Because of light changes and uneven reflections, the asphalt pavement image noise is rich, and the traditional crack segmentation methods are easy to lose the crack boundary. Therefore, proposes asphalt pavement image segmentation method based on optimized Markov Random Field. Comparing multiple wavelet domain threshold denoising algorithms, the BayesShrink wavelet threshold method is selected to preprocess the image to denoise. Meanwhile, comparing various initial segmentation methods, derived a adaptable initial segmentation for MRF segmentation method. The experimental results show that the initial segmentation will greatly reduce the noise interference before MRF segmentation, after BayesShrink denoising.
Asphalt Pavement Image Segmentation Method Based on Optimized Markov Random Field
22.10.2021
1534799 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Gaussian Markov random field based improved texture descriptor for image segmentation
British Library Online Contents | 2014
|A novel pixon-representation for image segmentation based on Markov random field
British Library Online Contents | 2008
|Color image segmentation using Markov random fields
IEEE | 1989
|