Related Experiment Video
Updated: May 2, 2026
Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment
Published on: May 7, 2019
Image dataset for foreign object detection in iron ore conveyor belt systems
Frederico L Martins de Sousa1, Thiago E Alves de Oliveira2, Saul E Delabrida Silva3
1Programa de Pós-Graduação em Instrumentação, Controle e Automação de Processos de Mineração (PROFICAM) - Universidade Federal de Ouro Preto (UFOP) and Instituto Tecnológico Vale (ITV) , Minas Gerais, Brazil.
This study introduces a new dataset of high-speed videos capturing iron ore on a conveyor belt, useful for detecting foreign objects in mining operations. The data aids in developing advanced computer vision systems for industrial material transport.
Area of Science:
- Materials Science
- Computer Vision
- Robotics
Background:
- Automated detection of foreign objects in industrial material transport is crucial for operational efficiency and safety.
- Existing datasets may not adequately represent the complexities of real-world mining environments, such as iron ore transportation.
Purpose of the Study:
- To present a novel, high-speed video dataset of iron ore on a conveyor belt.
- To facilitate the development and comparative evaluation of image-based detection algorithms for anomalies in industrial settings.
- To support research in computer vision for material handling and transportation.
Main Methods:
- Acquisition of high-speed (120 fps) video recordings using an NVIDIA Jetson TX2 with an OV5693 camera.
- Utilized a laboratory-scale conveyor belt (35 cm x 1.10 m) transporting hematite and contaminants (wood, plastic).
- Data organized into subsets for normal operation, anomalies, and object types, with frames extracted and sorted.
Main Results:
- A comprehensive dataset of synchronized video and object information was created.
- The dataset captures both regular iron ore flow and the presence of distinct foreign objects.
- Data is structured to allow for diverse applications in object detection and classification.
Conclusions:
- The presented dataset provides a valuable resource for advancing computer vision techniques in mining and industrial material transport.
- Enables robust testing and benchmarking of anomaly detection algorithms in a controlled, yet representative, environment.
- Facilitates future research into real-time monitoring and control systems for bulk material handling.
