Motion estimation for moving target detection

Abstract

This paper describes a suite of techniques for the autonomous detection of moving targets by processing electro-optical sensor imagery (such as visible or infrared imagery). Specific application scenarios that require moving target detection capability are described, and solutions are developed under the constraints imposed by the scenarios. Performance evaluation results are presented using a test data set of over 300 images, consisting of real imagery (visible and infrared) representative of the application scenarios.

Intravital leukocyte detection using the gradient inverse coefficient of variation

“…Let be a point on this clockwise ellipse, whose coordinate can be expressed as via a parameter . Note that the unit outward normal at can be written as (11) Let us now take discrete points on the ellipse determined through parameter values , then we can rewrite (7) and (8) respectively as their unbiased sample estimate within this finite sample (12) and (13) where is given by (14) Furthermore, we can express (10) as shown in (15) at the bottom of the page. Thus, the ellipse with the highest GICOV score is denoted by , and the maximal GICOV score corresponding to this case is denoted by .…”

“…There exist several methods for automatic object detection [4]- [11]. The Hough transform (HT) method [4] is a wellestablished approach, where the detected edge pixels vote for a shape according to a parametric representation in the parameter space.…”

“…A different approach based on image level sets has been suggested by Mukherjee et al where possible cells are captured through a search of the image level lines-boundaries of connected components within level sets [10]. The temporal information obtained by optical flow is used in low contrast object detection by Markandey et al [11]. The basic assumption of the optical flow technique is that the motion shift per frame should be less than the width of the transition of image edge.…”

Correlation Filter With Motion Detection for Robust Tracking of Shape-Deformed Targets

“…In our motion detection algorithm, frame differentiating [30], [31] is used to detect moving objects between two frames; i.e.,the greater the difference between the two frames, the greater the probability that the location of the moving object will be. However, lens shaking or camera moving will cause significant noise in frame differentiating, even the target does not move at all in the scene.…”

Independent moving target detection for aerial video surveillance

“…In [5], an active control mechanism has been incorporated to facilitate the detection of independent motions based on normal image flows, but the algorithm can only be applied to situations where the observer undergoes translational motion. Some methods [6,7] rely on the accurate computation of optical flow field, which is indeed an attractive solution if image flows can be estimated robustly, as sometimes they can be very useful in the detection of low contrast targets. However, in aerial surveillance applications, moving ground targets are usually so small with few textures that their motions can hardly be estimated accurately.…”