AI IMAGE PROCESSING

A key project executed by Azel IT for one of the world's largest and most reputed commercial vehicles manufacturer involved designing an AI-based software solution that would integrate and work with drone hardware.

The contracted solution was meant to command a drone to fly autonomously around chosen commercial vehicles and capture a series of high-definition photographs. These images were then to be processed using AI-based algorithms to detect and flag a variety of damages to the vehicle, including marks, dents, or scratches.

The project commenced in early 2022 and included a dynamic team of robotics engineers, software developers, a cloud architect, a project manager, as well as a business development manager.

The project during the course of its development involved several in-person interactive sessions with the customer team, and multiple demonstrations using the drones that resulted in obtaining the necessary feedback crucial for progress. There were also several brainstorming sessions held with the customer on addressing the various ground level challenges faced by the project.

Eventually, the project was successfully completed in the last quarter of 2022, resulting in the creation of a dynamic solution that comprised all of the following:

• An enhanced system that could provide instructions to a drone to autonomously navigate identified commercial vehicles and capture high-definition images in a variety of angles and then upload it to cloud-storage.
• A robotics-based component that used AI-based algorithms to scrutinize these stored images (based on its pre-learning using large data sets), to identify cases of vehicles that involved varying degrees of damages, with an accuracy of over 90%.
• The dynamic creation of 3-dimensional images of the vehicle using advanced image processing, which showed a comparative view of the vehicle from its original end-of-line to its current status.