Abstract:

This paper proposes an image fusion algorithm based on the multi-scale multi-direction Shearlet transform to improve the performance of multi-focus image fusion.First,the Shearlet transform was used to decompose the registered original images,thus the low frequency sub-band coefficients and high frequency sub-band coefficients of different scales and directions were obtained.The coefficients of high regional energy were taken as the low frequency sub-band coefficients to yield the high frequency sub-band coefficient.The fusion coefficients are determined by the multiple discriminant method combining the regional energy and regional definition as well as regional variance.Finally,the fusion image is obtained by the inverse Shearlet transform.The results show that the subjective visual performance and objective evaluation index of this algorithm are superior to other fusion algorithms.

Key words: Shearlet transformation;regional definition;regionalenergy;regional variance;multiple discriminant method