An Improved Image Restoration and Edge Detection Technique
Orvila Sarker, Rokeya Begum Jothi · Engineering International · 2016
Clustering and edge based segmentation are two basic image segmentation technique. This paper involves image clustering based restoration technique for finding out the set of consequential groups and restoring the original image from a noisy image. Previously, the feature of image cluster computing and restoration method is researched separately but now we combined the cluster and restoration method together. The k means clustering algorithm is applied on similar objects to create a cluster that separate noisy pixels and finally we use Gaussian filter to restore the noise corrupted image which enhanced the image quality. The simulation results show that the techniques are able to produce better output in terms of contrast and resolution. In case of edge based segmentation, canny edge detection algorithm is the optimal one because of its low error rate, good localization, only one response to a single edge etc. In this work, we have showed that applying double threshold in canny edge detection algorithm provides reasonably better output.