エッジを保った画像圧縮法 : GBTC法との比較(B.一般論文,開学20周年記念特集号)

概要

論文の詳細を見る
This paper proposes an image compression method suitable for digital image-based plotters. This plotter is an image processing system designed for aerial photogrammetry for map produciton. Since data volume is quite large, efficient data compression techniques are important. In aerial photogrammetry different pixel sizes are required in orientation of images and plotting object contours. In this aspect pyramidal image compression, including the GBTC (Generalized Block Truncation Coding) algorithm, looks most appropriate. But the GBTC algorthm does not preserve edges of images, and loses significant information when plotting object contours. The authors propose a method to store positions of edges instead of quantization levels in GBTC. The experiments illustrates the quality comparison of decompressed images with the GBTC and the authors' method. The result shows the latter is superior both in precision of idenfifying point pairs in stereo images and the compression rate.
福山大学の論文