Robot-assisted, source-camera-coupled multi-view broadband imagers for ubiquitous sensing platform

Non-destructive sensing and imaging using electromagnetic waves is crucial for industrial quality assurance, especially within the expanding Internet of Things (IoT). Current methods often lack the versatility to handle complex, arbitrarily located targets. This research aims to develop a ubiquitous sensing platform capable of comprehensive broadband monitoring independent of the environment. The platform needs to address several key challenges: flexible multi-view imaging with reflective and transmissive modes using uncooled broadband photo-absorbent thin films; built-in miniature photo-sources within the flexible imaging framework; and robot-assisted, high-speed, omni-directional monitoring for location independence. Previous work has advanced imaging modalities across the electromagnetic spectrum (visible light, infrared, terahertz, millimeter-wave), and flexible imagers enable multi-functional sensing on 3D surfaces. However, integrating these features into a portable, robot-assisted system remains a significant hurdle. This study addresses these challenges by developing a novel platform that combines these functionalities.

The researchers reviewed existing technologies in non-destructive sensing and imaging, highlighting advancements in various electromagnetic wavebands and flexible imager designs. They noted progress in functional unmanned remote sensing robotics, high-usability miniature photo-sources, flexible multi-view stereoscopic photo-imagers, and highly efficient uncooled broadband photo-absorbent materials. The thermo-phototronic effect, combining thermoelectric and photoelectric effects, was also considered, but its full integration into a ubiquitous sensing platform had not been achieved previously. The paper's authors identify a gap in the literature: a lack of integrated systems possessing flexible switching between reflective and transmissive modes, built-in miniature photo-sources, and robot-assisted operation for location-independent high-speed monitoring.

The study developed a robot-assisted multi-view broadband photo-monitoring platform with switchable reflective and transmissive modes. The core technology uses freely attachable photo-thermoelectric (PTE) devices based on flexible carbon nanotube (CNT) films. These films offer high mechanical strength, flexibility, and broadband photo-absorption, making them suitable for PTE applications. To improve photo-detection sensitivity, a multifaceted device design was implemented, including liquid-coated N-type chemical-carrier doping on P-type SWCNT films. The authors optimized the device design through steady-state thermal simulations using ANSYS, maximizing the temperature gradient across the CNT film channels. Selective suction CNT-film patterning and chemical-carrier doping techniques were used to create the photo-imagers. Silver-nanowire-based stretchable electrodes were employed. 3D printing and UV laser processing facilitated the integration of high-output-power miniature photo-sources into the compact sensing modules. The noise equivalent power (NEP) of the CNT-film based photo-imager was evaluated. Two types of modules were developed: a reflective multi-view stereoscopic capsule PTE imager and a transmissive multi-view stereoscopic PTE endoscope. The performance of these modules was tested on various industrial components (beverage bottles, water and gas pipes), and a multi-layered columnar object was used to demonstrate remote hierarchical image extraction using multi-frequency band sensing. Finally, the system was integrated with a robot-assisted multi-axis movable arm to conduct high-speed, omni-directional inspections on a miniature model of a defective winding road-bridge.

The researchers achieved a minimum NEP of 8.57 pWHz^-1/2 for the CNT-film flexible broadband photo-imager, comparable to state-of-the-art solid-state devices. The reflective multi-view stereoscopic capsule imager successfully detected various defects (metallic impurities, breakages, scratches) in beverage bottles through non-destructive inspection. The multi-frequency band (sub-THz and NIR) sensing enabled remote hierarchical image extraction from a multi-layered columnar object, demonstrating the ability to image both outer and inner structures non-destructively. The transmissive multi-view stereoscopic PTE endoscope also demonstrated effective non-destructive inspection of water and gas pipes, showcasing its potential for post-disaster assessment. The integration of miniature photo-sources via 3D printing and UV laser processing created a portable, 360°-view stereoscopic imager, capable of omni-directional imaging at 10 Hz. The robot-assisted multi-axis movable arm, equipped with the integrated photo-source and imager, successfully performed non-destructive unmanned, remote, high-speed, omni-directional inspection of a miniature aerial defective winding road-bridge.

The results demonstrate the successful development of a versatile, robot-assisted, ubiquitous sensing platform. The platform's ability to perform non-destructive inspections across various industrial components, coupled with its location independence and high-speed capabilities, significantly advances the field of non-destructive testing. The use of multi-frequency band sensing allows for detailed hierarchical image extraction, providing unprecedented insight into complex structures. The integration with a robotic arm provides accessibility to challenging environments previously inaccessible to traditional methods. The findings address the limitations of existing methods by combining previously disparate technologies into a single, efficient system. The portability and automation of the system offer significant advantages in terms of safety, efficiency, and cost-effectiveness.

This study successfully demonstrated a novel robot-assisted multi-view broadband photo-monitoring platform with switchable modes. The platform, built around a photo-thermoelectric design using flexible carbon nanotube films, achieved high sensitivity and versatile imaging capabilities. Future work could focus on optimizing scan speed, developing a comprehensive broadband spectral database, and exploring methods for detecting inner defects in various materials. Collaborations with studies on roll-to-roll printing, flexible high-density pixel integration, and AI-based image processing could further enhance the platform's performance and applications.

One limitation is the need to quantify the trade-off between scanning speed and image quality. A larger pixel configuration could improve speed, but this requires further investigation. A more extensive broadband spectral database of various materials would enhance the reliability of the system. Further research is also needed to optimize the detection of internal defects using different frequency bands.