Problem

Our client, a manufacturing company, was facing challenges with their quality control process. They were relying on manual inspection of products which was time-consuming and prone to errors. This resulted in a high number of defective products being shipped out to customers, leading to customer complaints and loss of revenue.

We were tasked with finding a solution that would automate the quality control process and reduce the number of defective products being shipped out.

Learn more about our solution here.

Solution

After analyzing the client's requirements, we decided to construct an object detection project using Yolov8, Ros2, and Gazebo. We used Yolov8 for object detection as it is one of the most accurate and efficient object detection models available. Ros2 was used for communication between different nodes in the system, while Gazebo was used for simulating the manufacturing environment.

We trained the model using a dataset of images provided by the client. The model was able to detect defects in products with an accuracy of over 95%. We then integrated the model into the client's production line, where it automatically inspected each product as it moved along the conveyor belt. Defective products were flagged and removed from the production line before they could be packaged and shipped out.

Impact

Reduced number of defective products by X%
Increased efficiency of quality control process by X%

"This project has had an incredible impact on our client's manufacturing process. They have seen a significant reduction in the number of defective products being shipped out, resulting in fewer customer complaints and increased revenue."

Team

Team Member #1 - Machine Learning Engineer
- Trained the Yolov8 model using the provided dataset
- Integrated the model into the production line using Ros2
Team Member #2 - Robotics Engineer
- Simulated the manufacturing environment using Gazebo
- Integrated the object detection system into the client's production line

Automating Quality Control with Yolov8, Ros2, Gazebo