Why did RAYHAWK choose to use a computer vision system?

Written by: Andrew Tremblay, Lead Software Developer

A common question we get asked during our demonstrations is why RAYHAWK decided to use computer vision to influence the operation of our machine. To understand why is to start with an understanding of what computer vision is!

Computer vision seeks to understand and automate tasks that a human visual system can do. These tasks consist of acquiring, processing, analyzing, and understanding to make informed decisions. In a railcar loading environment, these decisions can be formed by knowing where the latches and lids are on a railcar, identifying whether they are open or closed, and if they are able to be operated on.

RAYHAWK’s system has been trained against thousands of images of latches and lids (and counting) in various environments utilizing the latest in machine learning technologies. Using scalable cloud technologies, RAYHAWK can be fed curated images from our client’s environment to improve our computer vision. This also allows RAYHAWK to be adaptable to changes, allowing RAYHAWK to be flexible to address each client’s operational environment and equipment. This ensures RAYHAWK’s computer vision meets the demands of our client, whether that is within a dusty grain elevator or out in a windy and snowy open prairie.

Under the hood, the RAYHAWK computer vision is powered by two depth sensing cameras that enable us to see in three dimensions. These cameras give us real world measurements to know how far away an object of interest is within our gantry system. Our cameras are enclosed to ensure they can operate in dust or snow and ensure the integrity of the camera lens. Our camera system has been chosen to utilize equipment that is commercially available in the event of repair, they can be promptly replaced as needed or upgraded as newer models become available.

The choice for computer vision allows RAYHAWK to be highly adaptable to every client!