Cascade classifier as a tool for quick camera calibration

The mobile robotics development is very rapid nowadays. In the interest of companies and institutions is to transfer more and more task for mobile robots. The conversion from one application to other even similar is often associated with changes in hardware. The authors have performed a study witch effects possibility to significantly reduce the time of computer vision system calibration. The following paper presents multiple cameras calibration process for mobile robotics applications. The selection of applicable marker for cascade classifier training is described. Then the calibration of various cameras is shown. To perform this task the authors used application created by them especially for this. Results with emphasis for calibration errors are one of the most important parts of the paper. In the end, there are conclusions and prediction for further development and usage of this application.