Using convolutional filter for restoration of distorted image

Abstract

The main objective of this paper is the enhancement of the quality and the restoration of the acquired and distorted images. There are a lot of application areas and many methods of image restoration. In this work we apply Convolutional Filter for such case. This technique includes several stages. First stage was pre-processing phase before coding in terms for method selection and picture size adjusting and writing pseudo code for copying it to the script. After pre-processing, the next stage is compiling the code and observing results obtained from different methods. Two different solutions for image enhancement have been used in this work. First method is Homomorphic Filtering which has been applied to the distorted image. Obtained result is very satisfactory in terms of the quality of the outcome picture. The second method is the Blind Deconvolution method. The Blind Deconvolution algorithm is also very good in terms of the error range and the quality of the resultant image. On the other hand, it has been observed from the results that, the Homomorphic Filtering can be good choice for object detection, while the Blind Deconvolution algorithm should be applied where the adjusting of the brightness of the input image is required.