An algorithm for resolution enhancement of low-resolution patterns captured by a sensor array

摘要

In digital holography, diffraction fields of 3D objects can be captured by sensor arrays. Although, capturing high resolution diffraction field paves the way to reconstruct objects with sharper edges and wider viewing angles, there are some imposed limitations on the capturing devices. For instance, when the size of the sensors are decreased, power of the captured light will drop, conversely power of the shot noise will increase. Hence, the quality of the captured pattern may be plummeted. As a solution to capture high resolution diffraction fields, we propose a super-resolution method which is based on capturing multiple low resolution patterns. First, the relation between high and low resolution patterns are described by a system matrix. Special ordering is used on the pixels of the low resolution patterns, thus system matrix is formed as a square and band matrix. Then, inverse of the system matrix provided to calculate high resolution diffraction. Performance of the algorithm is observed by using noisy low-resolution patterns, imprecise shifts of the sensor array along the transversal axes. The proposed algorithm provides robust results and pave the way to recover some of the high frequency components. Also, the simplicity of the algorithm makes it easy to implement.