Vision analysis and prediction for estimation of pig weight in slaughter pens

Gang Jun Tu*, Erik Jørgensen

*Corresponding author af dette arbejde

Publikation: Bidrag til tidsskrift/Konferencebidrag i tidsskrift /Bidrag til avisTidsskriftartikelForskningpeer review

Abstract

The potential of using computer vision to estimate the weight of pigs has been demonstrated in several studies. In this paper, we propose a Vision Analysis and Prediction (VAP) system, which consisted of two main parts: (1) computer vision – a vision algorithm was developed to segment the pigs and identify which pig area could be used to estimate the corresponding weight; (2) statistical analysis – a statistical method for predicting pig weight using the area was presented, including BLUP model and linear regression as well as a prediction function. The results showed that our approach has successfully estimated the weight of growing pigs (approximate ranges from 20 to 105 kg) with an accuracy of weight estimation of 97.76% on average for the predicted mean weight at pen level. The proposed system does not require modifications to the building and shows a strong potential to be utilized in pig weight estimation in slaughter pens.

OriginalsprogEngelsk
Artikelnummer119684
TidsskriftExpert Systems with Applications
Vol/bind220
Antal sider9
ISSN0957-4174
DOI
StatusUdgivet - jun. 2023

Fingeraftryk

Dyk ned i forskningsemnerne om 'Vision analysis and prediction for estimation of pig weight in slaughter pens'. Sammen danner de et unikt fingeraftryk.

Citationsformater