For archaeologists and restorers, reconstructing the original painting and object from hundreds of relics and fragments with two-dimensional pictures, such as wall paintings and mosaics found because of archaeological excavations, is time-consuming and difficult process due to the large number of pieces, the corrosion on the pieces, and the erasing of patterns and colours on the pieces. In this article, we present a new automated application to obtain the original object by reassembling archaeologically fragmented two-dimensional objects using image processing methods. The application basically consists of image acquisition, image edge segmentation, feature extraction based on colour and edge curve, pairwise matching and general composition. In the image acquisition section, to keep the external environmental effects, camera position and angle constant while transferring two-dimensional fragments to digital image, a circuit consisting of ESP32-CAM Board and FTDI Programmer was designed, and digital images taken were transferred by Wi-Fi into the application. In the image edge segmentation section, the edge curve was divided into the edge sub-part curves according to the deflection points on it. In the feature extraction section, the table of the best match values was created by using colour difference values calculated from the colour distributions on the sub-part curves and the maximum edge curve matching value calculated geometrically from the image sub-part curves according to the neighbouring parts. In the last section, we focused on reassembling the image of original object using the best match values table.
Anahtar Kelimeler: Image acquisition, Image edge segmentation, Feature extraction, Pairwise matching, Image general composition
|
