Enhancing CLAHE with Interval-Valued Fermatean Fuzziness for Robust Low-Light Image Enhancement
Hongpeng Wang, Duanfa Wang, Zhiqin Wang, Chenglong Li · 2024
Image processing, as the foundation of computer vision, plays a crucial role in various important fields such as aerospace, medical imaging, and military operations. However, the images captured under extreme lighting conditions often suffer from issues like uneven brightness and contrast, loss of details, and color degradation, posing challenges for subsequent visual tasks such as semantic analysis, image segmentation, and object recognition. The mainstream Contrast-Limited Adaptive Histogram Equalization (CLAHE) method fails to achieve effective results when pixel intensities approach 0, and parameter selection remains a significant challenge. To address the problem of image quality degradation under extreme lighting conditions and the limitations of CLAHE, this paper proposes an Interval-Valued Fermatean Fuzzy CLAHE (IVFF-CLAHE) method for low-light image enhancement. The proposed method designs several Fermatean fuzzy generators and interval Fermatean generators to perform fuzzy processing and optimize the gray level distribution of the image. Subsequently, it applies fuzzy CLAHE to the values in the HSV color space for histogram equalization. For parameter selection, a modified Dynamic Weight Particle Swarm Optimization (DW-PSO) algorithm is utilized to improve the efficiency of optimal parameter search. Through comparative experiments with existing low-light image enhancement algorithms based on fuzzy principles, the proposed method demonstrates significant improvements in subjective contrast and brightness, with better preservation of details and texture, resulting in more natural and clean images. Objectively, it achieves the highest performance compared to the evaluated techniques in terms of information richness, contrast, brightness, colorfulness, and image quality metrics.