Artificial intelligence supported inventory tracking software for archaeological sites: ARKHESTOR

The objective of this thesis was to develop a software system for the digital recording, classification, analysis, and reporting of archaeological findings. The system's objective is to facilitate the optimization of processes associated with the recording, storage, classification, reporting, and analysis of findings from archaeological excavation sites for academic research. In the course of developing the software, Vue.js was utilized for front-end development, Laravel for back-end development, and MySQL as the database. Furthermore, for the identification of analogous findings, machine learning and image processing techniques were employed using Python and libraries such as OpenCV and TensorFlow, with the Cosine Similarity function and the EfficientNetB0 model being selected. It was observed that this model provided high accuracy in the process of feature extraction and database recording of the findings. Analyses conducted using the Cosine Similarity method achieved a 92% accuracy rate. Furthermore, the developed software facilitates the recording and tracking of findings through the use of QR code and barcode technologies. In conclusion, the developed software saves time for archaeologists and researchers in various areas, from the recording of findings in excavation sites to classification and reporting, thereby accelerating scientific analyses.