Image Stitching Based on Discrete Wavelet Transform and Slope Fusion

The fusion algorithm of traditional image stitching does not fully consider the differences of the clarity of the two images, and the conventional DiscreteWavelet Transform algorithm would blur the image when applied to image stitching. Owing to these, an improved method based on Discrete Wavelet Transform and Slope Fusion is proposed. The proposed algorithm firstly performs Haar wavelet transform on the image to be fused to obtain a low-frequency component and multiple high-frequency components. Subsequently, the Slope Fusion method is used for the obtained low-frequency component and the sub-regional Slope Fusion method is used for the high-frequency components. Finally, the fused image is obtained by using the Inverse Discrete Wavelet Transform for the new low-frequency component and high-frequency components. The proposed algorithm can retain the information of direction and detail while taking full account of differences in image sharpness, all of those benefits help improve the quality of the fused image effectively. The experimental results show that the proposed algorithm can make the fused image clearer and objectively enhance multiple fusion indicators of the fused image.