Differential Analysis of Two Model-Based Vehicle Tracking Approaches

Abstract

Abstract. An experimental comparison of 'Edge-Element Association (EEA) ' and 'Marginalized Contour (MCo)' approaches for 3D modelbased vehicle tracking in traffic scenes is complicated by the different shape and motion models with which they have been implemented originally. It is shown that the steering-angle motion model originally associated with EEA allows more robust tracking than the angular-velocity motion model originally associated with MCo. Details of the shape models can also make a difference, depending on the resolution of the images. Performance differences due to the choice of motion and shape model can outweigh the differences due to the choice of the tracking algorithm. Tracking failures of the two approaches, however, usually do not happen at the same frames, which can lead to insights into the relative strengths and weaknesses of the two approaches.

A Framework for Model-Based Tracking Experiments in Image Sequences

“…-OFThresh (see item 'Optical Flow Thresholding' in Section 3). -Wheelarch, the addition of wheelarches as already studied with a smaller sample size in Dahlkamp et al (2004b). -Shadow, the addition of visible shadow edges to the model.…”

Vehicle Detection From an Image Sequence Collected by a Hovering Helicopter

“…The objective is to find this template back in each image considered, (Ballard, 1981;Ferryman et al, 1995;Tan et al, 1998;Pece and Worrall, 2002;Hinz, 2003;Zhao and Nevatia, 2003;Dahlkamp et al, 2004;Pece, 2006;Ottlik and Nagel, 2008). The disadvantage of model-based methods is the high dependency on geometric details of the considered object, which in our case would require that vehicles should appear in the image sequence with many details and with clear boundaries.…”

Combining shadow detection and simulation for estimation of vehicle size and position

“…As said in (Dahlkamp et al, 2004(Dahlkamp et al, , 2007, switching from 2D tracking in the image plane to 3D tracking in the scene domain often results in a substantially reduced failure rate, because more prior knowledge about the objects can be utilized. For example, foreground regions are not always homogeneous, and we need some criteria to merge regions that are likely to belong to the same object.…”

“…(Dahlkamp et al, 2004) showed how a shadow added to their polyhedral model could improve the accuracy when fitting the model to edge features. (Song and Nevatia, 2007) give examples of how simulating shadows on a 3D box (with fixed size) can improve detection.…”

“…Many other models of different complexities have been proposed, ranging from a few 3D line features (Kim and Malik, 2003), rectangular 2D boxes (Magee, 2004), and more explicit models like polyhedral models (Nagel, 2004;Dahlkamp et al, 2004Dahlkamp et al, , 2007Lou et al, 2005). In our system the 3D box appears to be a sufficiently complex model, and it is simple enough to cover most types of vehicles.…”