An Improved Method to Estimate the Fundamental Matrix Based on 7-point Algorithm

The fundamental matrix plays an important role in the field of computer vision. How to estimate it accurately has become a major subject. In this paper, an improved approach based on the 7-point algorithm is presented. In this method, initial estimations of the fundamental matrix are provided by the 7-point algorithm. Then an optimal fundamental matrix will be selected which offers relatively low mean, low maximum and low standard deviation of the discrepancy between points and epipolar lines. Experimental results on synthetic data and real images show that the proposed algorithm is superior over the 7-point algorithm in terms of accuracy and anti-noise ability.