Research On The Theory And Technology Of 3D Information Acquisition In 3D Scanner

Three Dimensional (3D) information acquisition has numerous applications inmany areas. It is one of the research focuses in information technology. In this thesis,we study the theory and technology of 3D information acquisition and develop anAsymmetry Dual-Camera Structured Light 3D Scanner. Some significant achievementsare gained.First, we give a survey on the researches and advance of current 3D informationacquisition methods. The advantages and disadvantages of these methods are analyzed.Then we point out the difficult points and research focuses in the research area.Based on the introduction of camera model, we discuss two kinds of 3Dinformation acquisition algorithm in structured light system. The Planar Constraint DLTmodel and its calibration algorithm are presented.We study the shading problem in structured light system deeply. Analyzing thecause of unmeasurable area, we draw a conclusion that the unmeasurable area dependon three factors: parameters of structured light system, relative position between theobject and the structured light system, the shape of object. The criterion of measurablefield in structured light system is presented. We define the absolute unmeasurable areaand present the criterion of it. Then we propose some methods to overcome the shadingproblem, including parameters adjustment, more camera or more structured light source,scanning in multi-direction and multi-viewpoint. We analyzed the relationship betweenthe two cameras in the Dual-Camera Structured Light system, and proposed thecalibration algorithm.We present the systems error analysis of 3D information acquisition system. Forstructured light system, the quantization error is discussed. The formula of measurementrange and pixel resolution in range is present. We study the off-center error in the rotatescan structured light system and construct the off-center error model. Then we presentthe algorithm of off-center error calibration and correction. The result error caused bythe stripe position error is studied. We point out the factors that can influence thelocation of the stripe. Base on these analysis, we proposed the ways to reduce systemerror. Based on introducing the information fusion theory, we applied information fusionmethods in 3D information acquisition system to solve some inherent problems instandard structured light system. We present a scheme to obtain 3D information of thelow reflectance area by fusing the shape from contour and structured light. Theinformation fusion based method to acquire color information are proposed. The schemeto fuse the images of two cameras are introduced.All of these analysis are supported by experiments.At last, an Asymmetry Dual-Camera Structured Light 3D Scanner is developed.The system scheme and structure are introduced. We obtain some satisfied results. Sometentative for the future work are proposed.