Comparison of uncertainty analysis of the Montecarlo and Latin Hypercube algorithms in a camera calibration model

In this study, we performed a comparison between two statistical methods to examine the uncertainties of intrinsic and extrinsic parameters in the conventional digital camera calibration process. The first methods, Monte Carlo, involves the generation of samples for one variable by a simple random sampling method and the second method, Latin Hypercube, generates random samples that occur within equal probability intervals with normal distribution for each range. The behavior of both methods is analyzed taking into account the number of iterations, confidence level and time of the calculation of the uncertainty. Our results show the advantages of the Latin Hypercube method over the Monte Carlo method, taking into account the number of executions of the model, maintaining a 95 percent confidence level and reducing the execution time considerably.