Smart Camera Surveillance System: ITB Student Developers Achieved 2nd Place on the Schlumberger Agora Hackathon 2022

By Asep Kurnia, S. Kom

Editor Vera Citra Utami

BANDUNG, itb.ac.id — Two ITB students, Muhammad Sulthan Mazaya and Muhammad Naufal Aurora, with two other students from Institut Teknologi Sepuluh November (ITS), achieved 2nd place on the Schlumberger Agora Hackathon 2022 competition, Friday (23/09/2022). The contest was held by Schlumberger, an oil and gas company currently developing its new product, the Agora Gateway.

The students, Mazaya and Naufal, developed the Smart Camera Surveillance System (SCSS), a surveillance camera system for the purpose of mitigating accidents during projects. The team’s creation, named the Mjölnir Team, was a detector that can recognize personal protective equipment (PPE), fire blazes, even workers’ safety violation.

A bit distinguished from other IoT gateways, Agora Gateway can execute Edge Computing or pipeline data transformation storage directly from the sensor. The gateway works faster (low latency), safer, and more reliable as the contained data does not require processing on the cloud. In this hackathon, participants are expected to develop Edge Application, a pipeline data transformation that will be used in Agora Gateway.

SCSS consisted of two components, The Edge and The Cloud. The Edge is a machine learning model that will be implemented on the Agora Gateway, which will allow users to collect data from the sensor.

This data will be send through the MQ Telemetry Transport (MQTT) protocol and stored as well as displayed online and real time. When a date from the sensor is deemed dangerous, a warning email will be send to the registered email.

The students from ITB Engineering Physics and Petroleum Engineering asserted that the SCSS can reduce incidences through sending alerts before any accident occurs. In the long run, the implementation of SCSS will also profits financially.

“If we look deeper, the feature includes a flare detector model, utilizing machine learning model with the U-Net architecture and EfficientNet-B4 as the encoder; as well as a equipment detector model and breach safe zone detector model using YOLOv5,” explained Mazaya.

Reporter: Amalia Wahyu Utami (Engineering Physics, 2020)
Translator: Firzana Aisya (Bioengineering, 2021)