Related Experiment Video
Updated: Feb 17, 2026
04:05
Noninvasive EEG Recordings from Freely Moving Piglets
Published on: July 13, 2018
7.8K
Depth-Based Detection of Standing-Pigs in Moving Noise Environments
Jinseong Kim1, Yeonwoo Chung2, Younchang Choi3
1Department of Computer and Information Science, Korea University, Sejong City 30019, Korea. skykeeop@korea.ac.kr.
Sensors (Basel, Switzerland)
|November 30, 2017
Summary
This study presents a novel method for real-time, depth-based detection of standing pigs, even with significant nighttime "moving noises." The cost-effective approach achieves high accuracy, enabling continuous pig tracking.
Area of Science:
- Computer Vision
- Animal Science
- Agricultural Technology
Background:
- Real-time detection of individual pigs is crucial for 24-hour tracking in commercial farms.
- Nighttime "moving noises" in depth images present a significant, previously unreported challenge for pig detection.
Purpose of the Study:
- To develop a robust, real-time method for detecting standing pigs using depth imaging.
- To address and overcome the issue of "moving noises" in depth images during nighttime surveillance.
Main Methods:
- Applied spatiotemporal interpolation to effectively remove "moving noises" from depth images.
- Utilized undefined depth values around pigs for accurate detection.
- Employed a low-cost Kinect depth sensor for cost-effectiveness.
Main Results:
- Achieved high accuracy (94.47%) in detecting standing pigs.
- Demonstrated effectiveness even with up to 50% of the depth image occluded by noise.
- The method operates in real-time without requiring time-consuming processes.
Conclusions:
- The proposed depth-based detection method is effective and accurate for identifying standing pigs at night.
- Cost-effectiveness and real-time performance make this technique suitable for commercial pig farm surveillance.
- Successfully overcomes the challenge of "moving noises" for improved animal monitoring.
