Compressed sensing for ultra-weak light counting imaging

Since array detectors with sensitivity to single photon level were limited by sensitivity on each pixel and needed large number of measurements, an imaging system with high sensitivity was designed to realize under-sampling ultra-weak light imaging detection. This imaging system based on photon counting technique and compressed sensing theory employed a Digital Micromirror Device (DMD) to complete the random spatial light modulation, and used a single photon point detector to collect photons. The total light intensity was recorded by the form of photon counting. Then, the image of an object under ultra-weak light illumination could be reconstructed by an algorithm. The influences of the number of measurements, ultra-weak light intensity level and measurement time on the quality of imaging were investigated by experiments. Furthermore, the evaluation criterion of reconstructed image and the Signal to Noise Ratio (SNR) of the system were discussed to analyze the experimental data. The experimental results show that when the number of measurements is greater than 19.5 percent of the dimension of data, it can acquire a good reconstruction, the SNR of the system can be even decreased to 2.8438 dB, and the average count of photons on each pixel of the DMD can be lower than 1.106 count/s. Experiments also prove that the key of imaging lies in the fact that the fluctuation of signal should be greater than the fluctuation of noise. It concludes that this imaging system meets the demand of ultra-weak light imaging detection for ultra-weak light intensity, high sensitivity and few measurements.