Stationary object detection using multi-mode background modelling
Abstract STATIONARY OBJECT DETECTION USING MULTI-MODE BACKGROUND MODJELLING Disclosed is a method (1201) of processing a video stream, the method comprising selecting (1220), for each visual element in a set of visual elements derived from the video stream, a matching mode model (350) from a plurali...
Gespeichert in:
Hauptverfasser: | , |
---|---|
Format: | Patent |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | Abstract STATIONARY OBJECT DETECTION USING MULTI-MODE BACKGROUND MODJELLING Disclosed is a method (1201) of processing a video stream, the method comprising selecting (1220), for each visual element in a set of visual elements derived from the video stream, a matching mode model (350) from a plurality of candidate models, determining (1230) a 10 representative age measure from the matching mode models, comparing (1240) the representative age to a first threshold value; and determining (1260) if the set of visual elements is stationary, based on the result of the comparing step. 847749 f 1147235_1 280208 10'1 i/15 start e 00 Fig. 120 Are there unprocessed visual elements in the input frame? Yes 130 Select an input visual element and a corresponding vsual element model From the isual element model. select a mode model matching the input isual element 140 see Fig. 4) Determine whether the mode is foreground or background. Apply connected component --- analysis to foreground/background segmentation End Fig. i 847749 /1147237_1 |
---|