To effective integrate the bright-field and dark-field defect image information to solve the missed and false inspection problems due to incomplete information, according to the characteristic of strip steel surface defect image, a novel image fusion method is proposed. Firstly, the two original images are decomposed using Non-sub-sampled contourlet transform (NSCT) separately so that the low frequency coefficients and high frequency coefficients are obtained. To retain the important defect information, different methods are adapted to fuse high and low frequency coefficients. Finally, the fused image is obtained by performing the inverse NSCT on the combined coefficients. The experimental results show that the quantity of image information and image quality are highly improved. This method can effectively solve the problem of missed and false detection.