Comparison of algorithms for elimination the outliers in 3D modelling from video

In this paper, we make a comparison of different algorithms for the elimination of outliers in the process of 3D modeling from video. The process of 3D modeling from video uses specific algorithms for detecting and matching key points. The first step in this process is obtaining a cloud of points, then matching the key points, obtaining the network, and converting the network into a polygonal model. There are two types of points: inliers (points that describe the model) and outliers (unnecessary points). From the practical examples, we got results and conclusions for comparison of Random sample consensus (RANSAC), MEstimator Sample Consensus (MSAC), and Maximum Likelihood Sample Consensus (MLESAC) algorithms. Using different algorithms for eliminating the outliers in practical experiments we get different quality of the final 3D model.