Zhengxu Liu , Wei Peng , Chaoyuan Wang , Guoming Li , Boyu Ji , David Casey , Ning Lu , Yunxiang Zhao
{"title":"Automatic swine body measurements via two-dimensional imaging system and correlation analysis with conformation scoring","authors":"Zhengxu Liu , Wei Peng , Chaoyuan Wang , Guoming Li , Boyu Ji , David Casey , Ning Lu , Yunxiang Zhao","doi":"10.1016/j.atech.2025.101113","DOIUrl":null,"url":null,"abstract":"<div><div>The manual measurement of pig body dimension is not only labor-intensive and fallible but also causes stress to the animals. Moreover, the subjective and ambiguous visual assessment of pig body conformation by Chinese pig traders consistently impacted the sustainable profitability of pig farms. Considering long-term robustness, this study aimed to develop an automated measurement system for pig body measurements based on two-dimension (2D) machine vision. For the first time, pig traders from Chinese five regions (North, East, MidSouth, WestSouth, WestNorth) were invited to an objective survey of body conformation scoring. They were also invited to rate pig videos (Good-3, Medium-2, Bad-1) obtained by the system using specially developed software. Subsequently, correlation analysis and principal component analysis were conducted. The correlation coefficient for abdominal circumference (R<sup>2</sup>=0.85) and chest circumference (R<sup>2</sup>=0.84) by 2D imaging reflected strong agreements, improved by 0.13 (18.1 %) and by 0.15(21.7 %) respectively, compared with the similar 3D system. The 2D imaging also indicated a moderate discrepancy in the prediction of the above circumferences. The 2D imaging system should be more robust further considering the cost and maintenance difficulty in pig farm. In the survey, traders from all regions assigned consistently high importance on Body Length and Height. The variance was mainly from Fur&Color and Belly, where the traders from Mid-South put a higher value to Fur&Color, and East’s traders paid higher attention to Belly. Our study provided the first-ever objective data basis behind pig body conformation by Chinese traders. It was concluded that the belly_height by the 2D imaging showed the most significant positive impact on trader conformation scoring for most regions except West-South. Whereas the belly_y_size (diameter of the digitally defined belly part) by the 2D imaging exerted the most significant negative effect on conformation scoring in most regions, only that it did have negative effect in East and WestSouth though under the significant level. Figuratively, pigs perceived as ‘higher’ with ‘smaller' bellies showed stronger market preference among traders in China.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"12 ","pages":"Article 101113"},"PeriodicalIF":5.7000,"publicationDate":"2025-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Smart agricultural technology","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2772375525003466","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"AGRICULTURAL ENGINEERING","Score":null,"Total":0}
引用次数: 0
Abstract
The manual measurement of pig body dimension is not only labor-intensive and fallible but also causes stress to the animals. Moreover, the subjective and ambiguous visual assessment of pig body conformation by Chinese pig traders consistently impacted the sustainable profitability of pig farms. Considering long-term robustness, this study aimed to develop an automated measurement system for pig body measurements based on two-dimension (2D) machine vision. For the first time, pig traders from Chinese five regions (North, East, MidSouth, WestSouth, WestNorth) were invited to an objective survey of body conformation scoring. They were also invited to rate pig videos (Good-3, Medium-2, Bad-1) obtained by the system using specially developed software. Subsequently, correlation analysis and principal component analysis were conducted. The correlation coefficient for abdominal circumference (R2=0.85) and chest circumference (R2=0.84) by 2D imaging reflected strong agreements, improved by 0.13 (18.1 %) and by 0.15(21.7 %) respectively, compared with the similar 3D system. The 2D imaging also indicated a moderate discrepancy in the prediction of the above circumferences. The 2D imaging system should be more robust further considering the cost and maintenance difficulty in pig farm. In the survey, traders from all regions assigned consistently high importance on Body Length and Height. The variance was mainly from Fur&Color and Belly, where the traders from Mid-South put a higher value to Fur&Color, and East’s traders paid higher attention to Belly. Our study provided the first-ever objective data basis behind pig body conformation by Chinese traders. It was concluded that the belly_height by the 2D imaging showed the most significant positive impact on trader conformation scoring for most regions except West-South. Whereas the belly_y_size (diameter of the digitally defined belly part) by the 2D imaging exerted the most significant negative effect on conformation scoring in most regions, only that it did have negative effect in East and WestSouth though under the significant level. Figuratively, pigs perceived as ‘higher’ with ‘smaller' bellies showed stronger market preference among traders in China.