Processing Algorithm
Software-based image analysis operates to identify anomalies in radiographic or optical scans of electronic assemblies without manual operator intervention. Industrial platforms use automated defect recognition to process transmission images of solder joints to locate internal voids, bridging, and solder ball alignment issues. The process runs during high-throughput inline inspection where human inspection represents a cycle-time bottleneck.
Scan Requirement
High-resolution digital detectors capture the gray-scale density maps needed for the evaluation algorithms to function. Reliable automated defect recognition relies on consistent image contrast and a minimal signal-to-noise ratio. Small variations in substrate thickness or component density can cause false calls if the scan parameters are not calibrated against a known standard.
Defect Classification
Decision matrices categorize detected anomalies based on pre-programmed acceptance criteria from IPC standards. The system calculates the area percentage of each void within a ball grid array to determine if it exceeds the allowable limit. Rejection signals route the affected boards to a rework conveyor automatically, while the system logs the coordinate data for process correction.
Continuous feedback helps engineers isolate drifting stencil printer or reflow oven parameters before defects occur. By tracking these trends, assemblies avoid escape failures that escape manual visual inspection.
