**3. Methodology**

This section describes the preprocessing of stereo images and LiDAR point clouds, the structure of our proposed network, the loss function for network training, and the postprocessing of the network output. These descriptions are commonly applied to the calibration of the LiDAR-stereo camera and LiDAR-LiDAR combinations. We chose a LiDAR as the target sensor in the LiDAR-stereo camera and LiDAR-LiDAR combinations, and the rest of the sensors as the reference sensor.
