Realtime Depth Estimation and Obstacle Detection from Monocular Video
This paper deals with the detection of arbitrary static objects in traffic scenes from monocular video using structure from motion. A camera in a moving vehicle observes the road course ahead. The camera translation in depth is known. Many structure from motion algorithms were proposed for detecting moving or nearby objects. However, detecting stationary distant obstacles in the focus of expansion remains quite challenging due to very small subpixel motion between frames. In this work the scene depth is estimated from the scaling of supervised image regions. We generate obstacle hypotheses from these depth estimates in image space. A second step then performs testing of these by comparing with the counter hypothesis of a free driveway. The approach can detect obstacles already at distances of 50m and more with a standard focal length. This early detection allows driver warning and safety precaution in good time.
KeywordsImage Region Image Space Depth Estimation Large Time Scale Obstacle Detection
Unable to display preview. Download preview PDF.
- 3.Bruhn, A., Weickert, J., Schnörr, C.: Lucas/Kanade meets Horn/Schunck: Combining local and global optic flow methods. International Journal of Computer Vision 63(3), 211–231 (2005)Google Scholar
- 6.Förstner, W.: 10 pros and cons against performance characterization of vision algorithms. Performance Characteristics of Vision Algorithms (1996)Google Scholar
- 7.Franke, U., Joos, A.: Real-time stereo vision for urban traffic scene understanding. In: Proc. IEEE Conference on Intelligent Vehicles, pp. 273–278 (2000)Google Scholar
- 8.Hager, G., Belhumeur, P.: Efficient region tracking with parametric models of geometry and illumination. IEEE Transactions on Pattern Analysis and Machine Intelligence 1025–1039 (1998)Google Scholar
- 9.Ke, Q., Kanade, T.: Transforming camera geometry to a virtual downward-looking camera: Robust ego-motion estimation and ground-layer detection. In: Proc. IEEE Conference on Computer Vision and Pattern Recognition, pp. 390–397 (2003)Google Scholar
- 10.Longuet-Higgins, H., Prazdny, K.: The interpretation of a moving retinal image. In: Proceedings of the Royal Society of London. Series B, Biological Sciences, vol. 208, pp. 385–397 (July 1980)Google Scholar
- 11.Lucas, B., Kanade, T.: An iterative image registration technique with an application to stereo vision. In: Proceedings of the 7th International Joint Conference on Artificial Intelligence, pp. 674–679 (1981)Google Scholar
- 12.Rabe, C., Volmer, C., Franke, U.: Kalman filter based detection of obstacles and lane boundary. Autonome Mobile Systeme 19, 51–58 (2005)Google Scholar
- 14.Sun, Z., Bebis, G., Miller, R.: On-road vehicle detection using optical sensors: A review. In: IEEE International Conference on Intelligent Transportation Systems, vol. 6, pp. 125–137 (2004)Google Scholar