Multimodal Image Fusion with 3dIV/Fusion

Non-destructive reverse engineering (RE) of electronics (CCAs and/or PCBs) often relies on imaging techniques to capture both the internal and external details of the asset. For instance, PCBs pose a specific challenge in regards RE in that they are laminated structures comprised of several alternating layers of di-electric and conductive materials. The internal conducting layers (typically comprised of Cu) effectively hold the circuit design information and are connected to other internal circuit layers by conductive ‘holes’ (called vias) and finally to external layers and the electrical components (e.g. ICs, resistors, capacitors, etc.) mounted on the PCB outer layers. Thus, the circuit design information is not readily available through external optical inspection. Forming a clear picture of these internal connections requires non-destructive imaging techniques such as X-ray Computed Tomography (CT) or Terahertz Imaging.  However non-destructive techniques when applied alone often fail to capture certain critical information. That is each technique has its own advantages and oftentimes must be coupled together with other NDI or with optical information to gain a full picture of the data needed to perform high-fidelity non-destructive RE. Government engineers need a software tool to help them perform non-destructive reverse engineering of electronics assets that can effective combine all sources of information. The tool must be capable of addressing common problems in non-destructive electronics reverse engineering such as: a) dewarping 3d CT data of PCBs/CCAs, b) stitching multiple overlapping CT volumes into a single volume, c) correcting for imaging artifacts such as X-ray beam hardening, most importantly d) fusing image data from multiple modalities (e.g. X-ray, optical, Terahertz imaging) together into a cohesive view. Such a tool would allow users to extract 3d solid model (e.g. CAD-like models) of features/materials of interest based on individual user specifications. Recent advances in AI/ML are making multimodal image fusion, most commonly applied in medical fields, an attractive option for non-destructive RE of electronics.