A Framework for Segmentation of Thermal Images using Unsupervised Fuzzy Based Clustering Methods

Image segmentation is a primary task in lots of computer vision applications, including thermal image segmentation. Thermal images are typically characterized by high levels of noise and low contrast, which can make segmentation challenging. In this paper, we propose an outline for thermal image segmentation using unsupervised fuzzy-based clustering methods. The proposed framework consists of three main steps: preprocessing, clustering, and post-processing. In the preprocessing step, the thermal image is enhanced to improve the contrast and remove noise. In the clustering step, a fuzzy clustering algorithm is used to cluster the pixels in the image into different regions based on colors. In the post-processing step, the results of the clustering algorithm are refined to improve the segmentation accuracy. The proposed work utilized three fuzzy based methods as Fuzzy C-Means clustering (FCM), Possiblistic Fuzzy C-Means clustering (PFCM) and Modified Fuzzy C-Means clustering (MFCM) for the thermal image segmentation. The proposed framework is evaluated on a number of thermal images. The outcome demonstrates that the projected framework can achieve good segmentation accuracy, even in the presence of noise and low contrast. The proposed framework is also robust to changes in illumination and can be used to segment thermal images with different levels of details.