Inspector Digital Scrap challenge
Description
The dataset comprises details of the challenge shared by GERDAU on the HeroX crowdsourcing platform, outlining the complexities of automating scrap material inspection. It also encompasses the specifics of our proposed solution, "INSPECTOR: Artificial Intelligence Scraping Steel," elucidating the innovative approach adopted. Additionally, the dataset includes visual representations such as images capturing the solution in action, providing a holistic view of the proposed methodology within the context of the challenge.
Files
Steps to reproduce
Initially, we conducted a comprehensive analysis of the existing literature to understand the current state of related technologies and identify best practices. Additionally, we conducted field research at relevant industrial facilities, where we could directly observe the processes of handling and inspecting recyclable materials. During these visits, we used a variety of instruments, such as high-resolution cameras and measuring devices, to collect visual and metric data on the materials in question. Regarding laboratory experiments, we prototyped a scanner process on a scale, while training a model using darknet and YOLO with videos of the original scrap selection process. We presented the results in a functional prototype made available online, a video of the solution and laboratory tests, and a document, attached in this dataset. Video: https://youtu.be/pQS2zkJ9KkY