Efficiency analysis of photogrammetric system by simulation modelling

We developed a universal simulation model of photogrammetric systems that utilizes artificial target points and fiducial marks (coded targets). The model allows to analyze the efficiency of the system in particular, measurement errors as well as system performance (scalability). Simulation modeling allows to estimate how different factors influence the measurement error and the performance of the system. It also allows to vary the values of factors easily and in a broad range. It reduces organizational, time, and financial cost of the testing system as well. We have implement the proposed simulation model in several configurations with GNU Octave. We have also run a series of simulation experiments, thus estimating how the measurement error of the system depends on errors in various input factors. It was determined that the measurement error of pixel coordinates of circle targets is the key factor that influences the resulting measurement error of targets' 3D coordinates. Other factors, such as deviations of the parameters of the camera model from its initial calibrations, or uncertainties in estimation of cameras' initial pose, do not influence the resulting measurement error significantly due to effect of system automatic calibration. We have also estimated the impact of the linear size and instrument error of scale bar on system accuracy. The proposed simulation model may be also used for the verification of various algorithms in the field of computer vision in conditions that are hard to implement during the process of natural experiments.