A New Multi-Focus Image Fusion Techniquefor an Efficient Surveillance

摘要

The paper presents a new multi-focus based image fusion method for enhancing the image processing component of the central computer system in the visual sensor network (VSN) for efficient surveillance and motion tracking. The proposed method is based on decomposing the input multi-focus images using stationary wavelet transform (SWT). The wavelet coefficients are fused on the basis of the selected threshold value. The threshold value is selected on the basis of various parameters such as loss of correlation, luminance distortion, and contrast distortion. On the basis of these parameters and residual information of the input images, the fusion of wavelet coefficients is performed. The proposed fusion approach is analyzed on the basis of qualitative and quantitative measures. It is also compared to some of the traditional and non-traditional methods. The proposed method is capable of being incorporated in the real-time application of surveillance in VSN.