Wavelet Packet Denoising for Pavement Surface Cracks Detection

Abstract

Removing noise from pavement images to improve the quality of images is an important step of pavement image processing. A novel noise reduction method based on wavelet packet for pavement images is presented. Firstly, the image is transformed into a wavelet packet domain. Then using the semi-soft threshold to process wavelet packet coefficients can keep important image information and remove noise. Last, we reconstruct the image. Performance of the method has been tested on actual pavement images. From the experiment results, we see that the method effectively reduces the noise while preserving the edges in the original image.