Camera self-calibration is a fundamental problem in 3D reconstruction from un-calibrated image sequences and has been investigated by many researchers in recent years. In order to simplify the camera self-calibration process

the skew factor of the camera is usually assumed to be zero to get a group of linear equations where initial values of the camera's intrinsic parameters can be estimated

then a non-linear algorithm initialized with these values can be used to optimize the self-calibration result. However

when the camera is a complete perspective model

i.e. when the skew factor of the camera is not zero

the error of the self-calibration with the above assumption will be very large and sometimes the self-calibration algorithms will fail to get any result at all. In order to overcome the above problem

this paper addresses a self-calibration algorithm that can successfully compute the intrinsic parameters when the skew factor of the camera is not zero but known. Experiment results of synthetic data as well as real image sequences show that this algorithm can estimate the camera's intrinsic parameters precisely and make the successive 3D reconstruction more easily.