Segmentation of infrared image using fuzzy thresholding via local region analysis

According to the characteristic of infrared images, a target extraction method based on fuzzy thresholding is proposed for vehicle target images. A membership function composed of the modified bi-modality and the inverse S adjacency is used. In order to meet the requirement of real time, the bi-modality measure is calculated only in the boundary regions so that the execution time can be greatly reduced. The inverse S adjacency function is used to take full advantage of the position information of the pixels in the reference region. Our method is processed as follows. First, we calculate the membership values consisting of the modified bi-modality and the new adjacency. And then we perform the fuzzy thresholding and the post-processing to extract the precise target from the background. In order to evaluate the performance of our method, the proposed method is compared with other segmentation methods. The results of experiments prove that the presented algorithm is fast and has a good segmentation performance.