Object Detection of PCB faults and PCB inspection

Development Insights:

• CoreFragment Role : ML development, firmware development
• Camera Setup: We first set up and calibrate all the cameras by adjusting their settings like frame size, color, exposure, and more. We use special boards with patterns to help align the cameras and make sure they capture clear, accurate images.
• Vision Profiles: After calibration, we choose which cameras to use for taking pictures. We apply the settings from calibration to these cameras, so they work together to capture and stitch images accurately.
• Project Creation: We create a project where we set up how images should be transformed and analyzed. This helps in inspecting objects, like a PCB, step by step by focusing on specific areas (ROIs).
• Object Detection: For detecting defects like bubbles, we label the objects in images, train a model to recognize them, and then test the model on new images to ensure it works correctly.