Precision livestock farming cattle identification based on biometric data tarek gaber


Published on

Published in: Business, Technology
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Precision livestock farming cattle identification based on biometric data tarek gaber

  1. 1. PRECISION LIVESTOCK FARMING: CATTLE IDENTIFICATION BASED ON BIOMETRIC DATA By Tarek Mahmmed Gaber, PhD Faculty of Computers and Informatics Suez Canal University 08/04/2014 – Faculty of Agriculture, Ismailia, Egypt
  2. 2. Scientific Research Group in Egypt
  3. 3. Overview  Introduction  Current work (cattle identification)  Proposed System  Experimental Results  Conclusion
  4. 4. Introduction: What is the Problem  Worldwide demand for meat is expected to increase with >40% in next 15 years  Health: Relationship between animal health and healthy food  Animal welfare  Economic importance  Others …….  Source: [TIVO-project]
  5. 5. Introduction: Livestock Farm Livestock farming in the past The farmer spends some time for noticing and monitoring
  6. 6. Livestock Farming Today Experts do audio- visual scoring by visiting farms and looking to 0behavior of animal.
  7. 7. Precision Livestock Farming (PLF) “ Management of livestock farming by continuous automated real-time monitoring/controlling/tracing of production/reproduction, health and welfare of livestock.”
  8. 8. Benefit for Farmers from PLF  By automating the farming process, the farmer is able to receive real-time information on his livestock, so can:  Manage and optimise animal production and welfare in a fast and accurate way.
  9. 9. Research Area in PLF  Examples of research points of PLF  Monitoring feed times  Feed in-take  Condition scoring  Real-time analysis of sound
  10. 10. Animal Tracing: Animal Identification  Radio Frequency Identification (RFID) is currently the most well used method for animal identification.  Ear tag or as a microchip the skin.  Problems:  Invasive and religious matters
  11. 11. Animal Biometric-based Solutions  Can produce accurate results of cattle recognition in real production conditions.  Do not need to attach any additional elements with or within the animals.  Comply with most countries legal rules (e.g. the current EU legislation) for beef traceability in slaughterhouses.
  12. 12. Unique Features of Cattle  Breeds muzzle pattern or nose print has been investigated and proven to be unique for each cattle  It is then concluded that muzzle print is similar to the human's fingerprint
  13. 13. Precision Livestock Farming: Cattle Identification based on Biometric Data  Training phase  Collecting all training muzzle print images.  Extracting the features  Representing each image by one feature vector.  Applying a dimensionality reduction (e.g, LDA) to reduce the number features in the vector  Testing phase  Collecting the muzzle print image,  Extract the features  Feature vector is projected on LDA space.  Applying machine learning techniques for classifying the test feature vector to decide whether the animal is identified or not).
  14. 14. Results Accuracy results (in %) when applying our proposed algorithm using different training images
  15. 15. Conclusion  Precision Livestock farming could  Increase the efficiency and sustainability for farming and livestock production by monitoring (individual) animals  Our proposal approach for cattle identification could  Deliver quantitative information and complete traceability of livestock in the food chain.  Image-based identification could be a promising non- intrusive method for cattle identification
  16. 16. Thanks