Abstract
Infrared (IR) video sequences Improvement is a difficult task because of many factors for example low dynamic range, noise and Non-Uniformity (NU) effect. The NU is a time-dependent noise that appears due to the lack of sensor equalization. So, it is necessary to apply Non-Uniformity Correction (NUC) in video sequences based on the original sight. This paper suggests two schemes for the IR video sequences improvement. The first proposed scheme is a scene-based NUC technique based on Histogram Matching (HM). The noise effect on each picture in the sequence causes some drift in the histogram of that image. Hence, this technique depends on the HM notion for correction the histogram of each image in the sequence relied on the histogram of the original sight. As the original sight is free from the thermal noise effect. The second proposed scheme relies on Contrast Limited Adaptive Histogram Equalization (CLAHE) based NUC. The metrics are adapted to assessment are entropy, contrast factor, average gradient, and edge intensity. From the obtained results, it is clear that the two suggested techniques have succeeded in getting the best results in removing the noise and enhancing image quality. Simulation results prove that the second scheme is better than the first algorithm from point views the performance metrics.