Circuit Board Grader and Classifier

---

Objective: Develop an automated system for identifying and classifying electrical components and materials mounted on the surface of a printed circuit board, and generate a 3D representation of the surface for component height information using only a single 2D image. The primary purpose of this system is for the segregation of components for material recovery, analytics for PCB grading, augmented or robotic sorting, and eventually automated demanufacturing.

Solution: Deployed an IntelSight-based vision system on a self-contained fixture. Utilized custom-trained neural network to mask board images with a height-map and segmentation by material type.

Status: IntelSight vision system can now accurately classify components on PCBs and generate a 3D mesh from a single 2D image. 

Future: Utilizing an IntelSort system to load boards into different post-processing stations based on their perceived composition, monetary value, and ideal method of material recovery.