Spatial-frequency encoded imaging of multangular and multispectral images

Abstract

Modern imaging techniques increasingly require signals to be collected from multiple viewpoints and spectral bands to realize multi-dimensional and multi-species detections. For this purpose, multiple cameras are commonly required. Each camera collects signals from one viewpoint or one spectral band, resulting in a considerable experimental cost. Based on frequency modulation, this work proposes an encoded-imaging technique that can record multangular and multispectral images in one acquisition. The signals recorded from different viewpoints and spectral bands are superimposed in the spatial domain, while being separate in the frequency domain. This allows us to extract individual images based on their respective frequency components. In this work, a proof-of-concept experiment was conducted. The high correlation coefficient between the superimposition of the extracted images and a normal superimposed image demonstrates the effectiveness of this technique. In addition, an improved mathematical formulation was proposed to describe the higher spatial-frequency components, which were considered merely to be residual lines in previous studies. The proposed encoded-imaging technique may have potential for multangular and multispectral imaging, which is especially useful for tomographic reconstructions.