Research of vision-based global localization for mobile robot

The robot vision system is the critical component of the mobile robot, and it perceives the most of the information from the vision system. Because of the variable illumination conditions, the traditional image segmentation method based on color information is not satisfactory. Based on the color information and shape information of the object, this paper proposes an object recognition algorithm that combining color image segmentation with edge detection. Furthermore, this article applies evolution strategy to particle filter, and combines with the scheme of adaptive sampling again to realize localization and map-creating simultaneously for indoor mobile robot. This article establishes a robust perception model to extract environmental features and updates feature. Experiments show that this algorithm can recognize the object exactly in the different illumination conditions, satisfy the requirement of the real-time and can improve the success rate and precision of the mobile robot global localization.