研究目的
To develop an augmented reality (AR) aided smart sensing technique for in-line condition monitoring of IGBT wafers, enhancing defect detection and user interaction through non-destructive testing methods.
研究成果
The proposed AR aided smart sensing system effectively enables in-line condition monitoring of IGBT wafers with high throughput and non-destructive capabilities. It enhances defect detection through thermal imaging and AR visualization, potentially improving yield management and production efficiency. Future work should focus on expanding defect types and integrating with existing electrical testing methods.
研究不足
The technique is in an infant stage and cannot replace electrical testing entirely. It is currently limited to subsurface defects like metallization issues and may not detect other defect types such as pinholes or dislocations. The system requires further optimization for broader applicability and real-time implementation in industrial settings.
1：Experimental Design and Method Selection:
The study uses electromagnetic thermography with inductive heating to detect subsurface defects in IGBT wafers. Signal processing algorithms like independent component analysis (ICA) and multichannel morphological component analysis (MMCA) are applied for defect identification. Infrared-visible-fusion (IVF) and augmented reality (AR) techniques are integrated for enhanced visualization.
2：Sample Selection and Data Sources:
An IGBT wafer model IRGC75B120KB from International Rectifier is used, which has undergone static electrical testing. Defects include subsurface gate runner metallization issues.
3：List of Experimental Equipment and Materials:
Equipment includes a function generator for pulse signals, an induction heater for inductive heating, an FLIR SC7000sc IR camera for thermal imaging, a PC for data processing, and software like Matlab for image analysis. Materials include the IGBT wafer and cooling systems for the induction coil.
4：Experimental Procedures and Operational Workflow:
The process involves acquiring visible and IR images of the wafer under inductive heating, processing thermal images with ICA/MMCA to reveal defects, performing feature mapping between IR and visible images using SURF and RANSAC algorithms, fusing images with wavelet fusion, and embedding fused data into a 3D point cloud for AR overlay.
5：Data Analysis Methods:
Thermal responses are analyzed using ICA and MMCA to separate defect layers. Image registration and fusion use SURF for feature detection and RANSAC for matching, with wavelet decomposition for IVF. Statistical methods assess defect identification and system performance.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
