The characteristics for propagation of quantization errors for the question of pose estimation from line correspondences are introduced in this paper because quantization errors are the primary factors that affect the accuracy of pose estimation. Mathematical model of the propagation of quantization errors from images of line features to results of pose estimation could be set up in the form of closed-form solution. The research has some theoretical value. Furthermore, it has some practical value. One of important applications is to find the optimum line configuration. In priority, the paper presents the method to find the optimum line configuration based on quadrangular frustum pyramid model and determine the optimum line configurations on the analysis of the quantization errors. This analysis can increase the accuracy of pose estimation in computer vision.