Computational ghost imaging with low-density parity-check code

Ghost imaging (GI) offers a robust framework for non-local sensing across various wavebands. While most GI research focuses on reconstruction from undersampled signals, the use of information redundancy to enhance reconstruction reliability in complex environments has rarely been explored. To address this, we introduce a computational GI system employing LDPC-coded radiation fields. By modeling the imaging process as a matching fading channel, we demonstrate that signal redundancy can be exploited to suppress environmental noise and non-ideal factors. Furthermore, we establish an analytical lower bound for the bit error rate (BER) of the proposed system. Numerical and experimental results confirm that the LDPC-coded GI framework significantly enhances performance in challenging sensing conditions.