Skip to main content
ARS Home » Research » Publications at this Location » Publication #375149

Research Project: Characterizing Antimicrobial Resistance in Poultry Production Environments

Location: Location not imported yet.

Title: A machine vision-based method for monitoring broiler chicken floor distribution

Author
item GUO, YANGYANG - University Of Georgia
item CHAI, LILONG - University Of Georgia
item AGGREY, SAMUEL - University Of Georgia
item Oladeinde, Adelumola - Ade
item JOHNSON, JASMINE - University Of Georgia
item ZOCK, GREGORY - University Of Georgia

Submitted to: Sensors
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/1/2020
Publication Date: 6/3/2020
Citation: Guo, Y., Chai, L., Aggrey, S., Oladeinde, A.A., Johnson, J., Zock, G. 2020. A machine vision-based method for monitoring broiler chicken floor distribution. Sensors. http://doi.org/10.3390/s20113179.
DOI: https://doi.org/10.3390/s20113179

Interpretive Summary: Proper distribution of chickens in the house is an indication of a healthy flock. Routine inspection of broiler chickens’ floor distributions is done manually in commercial grow-out houses every day, which is labor intensive and time consuming. This task requires an efficient and automatic system that can monitor chicken’s floor distributions. In this study, a machine vision-based method was developed and tested in an experimental broiler house. For the new method to recognize birds’ distribution on images, the pen floor was virtually defined/divided as drinking, feeding, and rest/exercise zones. After training and testing the model using >2000 images collected at different chicken ages. The model showed high accuracy (> 94 %) in detecting the number of chickens and their distribution in the house floor. This study provides the basis/proof-of-concept for developing an automatic system to monitor flock health in a commercial production system.

Technical Abstract: Proper distribution of chickens in the house is an indication of a healthy flock. Routine inspection of broiler chickens’ floor distributions is done manually in commercial grow-out houses every day, which is labor intensive and time consuming. This task requires an efficient and automatic system that can monitor chicken’s floor distributions. In the current study, a machine vision-based method was developed and tested in an experimental broiler house. For the new method to recognize birds’ distribution on images, the pen floor was virtually defined/divided as drinking, feeding, and rest/exercise zones. As broiler chickens grew, images collected from each individual day are analyzed separately to avoid the bias caused by change of body weight/size over days. About 7000 different chicken areas/profiles extracted from images collected from 18 to 35 days of age to build the BP neural network model for floor distribution analysis and another 200 images were used to validate the model. Results showed that the identification accuracy of birds’ distribution in the drinking and feeding zones was 0.9419 and 0.9544, respectively. The correlation coefficient (R), mean square error (MSE), and mean absolute error (MAE) of the BP model was 0.996, 0.038, and 0.178, respectively, in analyzing broiler distribution. The missed detection was mainly caused by interferences of equipment (e.g., the feeder hanging chain and water line) on chickens’ floor images. Studies are ongoing to address those interference issues. This study provides the basis for devising a real-time evaluation tool to detect broiler chickens’ floor distribution and behaviors in commercial facilities.