Abstract: Blur can be defined as unsharp image region that happen through camera motion, object motion and through image acquisition. This paper introduced restoration of Gaussian blurred images by inverse filtering. This filter is done on the spatial domain for image restoration. The key objective of this work is fusion process based on peak signal to noise ratio measure. The main idea is compared PSNR value between sub image with other sub image of another image. In the beginning convolve inverse filter with image to get on the blurred image. After that convolve the blurred image with inverse of filter that remove blur from image to obtain on the deblurred image. Finally fuse the optimal sub images based on peak signal to noise ratio measure that has higher PSNR.

Materials and Methods
The key idea of applying blur on the spatial domain is convolving inverse filter with image to get on the blurred image. Then apply inverse of this filter that remove blur from image and it convolved with blurred image to obtain on the deblurred image. After that has been calculated PSNR value between sub image of one with other sub image and then summed PSNR of each sub image to get on the optimal sub image. Finally select PSNR value of optimal sub image based on peak signal to noise ratio measure that has higher PSNR and then fused optimal sub images to obtain on the fused image.

Conclusions
This work presents removing blur from images
using inverse filtering method .The results show
that higher PSNR is the better for selecting sub image to consisting fused image. The fusion process is given better results than deblurring techniques without fusion process.