Digital image analysis has potential – on farm and in the auction market
Take Home Message: On-farm digital image analysis can be used to accurately predict animal liveweight, fat class and conformation therefore improving selection for slaughter.
Live animal digital image analysis systems are feasible under common on-farm management conditions and can be used to predict some important slaughter characteristics.
“But further studies are required to extend the digital image analysis (DIA) database to cover additional animal groups, such as bulls and heifers and different breeds, as well as more divergent fat and conformation grading combinations,” the SAC’s Jimmy Hyslop told delegates at this year’s British Society of Animal Science annual conference.
“More robust prediction equations could then be developed to cover this wider range of commercial cattle. The potential of live animal DIA systems in an auction market environment should also be examined,” he added, sharing the findings of his team’s work to assess the feasibility of an on-farm DIA system to predict a number of characteristics in beef steers at slaughter.
Previous work has suggested that dimensional data obtained from DIA has the potential to predict the slaughter liveweights of finished beef cattle. Both farmers and cattle buyers may benefit from being able to select animals to meet slaughter specifications more accurately on-farm.
During the trial, top, side and rear-view images of finished beef steers were obtained within one day of the animals going for slaughter from two pens, each containing twelve Aberdeen Angus x Limousin (AAxLIM) and twelve Limousin x Aberdeen Angus (LIMxAA) finishing beef steers during late summer/early autumn 2007 where slaughter characteristics were also determined.
Continuous images were taken using three multiplexed, high-resolution, colour digital cameras mounted directly above the animal looking straight down (top view), to the side of the animal looking straight on (side view) and above and to the rear of the animal looking at a 45o angle (rear view). The cameras were positioned in a ‘cubicle’ leading to the water trough in the corner of each straw-bedded pen in which the animals were housed.
Images were stored on computer and selected static images were subsequently processed to obtain simple dimensional data using image processing software by an experienced operator. Dimensions collected for the body of each of the animals included length, width, area and depth.
“And we encountered no major problems with the practical operation of the DIA system in an on-farm environment,” said Dr Hyslop.
“Significant predictive relationships were established for all slaughter characteristics and the figures represent an ability to predict the actual slaughter value with an error of +/- 1.0 - 7.9%.”
summary (pdf)
To view proceedings of all summaries presented at the Annual Conference 2009 http://www.bsas.org.uk/Publications/Annual_Conference_Proceedings/ To view all Powerpoint presentations http://www.bsas.org.uk/Members_Area/ For further information contact: BSAS on 0131 445 4508 or bsas@sac.ac.uk
