Senalp, Fatih MehmetOrhan, BatuhanCeylan, Murat2023-05-302023-05-3020231380-75011573-7721https://doi.org/10.1007/s11042-022-14169-0https://hdl.handle.net/20.500.13091/3958Thermal cameras can be used safely on living things because thermal camera systems provide harmless and contactless imaging. However, thermal cameras are high-cost systems, which limits their widespread usage. Therefore, low-cost thermal cameras facilitate the use of thermal imaging in different areas. Also, these thermal cameras can create blurry thermal images with low detail information. The purpose of this work is to provide an alternative to the high-cost problem and to create a system that can be easily accessed via the internet. Here, it is important to use super resolution techniques on these low quality images. In this study, a new dataset was created using two different thermal cameras. This dataset was obtained in different way from the datasets used in traditional image enhancement implementations. Here, the low-resolution thermal face images are obtained by means of the Flir One Pro (c) thermal camera, which can be easily integrated into the smartphone. The high-resolution (ground truth) thermal face images are obtained with the Variocam HD (c) thermal camera, which is an HD (high definition) thermal imaging system. In addition, the TSRGAN+ deep network model is proposed as a new approach for super resolution application on the new dataset. The obtained results were compared with state-of-the-art models using peak signal to noise ratio (PSNR) and structural similarity index measure (SSIM) image quality metrics. When the results were evaluated, it was observed that the success performance of the proposed model was superior to other models. Here, considering the PSNR and SSIM values, it was seen that the proposed model achieved approximately 0.5 dB and 6% more successful results than other models, respectively. After these studies, the proposed super resolution model was run in the cloud environment and the system can be easily accessed from anywhere through the developed Android interface software.eninfo:eu-repo/semantics/closedAccessCloud environmentThermal imagingSuper resolutionDeep learningNew datasetsSuperresolutionArticle10.1007/s11042-022-14169-02-s2.0-85142160710