Akademik Veri Yönetim Sistemi
5th International Conference on Informatics and Software Engineering, IISEC 2026, Ankara, Türkiye, 5 - 06 Şubat 2026, ss.445-450, (Tam Metin Bildiri)
Thermal imaging technology provides reliable imaging independent of environmental conditions in many critical areas and significantly contributes to object detection and scene interpretation processes in these critical areas. Although these sensors have advantages in terms of low light independence and wide operating range, difficulties arise during object detection and feature extraction from thermal images due to high noise levels, thermal drift, multispectral light, low signal-to-noise ratio (SNR), and sensor-induced artifacts. This situation both hinders accurate detection and reduces image quality. This study aims to compare the noise reduction performance of basic image processing filters (Gaussian, Median, Bilateral, Wiener, and Non-Local Means (NLM) filters) on thermal images to overcome these challenges. Controlled levels of thermal noise were added to the thermal images in the dataset using Gaussian, salt-and-pepper, and power spectrum-based methods. The filters were then applied to the noisy thermal images. The improvement performance of the filters was evaluated using the Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index Measure (SSIM) quantitative metrics. The study revealed which filter type yielded more effective results under different thermal sensor conditions.