Computer Vision Project Implementation in Elixir for Construction Sites

Walid Salah from Evercam presented at ElixirConf EU 2024 on their project involving end-to-end computer vision implemented in Elixir to manage and monitor construction sites. Salah's team uses cameras to capture live feeds and other tools for enhancing site productivity and communication. The hardware setup typically includes a Raspberry Pi or Jetson Nano, a 4G router, and power management systems. The software side employs Elixir and Phoenix for managing highly distributed systems, with additional support from libraries like Membrane and Evision for video processing and OpenCV functionalities. Salah highlights challenges in hardware management and emphasizes the simplicity of adding functionality to the system, such as motion detection and AI model implementation using Bumblebee. The project, named ex_mvr, is open source and welcomes contributions from the community.