Based On The Scattered Point Cloud Of Cigarette Case Structure Light Vision Detection System Design And Research

The rapid development of science and technology and tremendous changes in the market forces the Cigarette Industry to optimize the automatic product line. The universal and non-contact structured light visual detection method of cigarette case is presented in this dissertation, which can be used to distinguish and remove the unqualified cigarette case on the automatic product line. The application of the detection method has greatly increased the automation degree of the whole product line, reduced the labor intensity, improved the product testing efficiency, and also has a broad application prospect in other industry area.Based on the machine vision and3D reconstruction technique, the system acquires the3D coordinate data of the cigarette case’s surface with the laser triangulation method, analyzes the point cloud data, and implements the3D reconstruction of the cigarette case’s surface in order to judge the cigarette package’s quality.Firstly, the theory and structure of the online detection system as well as the point cloud data acquirement and processing method are discussed in this dissertation. According to the system requirements, the system experimental platform is built with the hardware including the laser projector, CCD camera, projector and camera stand, horizontal mobile platform, etc. By applying a naval point cloud data pretreatment method which is combined with manual intervention and automatic process, the implementation of the point cloud data processing is given.Then the software design of the3D reconstruction and appearance deformation detection of the cigarette case is proposed, with the employment of the Delaunay triangulation algorithm, NURBS surface fitting and error analysis method. Lastly, the design and development of cigarette case3D display on the monitoring interface is carried out. The test and experiments validate the system with effectiveness and reliability.