Artificial intelligence based data processing algorithm for video surveillance to empower industry 3.5

• Published in: Computers & industrial engineering Vol. 148; p. 106671
• Main Authors: Nguyen, Minh T., Truong, Linh H., Tran, Trang T., Chien, Chen-Fu
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.10.2020
• Subjects: Artificial intelligence; Background modeling; Convolution neural networks (CNN); Industry 3.5; Region of interest (ROI); Video surveillance (VS); Region of interest (ROI); Convolution neural networks (CNN); Artificial intelligence; Video surveillance (VS); Industry 3.5; Background modeling
• ISSN: 0360-8352; 1879-0550
• DOI: 10.1016/j.cie.2020.106671

•AI based algorithm for data processing for camera surveillance is proposed.•The developed solution can reduce data transmission significantly.•The solution can empower smart manufacturing via camera surveillance.•Simulation results have validated practical viability of this approach. Nowadays, the demand of camera surveillance systems (CSS) has been increasingly adopted in various industries for smart manufacturing. However, the increase of utilizing CSS will pose many drawbacks in capacity storage and overload the transmission bandwidth. Focusing on the needs in real settings, this study aims to develop a novel algorithm based on artificial intelligence (AI) for data processing. Artificial intelligence (AI) is very helpful to process a large number of videos recorded by the CSS, while computer vision algorithms can also be employed to detect abnormal behaviors or noticeable objects, thus reducing the manpower. Since applications of AI for handling the above problems consume a lot of computational resources. This paper proposes a method to solve the above CSS issues. The idea is that focus on processing the valid background and moving object in the scene. After that, it will be transmitted to the server sides for further purposes. Indeed, the proposed method significantly reduces data transmission and storage and also improves the performance. The experimental results show that suggested method reduces storage capacity up to 80% and shows promising performance in which the number of calculations is reduced several times at the sever side compared to existing methods. Towards this end, the study proposes a method to solve the above drawbacks. It would be considered to apply for Industry 3.5 which is a mixture strategy in between the best practices of Industry 3.0 and to-be Industry 4.0.