Improved ViBe algorithm based on multi-frame combined with adaptive threshold

Abstract For moving human target detection under video surveillance, the traditional vibe algorithm is often disturbed by environmental changes, resulting in “ghosts”, incomplete results and “holes” body. An improved based on multi frame combined with adaptive threshold proposed. Firstly, sample set of expanded to 24 fields reduce possibility pixel misclassification; Secondly, historical queue introduced, initialization background model without ghost obtained according change foreground pixels time domain; Finally, distance determination dynamically adjusted using gray characteristic convergence divergence, update factor calculated introducing two fixed parameters optimize rate model. The experimental show that not only suitable for general scenes, but also dynamic complex it can accurately detect targets, eliminate obtain complete avoid appearance “holes”, improve accuracy detecting targets environments, enhance robustness algorithm.