Video pre-processing for the improvement of an automated lameness detection system for dairy cows

C. E.B. Romanini, S. Viazzi, T. Van Hertem, A. Schlageter-Tello, I. Halachmi, K. Lokhorst, C. Bahr, D. Berckmans

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Computer vision and image analysis are techniques with potential for use in animal health and welfare monitoring systems, since they are relatively inexpensive and do not interfere with the animals' environment. In an automatic lameness detection system for dairy cows, in particular, the performance of the model relies on the quality of the 3D input videos and therefore video pre-processing is required. This paper describes video pre-processing routines for automatic filtering of "video quality" and "cow's walking speed". Decision tree algorithms J.48 were used to classify model performance. The results show that the "video quality" filtering model correctly classified 75.7% of cases while the "cow's walking speed" achieved 91.3% accuracy. Detailed results are quantified in the form of a confusion matrix. Further optimisation of the 3D automated lameness detection system through pre-processing is needed before it can be used on commercial dairy farms.

Original languageEnglish
Title of host publicationPrecision Livestock Farming 2013 - Papers Presented at the 6th European Conference on Precision Livestock Farming, ECPLF 2013
PublisherKatholieke Universiteit Leuven
Pages479-487
Number of pages9
ISBN (Print)9789088263330
StatePublished - 1 Jan 2013
Externally publishedYes
Event6th European Conference on Precision Livestock Farming, ECPLF 2013 - Leuven, Belgium
Duration: 10 Sep 201312 Sep 2013

Publication series

NamePrecision Livestock Farming 2013 - Papers Presented at the 6th European Conference on Precision Livestock Farming, ECPLF 2013

Conference

Conference6th European Conference on Precision Livestock Farming, ECPLF 2013
Country/TerritoryBelgium
CityLeuven
Period10/09/1312/09/13

Keywords

  • 3D image
  • Cow's features extraction
  • Filtering
  • System implementation
  • Video quality

ASJC Scopus subject areas

  • Animal Science and Zoology

Fingerprint

Dive into the research topics of 'Video pre-processing for the improvement of an automated lameness detection system for dairy cows'. Together they form a unique fingerprint.

Cite this