Detection and Tracking of Large Number of Targets in Wide Area Surveillance

Abstract

Abstract. In this paper, we tackle the problem of object detection and tracking in a new and challenging domain of wide area surveillance. This problem poses several challenges: large camera motion, strong parallax, large number of moving objects, small number of pixels on target, single channel data and low framerate of video. We propose a method that overcomes these challenges and evaluate it on CLIF dataset. We use median background modeling which requires few frames to obtain a workable model. We remove false detections due to parallax and registration errors using gradient information of the background image. In order to keep complexity of the tracking problem manageable, we divide the scene into grid cells, solve the tracking problem optimally within each cell using bipartite graph matching and then link tracks across cells. Besides tractability, grid cells allow us to define a set of local scene constraints such as road orientation and object context. We use these constraints as part of cost function to solve the tracking problem which allows us to track fast-moving objects in low framerate videos. In addition to that, we manually generated groundtruth for four sequences and performed quantitative evaluation of the proposed algorithm.

[Paper] Visual Tracking Method of a Quick and Anomalously Moving Badminton Shuttlecock

“…6 (a 4) ) are blended. The observed color of the shuttle region is found on the (9) (8) (Fig. 6 (a 1 )~6 (a 6 ).…”

Recognizing Humans in Motion: Trajectory-based Aerial Video Analysis

“…Reilly et al detected moving objects by median background model, and objects are tracked using bipartite graph matching with the combination of road orientation and object context [12]. Xiao et al proposed a joint probabilistic relation graph approach to track a large number of vehicles [17].…”

“…It may be distributed unchanged freely in print or electronic forms. In general, the technique of tracking objects consists of the following three steps: (i) video images, which are captured by an airplane / unmanned aerial vehicle (UAV), are stabilized at the ground plane, (ii) object areas are extracted by background subtraction techniques or appearance-based detectors, and (iii) detected objects are tracked in sequential images by several approaches, such as Kalman filters, particle filters, a graph-based model etc [10,11,12,18]. Due to the ego-motion of the camera on the platform as shown in Fig.…”

Motion Component Supported Boosted Classifier for Car Detection in Aerial Imagery

“…(Kozempel and Reulke, 2009) provide a very fast solution which takes four special shaped edge filters trying to represent an average car. Another approach of (Reilly et al, 2010) shows a method which is based on background subtraction. The background is computed by a 10 frame median image.…”