Compressed Sensing for Imaging Applications

Compressed sensing is a new sampling theory which allows reconstructing signals using sub-Nyquist measurements. This can significantly reduce the computation re­quired for both image and video whether during acquisition or encoding, especially at the sensor. Compressed sensing works on the assumption of sparsity of the sig­nal in some known domain, which is incoherent with the measurement domain. We exploit this technique to build a single pixel camera using an optical modulator and a single photosensor. Random projections of the signal (image) are applied to the optical modulator, which has a random matrix displayed on it corresponding to the measurement domain (random noise). This random projected signal is focused and summed at the photosensor and will be later used for reconstructing the signal. In this scheme, a tradeoff between the spatial extent of sampling array and a sequential sampling over time with a single detector is performed. In addition to the single sensor method, we will also demonstrate a new design which allows compressive im