Open AccessMaking data count
Author(s) -
Z.J. Len,
Stanbury,
D. M. Suckling,
Ashraf M. ElSayed
Publication year - 2015
Publication title -
proceedings of the new zealand weed and pest control conference/new zealand plant protection/proceedings of the ... national weeds conference/proceedings of the new zealand weed control conference/proceedings of the new zealand plant protection conference
Language(s) - English
Resource type - Journals
eISSN - 0370-2804
pISSN - 0370-0968
DOI - 10.30843/nzpp.2015.68.5862
Subject(s) - software , data collection , computer science , process (computing) , task (project management) , object (grammar) , variety (cybernetics) , artificial intelligence , computer vision , systems engineering , engineering , statistics , operating system , mathematics
The plethora of hightech data collection methods available to the modern biologist presents opportunities but also challenges While tools such as automated camera traps can vastly increase the efficiency of data collection and storage allowing for the collection of massive amounts of photographic data extracting quantitative data from these images manually can be a timeconsuming and tedious task Prototype counting software currently in development at Plant Food Research that can be used to automatically process large numbers of photographs or scanned images is presented The user trains the software to recognise objects to be counted such as insects or seeds allowing the software to be used across different image types qualities and conditions The abilities of the counting software to extract object counts from a variety of images is discussed and data presented comparing its accuracy to manual counting in different situations The viewer is also given an opportunity to try the counting software in a handson demonstration