Title :
Using image profiles and integral images for efficient calculation of sparse optical flow fields
Author :
Sünderhauf, Niko ; Protzel, Peter
Author_Institution :
Dept. of Electr. Eng. & Inf. Technol., Univ. of Technol., Chemnitz, Germany
Abstract :
We propose a novel efficient algorithm for the calculation of sparse optical flow fields. Our algorithm uses profiles of summed pixel values of the rows and columns of overlapping image regions. It exploits the fast area summation that is provided by integral images. We describe the development of the algorithm, its application to UAV control and navigation and examine its runtime behavior and complexity.
Keywords :
mobile robots; path planning; remotely operated vehicles; robot vision; telerobotics; UAV control; image profiles; integral images; overlapping image regions; pixel values; sparse optical flow fields; Biomedical optical imaging; Cameras; Image motion analysis; Image sensors; Optical sensors; Pixel; Runtime; Sensor phenomena and characterization; Sensor systems; Unmanned aerial vehicles;
Conference_Titel :
Advanced Robotics, 2009. ICAR 2009. International Conference on
Conference_Location :
Munich
Print_ISBN :
978-1-4244-4855-5
Electronic_ISBN :
978-3-8396-0035-1
