Accuracy of Broiler Activity Index as Affected by Sampling Time Interval

Area of Science:

• Animal Science
• Agricultural Engineering
• Ethology

Background:

• Broiler activity index (AI) is crucial for welfare and productivity assessment.
• Previous studies used varying image sampling intervals for AI, but accuracy impacts remain unclear.

Purpose of the Study:

• To evaluate the effect of different sampling time intervals (0.04s to 300s) on broiler AI accuracy.
• To analyze AI variations across bird ages, locations (feeder, drinker, open areas), and times of day.

Main Methods:

• Utilized a ceiling-mounted camera for top-view video capture of broilers.
• Calculated broiler AI using image sequences at diverse sampling intervals.
• Analyzed AI data considering bird age, location, and time of day.

Main Results:

• The 0.04s sampling interval provided the highest broiler AI, capturing maximum motion detail.
• Broiler AI at longer intervals (1-99%) deviated from the 0.04s baseline.
• A 0.2s interval offered acceptable accuracy with 80% reduced computational load.
• AI generally decreased with age, but increased post-week 4 near drinkers.
• Location-specific AI patterns emerged: higher in open areas (weeks 1-4) and feeder/drinker areas (weeks 5-7).

Conclusions:

• Sampling time interval critically influences broiler AI accuracy.
• Broiler AI exhibits significant temporal and spatial variations in commercial settings.