Application of nonlinear filter in agriculture image de-noising

Agriculture image de-noising is the important precondition for feature extraction and recognition. It is very important for computer-aided crop disease and insect pests´ recognition to get a high quality image which is preprocessed. Traditional linear filter used for de-noising may lead to blurry edge; consequently, it lost the key factor of the image-edge feature. To solve this problem, in this paper, two kinds of nonlinear filters such as anisotropie diffusion filter and bilateral filter are proposed, also, they were applied to agriculture image de-noising. To verify the feasibility and effectiveness of our algorithm, the proposed algorithm is compared with linear filter. Experimental results show that our method is superior to linear filter algorithm.