Measuring traffic flow and classifying vehicle types: A surveillance video based approach

The paper presents a vehicle counting method based on invariant moments and shadow aware foreground masks. Estimation of the background and the segmentation of foreground regions can be done using either the Mixture of Gaussians model (MoG) or an improved version of the Group Based Histogram (GBH) technique. The work demonstrates that, even though the improved GBH method delivers performance just as good as MoG, considering computational efficiency, MoG is more suitable. Shadow aware binary masks for each frame are formed by using background subtraction and shadow removal in the Hue Saturation and Value (HSV) domain. To determine new vehicles in the subsequent frame (in addition to those in the current frame), invariant moments are used. For vehicles which are the same model and brand, color information and distance between center of mass and an imaginary reference line need to be considered. As for classification, the paper proposes a new method based on perspective projection of the scene geometry. The classification is grouped into three major tracks: bikes, saloon cars, and long vehicles. For each category, a lower and an upper bounding curve are developed to show the extent of their associated modality regions.

Anahtar Kelimeler:

Group based histogram, mixture of Gaussians, cast shadow removal, convex hull fitting, classification, modality regions

Measuring traffic flow and classifying vehicle types: A surveillance video based approach

Keywords:

Group based histogram, mixture of Gaussians, cast shadow removal, convex hull fitting, classification, modality regions,

PDF

___

locations in Famagusta city indicate that this approach is computationally eﬃcient. With MATLAB processing, computation of moments for one frame on the average took a little more than half a second. Since the proposed method needs to compare two consecutive frames the full processing on the average will require anywhere between 1–2 seconds.
However, since the most time consuming part is still the BE/FS, a full real time implementation would not become possible without using either a high level language or by moving towards more eﬃcient customized solutions such as ASICs.
Finally, also proposed was a new framework for classifying vehicles in traﬃc based on their modality curves. For each category a lower and an upper bound is generated. It is demonstrated that a new test vehicle can be categorized based on its area versus distance values and its Euclidean distance from the diﬀerent curves. The present paper also derives a mathematical formula for calculating the area occupied on the image plane by a particular vehicle on ground.
Even though the primary results on classiŞcation are promising more work has to be done to obtain the intrinsic parameters of the camera following the work presented by Bowen in [19].