Motion and intensity-based segmentation and its application to traffice monitoring

Badenas, Jorge; Bober, Miroslaw; Pla, Filiberto

doi:10.1007/3-540-63507-6_238

Cited by 10 publications

(3 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…If one of these errors for a region is lower or similar to the error that it had with its estimated motion, then both regions are merged. For a complete description of the frame to frame segmentation we refer the reader to [3].…”

Section: O T I O N S E G M E N T a T I O N U S I N G Two C O N S E mentioning

confidence: 99%

See 1 more Smart Citation

Applying computer vision techniques to traffic monitoring tasks

Badenas

Pla

1998

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

A b s t r a c t . This paper presents a method I for tracking and segmenting vehicles in a traffic scene. The approach is based on a frame to frame segmentation followed by a tracking process. As opposed to usual segmentation methods, our method feedbacks segmentation with tracking information for improving results. Several facilities are provided for traffic monitoring such as vehicles trajectories surveillance, segmentation of vehicle shape, measuring the mean velocity of the traffic, counting the vehicles that are moving on the lanes of a road or a motorway, counting the vehicles that stop at a junction and detection of events such as a vehicle stops on a road or a possible accident.Key Words : Region tracking, motion analysis, motion segmentation, traffic monitoring, vehicle surveillance I n t r o d u c t i o nIn recent years, as result of advances in information techniques both in terms of computational power and cost, it has become possible to use computer vision to perform many everyday tasks. This paper shows how it can be applied to traffic monitoring substituting existing methods such as Inductive Loop Detectors and Microwave Vehicle Detectors which are more expensive, and have more limited usefulness.A typical traffic scene consists of untextured objects with a regular shape moving over a surface which is also untextured. In this context, a traffic monitoring system has to survey moving objects detecting anomalous behaviors and situations, and producing measures.Several works have been reported on the application of computer vision to traffic monitoring. One of the early works [1] describes a system with motion analysis based on simple frame differencing. This simple approach can not provide enough information and produces poor results due to it is very sensitive to noise. Dubuisson and Jain [2] proposed a technique for segmenting vehicles also based on image substraction, but combined with color segmentation of regions. [5], where the edges extracted from the image are matched to 3-D segments of a generic car model that is projected onto the image plane. This matching is only applied to edges moving with coherent motion, to avoid mixing up edges from different vehicles. 2-D models have a lower computational cost, but present problems when the size and the number of vehicles increase. In [7], B-splines with four control points are fitted to vehicles projection. Weber at al.[6] method describe the contours by a closed cubic spline and employs two Kalman filters, one for estimating the affine motion parameters and other one for estimating the shape of the contours.A different group of methods for motion analysis tracks single tokens such as points or lines, that are extracted fl'om the image,[8], [9]. These methods present problems due to they do not provide explicit grouping of tokens moving with coherent motion and are quite sensitive to noise.We propose in this paper an approach based on tracking regions which integrates information recorded over a sequence on a frame by frame basis for improving se...

show abstract

Section: O T I O N S E G M E N T a T I O N U S I N G Two C O N S E mentioning

confidence: 99%

“…Hence, it distinguishes moving regions from stopped regions, but not regions with different motions. Badenas et al [3] presented another method based on regions. This method is a combination of several techniques which produces motion estimation and segmentation of vehicles.…”

Section: Introductionmentioning

confidence: 99%