摘要
Traditional Sobel algorithm is deeply influenced by Gaussian noise,therefore,before boundary extraction,mean filter should be done. But the filtering process is always over-smooth images'details of certain directions,so that images'edges will not be extracted correctly. Aiming at this problem,this paper puts forward a detection algorithm based on edge-preserving characteristics,by matching edge mould of different directions to definite edge preserving directions. Instead of the mean filter process,this algorithm improves the performance of traditional algorithms,and provides the simulation results. The experiment results prove that this algorithm preserves more images'edge information when canceling noise.
Traditional Sobel algorithm is deeply influenced by Gaussian noise, thereibre, beibre boundary extraction, mean fiher should be done. But the filtering process is always over-smooth images＇ details of certain directions, so that images＇ edges will not be extracted correctly. Aiming at this problem, this paper puts forward a detection algorithm based on edge-preserving characteristics, by matching edge mould of different directions to definite edge preserving directions. Instead of the mean filter process, this algorithm improves the performance of traditional algorithms, and provides the simulation results. The experiment results prove that this algorithm preserves more images＇ edge information when canceling noise.
基金
Sponsored by the Nation Nature Science Foundation of China(Grant No.61201237)
the Nature Sciece Foundation of Heilongjiang Province of China(Grant No.QC2012C069)
the Fundamental Research Funds for the Central Universities(Grant No.HEUCFZ1129,HEUCF130810,HEUCF130817)
