In the aviation industry， the demand for high accuracy of the airplane product is growing， which makes precise production of airplane parts and accurate assembly become more and more important. Three-dimensional laser scanner is widely utilized to capture the local shapes， represented by 3D point clouds， of a aircraft product from different viewpoints， followed by aligning various point clouds into the same coordinate system so as to form the entire geometry of the product， which is referred to as registration. Multiview registration of point clouds is therefore a critical step for reconstruction and inspection of aircraft products. The input of the system is the point cloud data with targets obtained by laser scanners at different stations. The point cloud data are in different measurement coordinates systems and have different degrees of overlap. Furthermore， we formulate multiple scan registration refinement as an optimization problem on an undirected graph， where each node and each edge represent a single scan and a connection between two overlapped scans， respectively. The multiple scan registration problem is solved via hierarchical optimization， and thereby the entire geometry shape of the aircraft is accurately obtained. In order to register these point cloud data to a unified coordinate system， a series of methods are proposed to improve the accuracy and efficiency of point cloud registration.