A New Edge Fusion Algorithm Based on Finite Element Method

Edge detection based on multiscale wavelet transform is one of the important algorithms in images edge detection. In order to give attention to both of strongly anti-noise in large scale and accurately locate edge in small scale, a new edge detection method based on fields´ fusion was developed. Firstly, the effective edge envelope of different edge model based on compact, symmetry and smooth function were discussed, Secondly, in the range of effective edge envelope of different edge model, a new edge detection, that was finite element methods edge fusion algorithms, was adopted for local edge of different scale edge field. The validity of the algorithm was proved by experiments on natural data.