Compressive sensing has been widely used in single photon imaging systems because of its advantages of high efficiency and low cost. However, when the received photon flux is large, some photons cannot be recorded by single photon detectors due to the dead time effect, which introduces nonlinear errors between the measurement results and actual values and further damages the imaging quality. In this paper, a photon counting correction method specific to paralyzable detectors is proposed to improve the quality of reconstructed images in single photon compressive imaging systems. To verify this method, a single photon compressive imaging system is built, which uses a digital micromirror device (DMD) to modulate the light and a PMT as the single photon detector. The Monte Carlo simulation is also implemented to double validate the performance of the proposed method and the results from the experiment. Peak signal-to-noise ratio (PSNR) is used as the imaging quality evaluation standard. The experimental and simulation results indicate that our method can overcome negative effect of the dead time and accurately recover the intensity and waveform shape of echo signal, which can significantly improve the quality of reconstructed images and has a better performance than traditional methods in the single photon compressive imaging system.