Multiple Moving Object Detection From UAV Videos Using Trajectories of Matched Regional Adjacency Graphs

ABSTRACT:

Image registration has been long used as a basis for the detection of moving objects. Registration techniques attempt to discover correspondences between consecutive frame pairs based on image appearances under rigid and affine transformations. However, spatial information is often ignored, and different motions from multiple moving objects cannot be efficiently modeled. Moreover, image registration is not well suited to handle occlusion that can result in potential object misses. This paper proposes a novel approach to address these problems. First, segmented video frames from unmanned aerial vehicle captured video sequences are represented using region adjacency graphs of visual appearance and geometric properties. Correspondence matching (for visible and occluded regions) is then performed between graph sequences by using graph matching. After matching, motion similarity is performed for detection of moving object. Experiments conducted on several DARPA VIVID video sequences as well as self-captured videos show that the proposed method is robust to unknown transformations, with significant improvements in overall precision and recall compared to existing works.

PROJECT OUTPUT VIDEO:

EXISTING SYSTEM:

• In existing, most moving object detection works use video footage from fixed cameras. This enables background stabilization and subtraction techniques to be used as the background is relatively the same throughout the frame sequences.

• Once background pixels have been identified, they can be removed allowing foreground objects to be detected. In this section, two general background subtraction categories are discussed. They are techniques based on background modeling and those based on image registration.

• Background modeling has long been applied in moving object detection where foreground objects are detected based on a reference (i.e., background) model/image. One idea is to calculate the difference between each frame sequence against the generated model where a thresholding procedure finally determines the results. Temporal differencing is another alternative that takes differences between two or three successive frames to model background pixels.

DISADVANTAGES OF EXISTING SYSTEM:

• Background based approaches are indeed flexible and fast. Nevertheless, they only work well in a fixed camera environment where the background is expectedly constant.

• In a moving camera setup, however, camera motion and scene transitions exist making such background models unsuitable. Moreover, aside from the unstable backgrounds, the presence of multiple moving objects at varying speeds, slow/rapid illumination changes and/or noise from poor quality videos will also cause object detection to be problematic.

PROPOSED SYSTEM:

In this paper, we propose a moving object detection framework without explicitly overlaying frame pairs. First, each frame is segmented into regions and subsequently represented as a regional adjacency graph (RAG). Correspondence matching on a group of consecutive frames is performed by using graph matching where one-to-one correspondences are discovered through appearance similarity and geometrical constraints. Once correspondences are identified, motion similarity is performed for detection of moving object interms of separate foreground and background.

ADVANTAGES OF PROPOSED SYSTEM:

• It is suitable for moving camera setup.

• Our method is flexible for detection of occluded vehicle.

SYSTEM REQUIREMENTS:

HARDWARE REQUIREMENTS:

• System           : Pentium IV 2.4 GHz.

• Hard Disk      : 40 GB.

• Floppy Drive : 1.44 Mb.

• Monitor          : 15 VGA Colour.

• Mouse            : Logitech.

• Ram                : 512 Mb.

SOFTWARE REQUIREMENTS:

• Operating system : Windows XP/7.

• Coding Language : MATLAB

• Tool                       : MATLAB R2013A

REFERENCE:

Bahareh Kalantar, Shattri Bin Mansor, Member, IEEE, Alfian Abdul Halin, Member, IEEE, Helmi Zulhaidi Mohd Shafri, Member, IEEE, and Mohsen Zand, Member, IEEE, “Multiple Moving Object Detection From UAV Videos Using Trajectories of Matched Regional Adjacency Graphs”, IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2017.