Abstract Longer‐wavelength (LW) computer vision (CV) monitoring techniques hold great potential as non‐destructive testing tools for daily necessities and industrial products, utilizing effective optical properties. Carbon nanotube (CNT) film photo‐thermoelectric image sensors facilities, such CV systems as broadband wavelength range operations over existing narrow‐band detectors at comparable sensitivity with them owing to the inherent advantageous material properties. Nevertheless, insufficient pixel integration of CNT image sensors requires extensive and time‐costing optical spatial scanning for CV measurements, limiting practical inspection applications at real‐time industrial sites. To this end, this work demonstrates LW computed tomography (CT, a representative CV method) monitoring system employing a 2D pixel‐array integrated CNT film PTE camera sheet and multi‐wavelength photo‐sources. The camera fabrication process utilizes air‐jet dispense‐printing and achieves high‐yield integration of CNT film pixels (20 mm‐15 mm‐sq. for pixel area, 60 pixels total). This work then performs structure reconstruction and material identification of a visibly opaque multi‐layered composite 3D object according to optical properties in the 976 nm–10.3 µm wavelength range by LW CT with the device. The presented CNT film camera completes its LW operation in ≈11 min for each wavelength, a 36.9 % time reduction compared to previous CT systems using 1D pixel‐array sensors.
Kubota et al. (Sun,) studied this question.