BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Automated modelling of industrial plants - Eva Agapaki\, Engineeri
 ng Department\, University of Cambridge
DTSTART:20180215T131000Z
DTEND:20180215T140000Z
UID:TALK96928@talks.cam.ac.uk
CONTACT:Lorena Escudero
DESCRIPTION:The cost of modelling existing industrial facilities is curren
 tly considered to counteract the benefits\nof the model in managing and re
 trofitting the facility. 90% of the modelling cost is typically spent\non 
 labour for converting point cloud data to the final model\, hence reducing
  the cost is only\npossible by automating this step. Previous research has
  successfully validated methods for\nmodelling specific object types such 
 as cylinders. Yet modelling is still prohibitively expensive.\nDuring this
  talk\, the most important object types of industrial facilities will be i
 dentified by ranking\nthem according to their frequency of appearance and 
 the man-hours required for modelling in a\nstate of the art software\, Edg
 eWise. This work is the first to rank objects according to their priority\
 nfor automated modelling. These are straight pipes\, electrical conduit an
 d circular hollow sections\nand constitute more than 80 % of industrial pl
 ants on average. This is significant because state-ofthe-\nart practice ha
 s achieved semi-automated cylinder detection saving 64 % of their manual\n
 modelling time for the case studies investigated. Automated detection and 
 semantic classification\nmethods for the recognition of the abovementioned
  objects will be analyzed.
LOCATION:The Richard King Room\, Darwin College
END:VEVENT
END:VCALENDAR