A Novel Pavement Distress Image Filtering Algorithm

摘要

Combined the Nonsubsampled Contourlet Transform (NSCT) with compressed sensing de-noising theory, this study proposed a compressed sensing image filtering algorithm in the NSCT transform domain. This algorithm adopted the transform domain filtering method. Firstly, the NSCT was used to do multi-scale and multi-directional transform on the input noisy pavement distress Image which was mapped to the transform domain. Then the compressed sensing theory was used to filter the high-frequency sub-band coefficients. The experiment results show that this algorithm improves the effectiveness and timeliness of the pavement distress image filtering.