Research On Map Construciton Method Based On Binocular Vision Oirentation

Mobile robot’s environment map construction is the key problem in mobile robotnavigation. At the same time, the precision of environment map construction has a greatinfluence for itself positioning of the mobile robot as well as the path planning. Therefore, theresearch for the building map of mobile robot has very significant theoretical significance aswell as application value. Currently, navigation in a totally-known environment is asophisticated technology，while navigation in a complicated totally-unknown environment isstill an open question and that is the main interest of the present thesis. Aiming at mapbuilding problem in the process of mobile robot navigation under unknown environment, thisarticle launches the research and the main work is as follows:First of all, it has carried on the induction and summary to the present situation of theenvironment map construction in the field of mobile robot in current domestic and externalresearch.Secondly, this paper studies the stereo matching algorithm and improves SURFalgorithm. The experimental results demonstrate that the improved arithmetic in rotation,illumination, dimension, noise, etc is superior to SURF algorithm. At the same time, therunning time of algorithm is less than SURF algorithm. Improved stereo matching algorithmis applied to implement the three dimensional positioning of objects and it has carried on thepositioning experiment under different distance measurement.The measurement results areanalyzed and the correction to get more precise three-dimensional coordinate information,ensuring the accuracy of the environment map construction.Finally, this paper puts forward the two-dimensional environment map building methodbased on distance. At the same time, using dynamic artificial potential field method is toconstruct the path planning under a two-dimensional environment map with dynamicobstacles under the dynamic environment which is built. Experimental results verify thefeasibility of two-dimensional environment map constructed.