Key Technologies Research On3D Surface Measurement System Based On Linear Structured Light Vision Sensor

Accompanied by the development of the Image Engineering Technology and Computer Vision (CV) Technology, More and more researchers put into the research and development work of the Three-dimensional (3D) Vision Measurement Technology (3D-VMT) in Reverse Engineering (RE). It played a huge advantage in the professional measuring&testing and engineering applications. On the basis of summing up the characteristics of the existing visual measurement methods and its measurement system, and in order to achieve measurement requirements of the integrity and high precision for the curved surface object, key technologies research on3D surface measurement system based on Linear Structured Light Vision Sensor (LSLVS) are studied in the thesis, and it effectively achieve the integrity measurement of the curve surface.The main research works of the thesis include the following aspects:(1) A three-dimensional surface structured light vision measurement system is established. Combined with the Coordinate Measuring Machining (CMM) as the motion drive-control equipment for linear structured light vision sensor, the precision motion control feature of the measuring machine is applied to the3D surface scan measurement. The mathematical model of the measurement system and its work process are proposed. The characteristics of the system and its influencing factors of precision performance are analyzed.(2) The influence factors of the accuracy in extracting the light stripe centers are comparatively analyzed and divided into four categories. The extracting centers methods of the linear structured light stripe are investigated. On the basis of the different characteristics of the extraction methods, a new idea is proposed that they are classified into two different methods, including the geometrical center method and the energy center method. A new extracting center method is proposed according to the image segmentation algorithm of GrabCut. It greatly reduced amounts and redundant datum operation of the calculation process, while improved the accuracy of extracting the center, based on extracting process after divide the complete pure light strip.(3) The Camera Calibration and the structural parameters calibration of the vision sensor are studied in the vision measuring system. Based on2D target calibration algorithm of Zhang for camera calibration, a new improved method proposed. Calibration results with higher precision are obtained, because of increased two distortion factors and global nonlinear optimization treatment. The conventional methods of the structural parameters calibration for the vision sensor are analyzed. Combined with the conditions of the plane-plane model for the linear structured light measurement, the structural parameters calibration based on free movement of the target plane is implemented, its experiments and error analysis are done.(4) The transformational relationship question between the light plane coordinate and the global measuring coordinate is solved, their transfer matrix is proposed. According to the application example experiments of two different models, the accuracy and completeness of the datum of3D surface measurement system based on Linear Structured Light Vision Sensor in the thesis are verified. Combined with the characteristics of the system, the applicability features of the system accuracy and its influencing factors and errors analyzed.