Simultaneous location and mapping by fisheye vision
The localisation of autonomous robots in industrial environments currently comes up against the problem of the different localisation systems required indoors and outdoors (laser, GPS, etc.).
Locavision’s objective is to develop a SLAM (Simultaneous Localization and Mapping) algorithm to obtain visuals via cameras, in an industrial context capable of managing the mobile’s passage between the indoor and outdoor environment (loss of points of interest or loss of GPS signal).
To solve the problems posed by the number of points of interest (uniform surfaces of buildings, loss of points when passing through doors), or the scale factor, SEGULA Technologies chose to integrate several sensors (stereoscopic system, IMU, odometry, etc.) into the LOCAVISION system. In addition, to optimise the mobile’s movements, the system is equipped with a learning algorithm.