3D reconstruction and defect pattern recognition of bonding wire based on stereo vision

Non-destructive detection of wire bonding defects in integrated circuits (IC) is critical for ensuring product quality after packaging. Image-processing-based methods do not provide a detailed evaluation the three-dimensional wire. Therefore, method 3D reconstruction and pattern recognition based on stereo vision, which can achieve non-destructive proposed. The contour features wires other electronic components depth image analysed to complete wires. Especially filter noisy point cloud obtain an accurate surface, segmentation spatial surface feature (SFD) was SFD extract more distinct from during process. Furthermore, defect process, directional discretisation descriptor with multiple local normal vectors designed combines global describe variation trends structural experimental results show that wires, average accuracy 96.47%, meets production requirements detection.