm is introduced to search for optimal code words;and the object image relationship is combined to estimate the Point Spread Function(PSF)of the moving target;using the background difference method to synthesize the motion blur superposition characteristics to achieve the separation of the target scene;using the Bayesian maximum posterior probability framework to achieve the restoration and reconstruction of the blurred image.An experimental platform is built for real shooting experiments.The experimental results show that this method can effectively solve the morbid problem of traditional motion blurred imaging restoration,greatly improve image resolution,and have better subjective and objective evaluation.In order to further study the coded exposure imaging,the coded exposure imaging technology is introduced into the multi-frame image restoration,and a multi-frame image joint restoration method based on the coded exposure imaging is proposed.The Golay pairing matrix is used to construct a set of complementary coding sequences of moderate length.Arbitrary coding sequence in the coding sequence sets satisfies the frequency domain complementarity criterion.A certain number of optimal coding sequences conforming to the coding exposure imaging are selected from the coding sequence set.The amplitude curve of the joint coding sequence has a larger minimum and a lower value.The variance value ensures that the obtained multi-frame image can retain more detailed information,and at the same time there is information complementary to each other.Finally,combined with Total Variation(TV)regularization model to achieve joint restoration of multi-frame images.Experimental results show that this method can retain more image detail information,and can effectively suppress the negative effects in the restoration process,and the restored image has better subjective and objective evaluation.