This paper addresses the problem of background modeling for foreground object detection in complex environments. A Bayesian framework that incorporates spectral, spatial, and temporal features to characterize the background appearance is proposed. Under this framework, the background is represented by the most significant and frequent features, i.e., the principal features, at each pixel. A Bayes decision rule is derived for background and foreground classification based on the statistics of principal features. Principal feature representation for both the static and dynamic background pixels is investigated. A novel learning method is proposed to adapt to both gradual and sudden "once-off" background changes. The convergence of the learning process is analyzed and a formula to select a proper learning rate is derived. Under the proposed framework, a novel algorithm for detecting foreground objects from complex environments is then established. It consists of change detection, change classification, foreground segmentation, and background maintenance. Experiments were conducted on image sequences containing targets of interest in a variety of environments, e.g., offices, public buildings, subway stations, campuses, parking lots, airports, and sidewalks. Good results of foreground detection were obtained. Quantitative evaluation and comparison with the existing method show that the proposed method provides much improved results.

No abstract

No abstract

No abstract

The first-order high frequency surface wave radar (HFSWR) cross section of the ocean surface is derived for the case of the transmitting and receiving antenna being mounted on a floating, but otherwise fixed, ocean platform. It is assumed that the sway component of the platform or barge motion is responsible for observed differences in the cross section compared to that for the fixed antenna case. Based on earlier work, a general expression for the bistatically received first-order electric field, which consists of a two-dimensional spatial convolution, is presented and reduced to integral form. Then, it is assumed that the surface can be described by a Fourier series whose coefficients are zero-mean Gaussian random variables, and from there the analysis proceeds for the backscatter case. The integrals are taken to the time domain, with the source field being that of a barge-mounted omnidirectional vertically polarized pulsed dipole antenna. Subsequently, the first-order monostatic radar cross section is developed and found to consist of Bessel functions. Simulation results for the new cross section are also provided to show the effects of barge motion under different sea states and operating frequencies. It is seen that the results have important implications in the application of HFSWR technology to ocean remote sensing.