Multi Computer Vision-Driven Testing Platform: Structural Reconstruction and Material Identification with Ultrabroadband Carbon Nanotube Imagers

Abstract

While non-destructive inspections of daily necessities and industrial products require material composition identifications and structural reconstructions of composite multi-layered 3D objects, the development of analyzing systems that fully satisfy those performances is still insufficient. For such situations, this work synergizes permeable material identifications with carbon nanotube (CNT) film photo-thermoelectric (PTE) imagers in millimeter-wave (MMW)–visible light (Vis) bands and conventional Vis-computer vision (CV) approaches mainly used for analyzing outer shapes of 3D objects. Specifically, this work demonstrates multi-CV-driven non-destructive structural reconstruction techniques by ultrabroad MMW–Vis band multi-wavelength optical monitoring using the CNT film PTE imager. Here, the CNT film PTE imager ultra-broadly performs photo-detection operations over the existing sensors in comparable sensitivities with the conventional narrowband devices at each region. Thus, the above imager facilitates utilizing ultrabroadband multi-wavelength photo-irradiation as playing an essential role in the presenting non-destructive CV system. Following such device performances, the MMW–Vis multi-CV system with the CNT film PTE imager first speculates spatial locations of each part for targets by visual hull measurements, then restores their detailed structures by tomography approaches. By incorporating these device and system properties, this work finally achieves non-destructive composition-identifying structural reconstructions for composite multi-layered 3D objects.