Abstract: The atmospheric coherence length r0 is an important parameter for measuring the strength of atmospheric optical turbulence along the integral path. To address the challenges of beacon absence and real-time detection requirements in practical application scenarios, a deep learning-based laser active illumination technique for r0 detection is proposed in this work. By constructing a laser active illumination atmospheric transmission simulation model, a one-to-one correspondence between the imaging spot images and atmospheric coherence length r0 is established, which is used to train a convolutional neural network (CNN) model. Additionally, a segmented model training and prediction optimization strategy is proposed, further enhancing the model's robustness under strong turbulence condition. The results demonstrate that the proposed model achieves mean relative errors between the predicted results and the true values consistently below 6% under varying surface roughness conditions, with a single prediction time of less than 0.49 ms. This preliminarily validates the feasibility of the proposed method, and provides a new approach for beacon-free, high-precision, high-robustness, and real-time atmospheric coherence length r0 detection. The study offers valuable insights for next-generation laser computational transmission research.

Key words: atmospheric turbulence, atmospheric coherent length, deep learning, laser active illumination