A Machine Vision-Based Method for Monitoring Broiler Chicken Floor Distribution

Abstract

The proper spatial distribution of chickens is an indication of a healthy flock. Routine inspections of broiler chicken floor distribution are 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 the 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 bird distribution in the images, the pen floor was virtually defined/divided into drinking, feeding, and rest/exercise zones. As broiler chickens grew, the images collected each day were analyzed separately to avoid biases caused by changes of body weight/size over time. About 7000 chicken areas/profiles were extracted from images collected from 18 to 35 days of age to build a BP neural network model for floor distribution analysis, and another 200 images were used to validate the model. The results showed that the identification accuracies of bird distribution in the drinking and feeding zones were 0.9419 and 0.9544, respectively. The correlation coefficient (R), mean square error (MSE), and mean absolute error (MAE) of the BP model were 0.996, 0.038, and 0.178, respectively, in our analysis of broiler distribution. Missed detections were mainly caused by interference with the equipment (e.g., the feeder hanging chain and water line); studies are ongoing to address these issues. This study provides the basis for devising a real-time evaluation tool to detect broiler chicken floor distribution and behavior in commercial facilities.

Keywords: animal behaviors; broiler chicken; health and welfare; precision farming.

Figure 1

Experimental setup for broiler chicken image data collection.

Figure 2

A top view of a pen and zone definition. The red box (1) represents the drinking area: the center of the nipple drinker is the center of the drinking area and its width is defined as one body length of a three-week-old broiler chicken; the yellow circle (2) represents the feeding area: the center of the tube feeder is the feeding area center; the radius of feeding zone is the tube feeder radius plus the body length of a three-week-old broiler chicken; and the cyan box (3) represents the overall detection area of the pen, so any area not included in drinking and feeding zones will be considered as the rest/exercise zone.

Figure 3

Comparison between different color spaces ((a) LAB, (b) RG, (c) RB, and (d) GB) in the classification process.

Figure 4

The image processing steps. (a) Nontarget area removal; (b) Generation of binary image; (c) Morphological corrosion operation and background removal.

Figure 5

The visualization efficiency of three different methods (a, b, and c represent images taken on d18, d24 and d30, respectively).

Figure 6

Input–output correlation in the newly developed BP neural network model for identifying broiler chicken floor distribution (a,b,c, and d correspond to the training set, validation set, test set and overall results, respectively. The horizontal axis Target is the actual number of chickens; and the vertical axis Output is the number of BP model output; “O” represents the input data of the model; “Fit” is the fitting relationship between input and output; “Y = T” means the training output value equal to the target value).

Figure 7

Number of chickens identified and their distribution determined with the newly developed BP model (chickens were three weeks of old in (a), (b) and 4 weeks old in (c), (d); cyan rectangles represent target extraction zone by BP method, and back rectangles represent missed target area (chicken); yellow numbers indicate true chicken number and behind indicate BP model recognized chicken).

Figure 8

Broiler chicken distribution in feeding and drinking zones as identified by the BP model. (a) total chicken tracked and identified; (b) chicken distribution in drinking and feeding zones. The big red rectangle is the drinking zone, and the small red rectangles in this zone indicate the detected broiler chickens. The yellow circle is the feeding area, and the yellow rectangles in the zone indicate detected broiler chickens.