ability is 
rked) ob- 
traced by 
entage of 
| percent- 
hm to all 
)% minus 
nd by the 
nd ETISEO 
vas espe- 
om N270 
ction and 
ie. digital 
tracking 
based on 
ectrums. 
ection of 
forming 
atiotem- 
esting on 
itlined. 
Igorithm 
Vision 2, 
nance of 
on 12(1), 
ies anal- 
://WWW- 
the Insti- 
ters. Int. 
g optical 
second- 
pp. 85- 
es/pets.html. 
spective. 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B3, 2012 
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia 
  
  
  
| Video name | Frame dimensions | Detection Type | Precision | Recall | FPS | 
PETS-2001-SEQ1-CAMI 768x576 Moving objects 80% (8/10) 100% (8/8) | 110 
PETS-2001-SEQ1-CAM2 768x576 Moving objects 88% (8/9) 100% (8/8) | 109 
PETS-2006-S1-T1-C3 720x576 Moving objects | 74% (26/35) | 85% (24/28) | 115 
ETISEO-VS2-BE-19-C2 768x576 Moving objects 100% (4/4) 100% (4/4) | 110 
ETISEO-VS1-AP-5-C5 720x576 Moving objects 88% (8/9) 100% (6/6) | 124 
ETISEO-VS1-AP-5-C7 720x576 Moving objects 100% (9/9) 100% (7/7) | 124 
ETISEO-VS2-BC-17-C1 640x480 New/diss. objects 66% (2/3) 100% (2/2) | 142 
ETISEO-VS1-BC-12-C1 640x480 New/diss. objects | 100% (1/1) 100% (1/1) | 144 
  
  
  
  
  
  
  
  
  
Figure 4: Sample frame of PETS-2001-SEQ1-CAMI video sequence. Objects, tracked on this frame: walking man (7), moving car (5) 
Table 1: Video analysis algorithms testing results 
  
and parking car (3). 
563