A computer vision system for the detection and classification of vehicles at urban road intersections

Messelodi, Stefano; Modena, Carla Maria; Zanin, Michele

doi:10.1007/s10044-004-0239-9

Cited by 137 publications

(83 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In the past decade, researchers have explored the more complex task of monitoring vehicles at urban areas which includes monitoring of intersections [9], [5] and pedestrians and two-wheelers such as mopeds and cyclists [10], [11], [12]. Monitoring urban traffic is challenging due to the density of the traffic, variable types of road users, and lower camera orientations which aggravates occlusion.…”

Section: Related Workmentioning

confidence: 99%

Detecting Road Users at Intersections Through Changing Weather Using RGB-Thermal Video

Bahnsen

Moeslund

2015

Advances in Visual Computing

View full text Add to dashboard Cite

Abstract. This paper compares the performance of a watch-dog system that detects road user actions in urban intersections to a KLTbased tracking system used in traffic surveillance. The two approaches are evaluated on 16 hours of video data captured by RGB and thermal cameras under challenging light and weather conditions. On this dataset, the detection performance of right turning vehicles, left turning vehicles, and straight going cyclists are evaluated. Results from both systems show good performance when detecting turning vehicles with a precision of 0.90 and above depending on environmental conditions. The detection performance of cyclists shows that further work on both systems is needed in order to obtain acceptable recall rates.

show abstract

Section: Related Workmentioning

confidence: 99%

Detecting Road Users at Intersections Through Changing Weather Using RGB-Thermal Video

Bahnsen

Moeslund

2015

Advances in Visual Computing

View full text Add to dashboard Cite

show abstract

“…For top down, the whole context is analysed simultaneously or used to verify a hypothesis during searching. Motion silhouettes are generated from background modelling and classification is performed based on motion silhouette measurement features [20,24,18]. This approach is vulnerable to inaccurate foreground segmentation, which is inherent to urban environments due to low camera angles, occlusions, etc.…”

Section: Related Workmentioning

confidence: 99%

“…The above 2D approaches can be extended to 3D for vehicle detection and classification as in [24,18] and Buch et al [3,4]. The motion silhouette outline is used for classification in [18,3] and for vehicle detection of a single size in [22]. Wire frames are matched to images in [26].…”

Section: Related Workmentioning

confidence: 99%

“…Our approach takes the good results from 3D models into account [24,18,3] and defines the local features and the spatial relationship between them in 3D world space. The top down solution from histogram of oriented gradients (HOG) using a 2D search window [6] is generalised to 3D by 'wrapping' the camera image around the models like in [25].…”

Section: Dhog Detector and Classifiermentioning

confidence: 99%

See 1 more Smart Citation

3D Extended Histogram of Oriented Gradients (3DHOG) for Classification of Road Users in Urban Scenes

Buch

Orwell

Velastín

2009

Procedings of the British Machine Vision Conference 2009

View full text Add to dashboard Cite

This paper proposes and demonstrates a novel method for the detection and classification of individual vehicles and pedestrians in urban scenes. In this scenario, shadows, lights and various occlusions compromise the accuracy of foreground segmentation and hence there are challenges with conventional silhouette-based methods. 2D features derived from histograms of oriented gradients (HOG) have been shown to be effective for detecting pedestrians and other objects. However, the appearance of vehicles varies substantially with the viewing angle and local features may be often occluded. In this paper, a novel method is proposed that overcomes limitations in the use of 2D HOG. Full 3D models are used for the object categories to be detected and the feature patches are defined over these models. A calibrated camera allows an affine transform of the observation into a normalised representation from which '3DHOG' features are defined. A variable set of interest points is used in the detection and classification processes, depending on which points in the 3D model are visible. Experiments on real CCTV data of urban scenes demonstrate the proposed method. The 3DHOG feature is compared with features based on FFT and simple histograms. A baseline method using overlap between wire-frame models and motion silhouettes is also included. The results demonstrate that the proposed method achieves comparable performance. In particular, an advantage of the proposed method is that it is more robust than motion silhouettes which are often compromised in real data by variable lighting, camera quality and occlusions from other objects.

show abstract

“…Background modelling in the literature has been applied in a variety of situations including: motorways (Unzueta et al, 2012;Mithun et al, 2012), road intersections (Messelodi et al, 2005;Ottlik and Nagel, 2008), car parks (Choeychuen, 2012(Choeychuen, , 2013, swimming pools (Eng et al, 2004;Nuno et al, 2009) and water channels (Bloisi and Iocchi, 2009;Bloisi et al, 2014), etc. In general, we categorise different types of scenes into two groups: land scenes and water scenes, as background dynamics in these contexts differ markedly.…”

Section: Introductionmentioning

confidence: 99%

Modelling land water composition scene for maritime traffic surveillance

Pang¹,

Zhao²,

Hartill³

et al. 2016

IJAPR

View full text Add to dashboard Cite

Background modelling, used in many vision systems, must be robust to environmental change, yet sensitive enough to identify all moving objects of interest. Existing background modelling approaches have been developed to interpret images in terrestrial situations, such as car parks and stretches of road, where objects move in a smooth manner and the background is relatively consistent. In the context of maritime boat ramps surveillance, this paper proposes a cognitive background modelling method for land and water composition scenes (CBM-lw) to interpret the traffic of boats passing across boat ramps. We compute an adaptive learning rate to account for changes on land and water composition scenes, in which a geometrical model is integrated with pixel classification to determine the portion of water changes caused by tidal dynamics and other environmental influences. Experimental comparative tests and quantitative performance evaluations of real-world boat-flow monitoring traffic sequences demonstrate the benefits of the proposed algorithm.

show abstract

A computer vision system for the detection and classification of vehicles at urban road intersections

Cited by 137 publications

References 20 publications

Detecting Road Users at Intersections Through Changing Weather Using RGB-Thermal Video

Detecting Road Users at Intersections Through Changing Weather Using RGB-Thermal Video

3D Extended Histogram of Oriented Gradients (3DHOG) for Classification of Road Users in Urban Scenes

Modelling land water composition scene for maritime traffic surveillance

Contact Info

Product

Resources

About