Motion and Intensity-Based Segementation and Its Application to Traffic Monitoring

This paper is concerned with an e cient estimation and segmentation of 2-D motion from image sequences, with the focus on tra c monitoring applications. In order to reduce the computational load and facilitate real-time implementation, the proposed approach makes use of simplifying assumptions that the camera is stationary and that the projection of vehicles motion on the image plane can be approximated by translation. We show that a good performance can be achieved even under such apparently restrictive assumptions. To further reduce processing time, we perform gray-level based segmentation that extracts regions of uniform intensity. Subsequently, we estimate motion for the regions. Regions moving with the coherent motion are allowed to merge. The use of 2D motion analysis and the pre-segmentation stage signi cantly reduces the computational load, and the region-based estimator gives robustness to noise and changes of illumination.