Motion detection of vehicles based on FPGA

Abstract

In this paper a hardware approach for evaluating motion detection of vehicles on transit roads was proposed. A motion detection method is used through computational vision with a fixed camera which is based on the difference between the current image and a reference image of the environment that is being monitored. Results are compared with a software approach in order to validate and show the effectiveness of a hardware approach. Experiments based on real vehicle images in roads were performed and results are 7.5 times faster by using a reconfigurable hardware approach as compared to the same application in a software approach. The results were also compared with another hardware approach for motion detection and a performance improvement of about 66% was observed in the image processing.