Abstract

GARCES-SOCARRAS, Luis M. et al. Implementation of a Motion Detector for smart cameras on embedded systems. EAC [online]. 2020, vol.41, n.3, pp. 53-65.  Epub Dec 01, 2020. ISSN 1815-5928.

This paper describes the implementation of a motion detector system for traffic control’s smart cameras on embedded devices that can be used in multiple application fields. The proposed system detects moving areas in video sequences, thus reducing the processing time of video analysis algorithms by selecting regions of interest. As part of this work, three motion detection modules, and auxiliary blocks, were developed based on the bio-inspired elementary motion detector presented by Reichardt & Hassenstein. These motion detection modules have been incorporated in the portfolio of modules in the XIL XSGImgLib video and image processing library, which aims to speed-up and ease visual image processing applications over Xilinx FPGAs and SoC FPGAs. To test the detectors, a spatio-temporal motion detection application has been implemented on a Spartan-6 LX45 FPGA using video sequences from a vehicular traffic control point, showing execution time improvements on the elementary motion detector block compared with similar developments reported in the consulted literature.

Keywords : Smart camera; embedded image processing; elementary motion detection; FPGA-based system; XIL XSGImgLib.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )