A Parallel Architecture for Improving the Performance of the Kriging Algorithm

Authors : Özgür TAMER, Ahmet Esat GENÇ

Pages : 463-471

Doi:10.29137/umagd.1165147

View : 54 | Download : 53

Publication Date : 2023-07-14

Article Type : Research Paper

Abstract :Estimating missing data values by using interpolation algorithms is a well-known technique. Kriging is an optimized interpolation method based on regression against evaluated values from the surrounding observation points, weighted according to spatially varying values according to the covariance between these observation points. It has been widely used for estimating the missing geological data of the areas based on the measurements in close proximity. In this work we use the Kriging to recover the missing pixels of digital images. Even though Kriging is considered as successful on estimating the missing pixels, the algorithm has a high operation load, causing delays especially for live streaming videos. In this paper we propose a parallel architecture to improve the performance and reduce the operation time of the Kriging Algorithm for estimating the missing pixels. The proposed method can be applied on Field Programmable Gate Arrays insert ignore into journalissuearticles values(FPGA); and considerable performance improvement have been achieved depending on the number of logic blocks available inside the FPGA.
Keywords : Kriging Algoritması, paralel mimariler, interpolasyon, görüntü tekrar inşası, FPGA