Hybrid regularization method for LRT image reconstruction under incomplete projections

Lidar reflective tomography (LRT) system is able to achieve long-distance target detection and collect laser projection data for target reconstruction. In target detection with laser, the projection data is often incomplete or missing, where the traditional reconstruction algorithms are no longer effective. Therefore, this paper applies the regularization model to LRT image reconstruction and proposes a hybrid regularization method for higher LRT imaging performance, to prevent the over-fitting problem of the model and improve the speed and stability of the proposed method. Meanwhile, the edge and the structural characteristics of the reconstructed target with LRT is better enhanced by using the proposed hybrid regularization method. By conducting the LRT outfield experiment and using the collected LRT projection data, the validity of the proposed method in incomplete projection reconstruction is well proved.