Three-Dimensional Photon Counting Axially Distributed Image Sensing

In this paper, we present a 3D photon counting axially distributed image sensing system using statistical approaches, such as a maximum-likelihood estimation and total variation maximum a posteriori expectation maximization, to enhance the visual quality of 2D photon counting images and obtain better 3D reconstructed images. Conventional photon counting integral imaging is implemented by using a lens array, moving an image sensor in lateral (horizontal and vertical) directions, or using an image sensor array to obtain elemental images. To avoid lateral movement of the image sensor, axially distributed sensing is applied to photon counting imaging. A single image sensor is moved along its optical axis to pickup multi-view 2D images with slightly different perspectives which are used for 3D visualization. Axially distributed sensing (ADS) with proper statistical processing can remedy the effect of partial occlusion of the 3D scene. We show that our method can improve the visual quality of 2D photon counting images computationally and obtain the enhanced 3D reconstructed images.