Reserch On Processing And Analysing Of Point Cloud Data Of Aircraft Sheet Metal Parts

The main body of modern aircraft is sheet metal riveted structure. In order to fulfill theaerodynamic characteristics and stealth requirements of aircraft, the forming quality of aircraft sheetmetal parts have to meet more strict demands. Some dedicated devices are used in the traditionaldetection of aircraft sheet metal parts, the detection results are generally qualitative conclusions thatare difficult to discrible the sheet metal parts accurately. The accuracy of traditional detection methodsis also affected by the complexity of parts’ surface, and it’s difficult to obtain the geometryinformation of some parts’ surface by using traditional detection way. Some point cloud dataprocessing techniques in the process of digital detection are studied in this thesis, including3D digitalmeasurement, feature lines extracting, point cloud segmentation and plane fitting, etc. The typicalengineering features of sheet metal parts would be analyzed with some important local featuresextracted from the point cloud data. A point cloud data processing and analysing system is realized.The main research contents are summarized as follow:(1) The Half-Edge Data Structure is proposed to store triangle meshes after analyzing thetopological relation of triangle meshes. Because of the existing of noise points and poor qualitytriangles in the initial point cloud acquired by various measuring instruments, some mesh editingfunctions are developed, including triangles selection, deletion, repairment, etc.(2) The local features of sheet metal parts are very important in the process of analying typicalengineering features. Some local features extraction methods are studied in this thesis. Some methodsare proposed for extracting ridges and boundaries on triangular surfaces. A region-growing method isused to segment the plane points from point cloud data of aircraft sheet metal parts.(3) The detection elements of some typical engineering features of aircraft sheet metal parts aredefined and analyzed in this thesis. Some specific evaluation methods are proposed in the process ofdetecting various elements in some typical engineering features, such as bending feature, sinkingfeature, section feature, etc.(4) A point cloud data processing and analysing system is realized based on the full-textalgorithms and techniques. The performance of the system is verified through some specificexamples.