Robots are coming to play important role in the traditional field of engineering for their exceptional mechanical capabilities and flexible movement abilities. The quadruped robots equipped with professional surveying instruments will significantly contribute in field investigations and mapping in more complex terrains. However, challenges persist in terms of integrating point cloud alignment of surveying observations with robotic motions due to the occurrence of registration errors and increased computational costs caused by millions or even billions of LiDAR points. To address these limitations, this paper proposes a novel system and an alignment assistant module focusing on introducing an optimal trade-off between accuracy and efficiency in point cloud processing of complex terrain environments. In contrast to rough alignment method between point cloud sets, it utilizes the transformation results based on SLAM techniques, which sets up initial values of gradient directions in the translation and rotation components. Experimental results show that, compared to the direct ICP registration method, an average error of 3.16 cm is achieved costing 4.11 s between two station-based point cloud sets.