Mobile mapping technology has grown exponentially the last ten years, particularly due to advances in computer and sensor performances. However, the very accurate positioning of data generated by such technique remains a crucial issue. The first part of this thesis presents the mobile mapping system that has been designed in the MATIS lab of IGN as well as its operational use. A detailed analysis of image data is proposed and data used for this work is discussed. The second part tackles the standard calibration procedure. First, camera calibration is performed by using a panoramic-based acquisition geometry, which allows not to required ground control points. Secondly, a full calibration procedure dedicated to the Stéréopolis V2is proposed so as to determine accurately the position and orientation of all the cameras. For that purpose, two procedures are explained : one requiring an area with points positioned with high accuracy ,and the other one based only the data acquisition. The third section details the compensation applied to the mobile mapping car that allows to improve poses of a large number of images. The mathematical formulation is proposed, and various cases of the method are explained. Data management is also presented since it is a mandatory step for efficient large amount of data management The fourth and final part of the thesis presents different registration scenarii, where methods developed in this work can be used individually as well as combined with other ones so as to bring higher coherence between sequences of distinct images
