Enhancing Image Clarity through a Comparative Analysis of Polarizing Filters and Homomorphic Filtering Algorithms
Sung Keun Cha, Si Woo Lee, Jae Wook Jeon · 2023
In the context of autonomous vehicles, camera-based lane recognition can be compromised by external factors like sunlight. This study evaluated two denoising methods: software-based denoising via the homomorphic filter algorithm and hardware-based denoising by attaching a polarizing filter in front of the camera lens. Images were analyzed under four conditions based on the combination of using the homomorphic filter and the presence of a polarizing filter. A comparative analysis revealed that a combined application of both filters achieved the highest Structural Similarity Index Measure (SSIM) and the lowest Mean Square Error (MSE), establishing it as the most effective denoising approach.