Computerized tomography technique for reconstruction of obstructed temperature field in infrared thermography

Infrared thermography is a rapid, non-invasive and full-field technique for non-destructive testing and evaluation (NDT&E). With all the achievements on IR instrumentation and image processing techniques attained, it has been extended far beyond simple hot-spot detection and becomes one of the most promising NDT&E techniques in the last decades. It has achieved increasing acceptance in different sectors include medical imaging, manufacturing component fault detection and buildings diagnostic. However, one limitation of IR thermography is that the testing results are greatly affected by object surface emissivity. Surface with various emissivities may lead to difficult discrimination between area of defect and area with different emissivity. Therefore, many studies have been carried out on eliminating emissivity, for example, the time derivative approach, lock-in processing and differential contrast measurements. In these methods, sequence of themo-data/images are recorded and being processed in order to eliminate differences of emissivity. Another problem of IR thermography is that any obstruction may limit stimulations and imaging which leads to the observation of unclear defect image. To solve this problem, this paper proposes an algorithm based on the principle of computerized tomography which permits the reconstruction of unavailable/partially available temperature distribution of the affected area using the measured surrounding temperature field. In the process, a set of imaginary rays are projected from many different directions across the area. For each ray, integration of the temperature derivatives along the ray is equals to the temperature difference between the boundary points intercepted by the ray. Therefore, a set of linear equations can be established by considering the multiple rays. Each equation expresses the unknown temperature derivatives in the affected area in terms of the measured boundary temperature data. Solution of the set of simultaneous equations yields unknown thermal distribution in the area which needs to be reconstructed. Based on the proposed computerized tomography reconstruction (CTR) technique, deviated temperature data due to missing temperature data hidden by obstacle can be reconstructed. With further development, this technique may have the potential to be applied in the reconstruction of any smooth physical fields like phase information in optical techniques.