Enhancing Transportation’s Images Quality using Anisotropic Diffusion Kuwahara Filtering for Noise Reduction and Edge Preservation

Abstract

This paper proposes a method to improve the segmentation of traffic images after edge preservation using anisotropic diffusion filtering. Anisotropic diffusion filtering is applied to the traffic images to preserve important edges and structures while reducing noise. However, the resulting images may still contain artifacts that can affect the accuracy of segmentation tasks such as object detection and lane delineation. To address this issue, we introduce a novel post-processing technique to refine the segmentation results. The proposed method leverages the edge-preserving properties of anisotropic diffusion filtering to enhance the boundaries of segmented objects and remove spurious artifacts. Experimental evaluations conducted on a variety of traffic scenes demonstrate the effectiveness of the proposed approach in improving the segmentation accuracy compared to traditional methods. The results highlight the potential of integrating edge preservation techniques with segmentation algorithms for enhanced performance in traffic image analysis tasks. The proposed methodology uses various phases such Noise Reduction, Edge Preservation, Improved Segmentation, Enhanced Visibility, Adaptive Filtering. Step by step each algorithm different operations on transportation’s Images. Using Anisotropic diffusion filtering and Two-Directional Two-Dimension Principal Component Analysis (2D2PCA) reduces more than 30% original image size and also preserve edges more than 95% of original images. The reduced size images are very useful for future work.