Automatic Regions-of- Interest Selection based on Pearson's Correlation Coefficient

Arthur de Miranda Neto (1) , Alessandro Correa Victorino (1) , Isabelle Fantoni (1) , Douglas Eduardo Zampieri (2)

1 Heudiasyc - Heuristique et Diagnostic des Systèmes Complexes [Compiègne]
2 UNICAMP - Universidade Estadual de Campinas = University of Campinas

Navigation of a Mobile Robot is based on its interaction with the environment, through information acquired by sensors. Particularly for Mobile Robot navigation in unknown environment, the type and number of sensors determines the data volume necessary to process and compose the image from the environment. Nevertheless, the excess of information imposes a great computational cost in data processing. Taking into account the temporal coherence between consecutive frames, a Discarding Criteria methodology based on Pearson's Correlation Coefficient (PCC) was proposed and applied as a Dynamic Power Management solution to a robotic visual-machine perception. In this context, this work proposes an environment observer method based on PCC that instead of processing all image pixels, it selects automatically only the regions-of-interest (ROI) and processes it in real time in order to perform a task: road detection and obstacle avoidance. This real-time visual perception system has been evaluated from real data obtained by two experimental platforms.

Domaines: Automatique / Robotique

Arthur de Miranda Neto, Alessandro Correa Victorino, Isabelle Fantoni, Douglas Eduardo Zampieri. Automatic Regions-of- Interest Selection based on Pearson's Correlation Coefficient. IROS Workshop on Visual Control of Mobile Robots (ViCoMoR), Sep 2011, San Francisco, United States. ⟨hal-00657529⟩