Discrete cosine single-pixel microscopic compressive imaging via fast binary modulation

Abstract

Single-pixel imaging technology has significant imaging advantages in the invisible wavelength and low light environment. Considering the energy compressive characteristic of discrete cosine spectrum, a single-pixel microscopic compressive imaging method based on the two-dimensional discrete cosine transform is proposed. In this method, a digital micro-mirror device is used for high-speed binary differential modulation of discrete cosine basis patterns; and the zigzag sorting is adopted to perform the under-sampling. The discrete cosine spectrum is generated by calculating the weighted sum of the light intensities measured by a single-pixel detector, and the object is reconstructed by two-dimensional inverse discrete cosine transform. The proposed method is verified by both computational simulations and laboratory experiments. The effects of different sampling rates, binary quantization level and compressive sampling modes on imaging results are discussed. The results show that the proposed method can realize high quality and high efficiency imaging. The method is suitable for microscopic imaging applications.