The need of the project was to develop an integrated quality control system that inspects products on a production line and discards any defective product (surface defect or dimensional defect) using a 4 DOF Delta robot.

Problem Description:

Develop an integrated quality control system that employs a 4-degrees-of-freedom (4-DOF) Delta robot to inspect products for defects and subsequently sorts the defective and non-defective products. The system utilizes a Convolutional Neural Network (CNN) algorithm, specifically MobileNetV2, for defect detection through machine learning. The non-defective products continue down the conveyor line, while the defective ones are accurately identified and picked aside by the Delta robot.

Workflow:

1. On system initialization, the system variarbles are first calibrated to their default values to ensure optimal operational conditions.

2. The conveyor is started and the system awaits the arrival of the product. 

3. Once the product is placed on the conveyor and detected by the first IR sensor, the system turns on the camera feed and captures the image of the product. 

4. The image is then run through the machine learning model and the prediction is published to the respective topic. 

5. If the product is found to be non-defective, the system allows the product to continue down the conveyor line regardless of the second IR sensor's input. 

6. If the product is found to be defective, the system waits for the triggering of the second IR sensor to pick it out from the conveyor line.

Benefits

a. High Accuracy:

MobileNetV2 ensures accurate defect detection, minimizing false positives and negatives.

b. Speed and Efficiency:

The 4-DOF Delta robot enables rapid and precise sorting of defective products, maintaining production efficiency.

c. Real-Time Processing:

The system operates in real-time, ensuring immediate identification and sorting of defects.

d. Adaptability:

The system can be adapted to various product types and sizes, making it versatile for different manufacturing scenarios.

e. Reduced Human Intervention:

Automation reduces the need for manual inspection, minimizing errors and increasing overall efficiency.

Conclusion

The integrated quality control system with a Delta robot and MobileNetV2 CNN algorithm provides a robust solution for defect detection and sorting in a manufacturing environment. The combination of advanced image processing, machine learning, and robotic automation ensures a reliable and efficient production line.