Font Size: a A A

Small Moving Targets' Detection In Image Sequences

Posted on:2004-12-26Degree:MasterType:Thesis
Country:ChinaCandidate:X J CaiFull Text:PDF
GTID:2168360152456999Subject:Information and Communication Engineering
Abstract/Summary:PDF Full Text Request
The detection of moving targets in image sequences is always the focus of the research work these years. To detect the small targets in image sequences, background suppression is a key step. Generally the irregular motion of the video appliance leads to the motion of the background. So it is necessary to get the global motion estimation before detecting process of small target. In this paper, we propose a new method to estimate the global motion parameters, which is called Center-Biased Motion Detection using Hough Transform (CB-MDHT). It derives from MDRHT. But during the procedure of processing, we get the optimal displacement estimation of inter-frame and the randomicity in MDRHT is removed away in our method. The experiments show that this method can estimate the large-scale displacements between frames and satisfactory results are achieved.There are five chapters in this paper. The main work can be summarized as follows:In chapter I, a brief discussion is made on the necessities and the characteristics of small target detection, and the related theorem-charts are introduced at the end of this section.In chapter II, we focus on the background suppression technology in single frame. We concisely introduce the conventional methods for the detection of small moving targets in fluctuating background. By simulation results we conclude that it is unsuitable to directly suppress the background within single frame while the background is moving and complicated.In chapter III, to suppress the moving and complicated background, global motion estimation algorithms are considered. Firstly, we introduce some classical methods of the global motion estimation. Then, a threshold partition method is used to convert the gray images into black-white ones for the convenience of the image sequences detection. Because the target appears in pairs in the post-processed images, target pairs are kept and most noises are removed.In chapter IV, the motion association is considered, and the target trajectory is tracked based on the moving characteristics of targets.In chapter V, the work of this paper is summarized, and the developing tendency of small target detection in future is discussed.
Keywords/Search Tags:Small Target Detection, Global Motion Estimation, Background Suppression, Image Sequences, Optical Flow, MDRHT
PDF Full Text Request
Related items