New trends on moving object detection in video images captured by a moving camera: A survey

Abstract

This paper presents a survey on the latest methods of moving object detection in video sequences captured by a moving camera. Although many researches and excellent works have reviewed the methods of object detection and background subtraction for a fixed camera, there is no survey which presents a complete review of the existing different methods in the case of moving camera. Most methods in this field can be classified into four categories: modelling based background subtraction, trajectory classification, low rank and sparse matrix decomposition, and object tracking. We discuss in details each category and present the main methods which proposed improvements in the general concept of the techniques. We also present challenges and main concerns in this field as well as performance metrics and some benchmark databases available to evaluate the performance of different moving object detection algorithms.

Panoramic Robust PCA for Foreground–Background Separation on Noisy, Free-Motion Camera Video

“…where u i and v i are the singular vectors of X, and we denote its singular values by σ i . Problem (40) seeks the best approximation of the latent low-rank signal matrix L by an optimally weighted combination of estimates of its left and right singular vectors. The truncated SVD (of rank r) and SVT are both feasible points for (40).…”

“…Problem (40) seeks the best approximation of the latent low-rank signal matrix L by an optimally weighted combination of estimates of its left and right singular vectors. The truncated SVD (of rank r) and SVT are both feasible points for (40). Indeed, the truncated SVD corresponds to choosing weights w i = σ i 1{i ≤ r} and SVT with parameter τ ≥ σ r+1 corresponds to w i = ( σ i −τ ) + .…”

“…Indeed, the truncated SVD corresponds to choosing weights w i = σ i 1{i ≤ r} and SVT with parameter τ ≥ σ r+1 corresponds to w i = ( σ i −τ ) + . However, (40) can be solved in closed-form (see [49]), yielding the expression…”

“…is the empirical mass function of the noise-only singular values of Y with q = min(m, n). By Theorem 1, OptShrink r ( X) asymptotically solves the oracle denoising problem (40).…”

Moving Object Detection from Moving Camera Image Sequences Using an Inertial Measurement Unit Sensor

“…Zheng et al proposed a vehicle detection method using image stripe features [18], and an image registration method was used by Hsieh et al to detect a vehicle in video sequences [19]. Yazdi and Bouwmans [20] compared various moving object detection methods with a focus on the case of a moving camera. Some studies utilized training techniques to detect objects from video [21][22][23].…”