A statistical method for estimating the accuracy of scene reconstruction using a collinear digital stereo vision system

摘要

We present a model and technique for the analysis of measurement capabilities of on-board collinear digital stereo vision systems. For this, we use a technique of determining statistical distributions of measurement errors and report the obtained results. The proposed model can be employed for pre-calculating the characteristics of on-board collinear digital stereo vision systems, while the results can also be used to estimate the accuracy of reconstructed maps. This is especially interesting for the oil industry, where continuous monitoring of the conditions of long pipelines using autonomous robots is necessary. With the help of appropriate stereo vision systems, it is possible to assess the conditions of pipelines, detect critical deformations, fix gust places, etc. to ensure safety in the oil infrastructure.