Chinese Optics Letters, Vol. 2, Issue 5, 05268 (2004)
A new noise erosion operator for anisotropic diffusion
Chao Cai*, Mingyue Ding, Chengping Zhou, and Tianxu Zhang
- Institute for Pattern Recognition and Artificial Intelligence, State Education Commission Key Lab for Image Processing and Intelligent Control, Huazhong University of Science and Technology, Wuhan 430074
A noise erosion operator based on partial differential equation (PDE) is introduced, which has an excellent ability of noise removal and edge preservation for two-dimensional (2D) gradient data. The operator is applied to estimate a new diffusion coefficient. Experimental results demonstrate that anisotropic diffusion based on this new erosion operator can efficiently reduce noise and sharpen object boundaries.