Single-photon 3D Reconstruction Based on Compressive Sampling Matching Pursuit

Traditional single-photon 3D reconstruction is often hindered by low-light conditions and backgroundnoise, making it difficult to accurately reconstruct the target object. To address these challenges, we propose a novel single-photon imaging algorithm that leverages the Compressive Sensing Reconstruction Algorithm, specifically the Compressive Sampling Matching Pursuit (CoSaMP) algorithm. Building upon existing imaging methods, our approach exploits the sparsity of the imaging scene to calculat e object depth information and filter out noise photons, thereby enhancing reconstruction accuracy. This method overcomes the li mitations of traditional algorithms regarding noise and temporal resolution. We innovatively apply CoSaMP to single-photon 3D reconstruction algorithms, achieving high-precision 3D scene reconstruction under low-light conditions. MATLAB simulation results demonstrate that, in human and sunflower data models, the me an absolute error (MAE) obtained by traditional single-photon 3 D reconstruction methods is approximately 35 cm, whereas our proposed method achieves an MAE of around 1.34 cm. These results indicate the feasibility of our approach in the realm of single-p hoton 3D reconstruction, offering new possibilities for application s in biomedical imaging and remote sensing.