Structured Light 3D Vision System Based On Time Encode Method

Along with economy development, community progress, and manufacture level enhance, a good many industries and fields, .g. equip manufacturing and so on, have stringent demand of 3D non-contact on-line inspect, which brings on fast development of optical 3D vision detection instruments represented by 3D vision detection instruments based on structured light. In all kinds of optical 3D vision detection technologies and instruments, structured light method has measuring speed and resolution advantages and is generally thought as the most practical method so far, which has been confirmed by the succeeded examples. Therefore, 3D vision detection instruments based on structured light have wondrously and important background in application. This thesis analyzes and reviews the home and abroad actuality and development of 3D vision detection instruments based on structured light, introduces its basic principle, and builds up its system using time encoding with two levels of grey. The camera model and the projector model is established, then the range sensor model is established through introducing geometry constraint condition to relative position of the camera with the projector. The method using normal and reverse encoded patterns is used to realize sub-pixel localization of the stripe edges, and to reduce the influence of environment light and surface change on the range measuring results. The parameters of the camera and the projector are simultaneously estimated by using non-linear least-square estimation model, whose outputs contain the estimated parameters and their accuracys. By comparison of theoretical precision of 3-D check points σ? X, σ? Y, and σ? Z with empirical accuracy RMSE X, RMSE Y, and RMSE Z, the our proposed model, estimated values of system parameters and their accuracy, the range measuring accuracy estimations and so on will be validated. The error ellipsoid whose the longest axis length and orientation are given is used to represent the covariance matrix of the range measurement. The relation of longest axis orientation of the error ellipse with the projector optical axis orientation and the relation of the longest axis of the error ellipse with the distance from the camera are researched. The calibration method is proposed and our system calibration experiment is accomplished. Our research supplies a method improving the range measurement accuracy with the structured light 3D vision system based on time encoding method, and enriches the structured light 3D vision theory. Our research will have practical guidance role for the design and application of the structured light 3D vision system.