Method for catadioptric vision based localization and mapping in a particle filter framework

摘要

A method for the forward vision sensor based on location estimation and mapping are disclosed in the particle filter framework. The method (a) from the current position of the moving object, obtaining an omnidirectional panoramic image for exploration environment, using the obtained omnidirectional panoramic image, the azimuth from the moving object to the feature points each position present in the sensing environment detecting a; (B) it based on the posterior probability distribution of the feature points and the detected azimuth pre-update location, updating the posterior probability distribution of the current position of the moving object; And (c) based on the detected azimuth angle, and the posterior probability distribution of the updated moving object location, comprising the step of updating the posterior probability distribution of the feature point location, and repeating the step (a) to the step (c) It characterized in that it performs. ; According to the invention, SLAM (Simultaneous Localization And Mapping: hereinafter referred to as SLAM) and the amount of calculation congestion problems that occur when the particles to be applied to the filter (Particle Filter), through the omnidirectional visual sensor to acquire color screen occupancy map can, can provide a multiplicity of information on the moving object exploration environment, it is possible to provide a position estimation algorithm using the default sensor camera Taiwan for moving the robot over the position estimate with the obtained color screen occupancy map .