A Robust Software Sensor to Identify People in a Video Scene

• Arthur Hernandez Perez UERJ
• Karran Cardoso de Araujo Lemos UERJ
• Evandro Luiz Cardoso Macedo UERJ https://orcid.org/0000-0001-5430-2168
• Alexandre Sztajnberg UERJ https://orcid.org/0000-0003-2018-6386

DOI: https://doi.org/10.5753/sbrc_estendido.2025.6794

Abstract

The development of smart cities and the Internet of Things paradigm allows the creation of more efficient, sustainable and safe cities. Among the innovations, smart cameras stand out for identifying objects and people, enabling continuous monitoring and real-time data collection for different environments. This article presents a smart software sensor for counting people in video scenes, developed for Computer Sciences Lab. and integrated with the LCC-IoT application. It allows monitoring and generating alarms in specific situations, such as the presence of people outside of permitted hours, while preserving users’ privacy. The sensor was implemented to maintain continuous and concurrent operation across multiple cameras, controlling data acquisition and consistent generation of telemetry through semaphores, in addition to recording errors, making it robust. The solution uses the OpenCV and YOLO for image processing, MQTT and HTTP application protocols for sending telemetry.

Keywords: Software sensor, Internet of Things, Application protocol, Identifying people in a video scene

References

Ahmed, E. et al. (2016). Internet-of-things-based smart environments: state of the art, taxonomy, and open research challenges. IEEE Wireless Comm., 23(5):10–16.

Bradski, G. (2000). The OpenCV Library. Dr. Dobb’s Journal of Software Tools.

Camlytics (2022). People counting solution. Web. [link] [Access: 07/2022].

Chen, C.-C., Lin, H.-H., and Chen, O. T.-C. (2011). Tracking and counting people in visual surveillance systems. In 2011 IEEE ICASSP, pages 1425–1428.

Eclipse Paho (2018). Eclipse paho javascript client. Web page. [link].

Eocortex (2022). People counting. Web. [link].

LCC, UERJ (2025). Sensor de pessoas em uma cena. Web. [link].

Motta, R. C., Batista, T. V., and Delicato, F. C. (2024). The Intersection of the Internet of Things and Smart Cities: A Tertiary Study. JISA, 15(1):325–341.

Mueller, V. (2021). Non-maximum suppression. [link].

Redmon, J. (2013–2016). Darknet: Open Source Neural Networks in C. [link].

Redmon, J. and Farhadi, A. (2018). Yolov3: An incremental improvement. CoRR, arXiv, [link].

Yang, Gonzalez-Banos, and Guibas (2003). Counting people in crowds with a real-time network of simple image sensors. In 9th IEEE ICCV, pages 122–129 vol.1.