Few rare, circular, concentric enclosure ditches called rondels were discovered in Slovakia, a country in Europe; within the ditches, material traces of Neolithic European culture can be excavated. For exca- vations to happen, archaeologists must first locate these rare structures. Most rondels were spotted on agricultural fields or searched for manually during aerial surveys in a time-consuming manner. With the release of a high-resolution multispectral aerial orthophotomosaic data set of Slovakia, detailed sites containing rondels may be discovered using machine learning techniques.

Machine learning techniques using convolutional neural network (CNN) models can be applied to the field of archaeology to search for excavation sites automatically. An obstacle remains: CNN models require a lot of training image data to be efficient at their task, which is to classify whether areas contain rondels or otherwise. There are only 20 visible rondels on the orthophotomosaic that can be used as training images for model input, creating an imbalanced data set of a class with a minority class of aerial images of rondels and a large majority class of aerial images without rondels. Sketches and recorded characteristics of rondels from current images and from archaeological publications were used to automatically and randomly replicate rondel appearances from above, resulting in a created balanced data set with sufficient rondel examples for CNN training.

Multiple ResNet-34 models and a ConvNeXt model with differing hyperparameters were trained. The most promising model, a modified ResNet-34, was selected based on validation loss from the cross- entropy loss function and on the number of correctly and incorrectly classified labeled images from a test set. The selected model is used to classify data from the orthophotomosaic for rondels using a sliding window technique. Over 9510 square kilometers of agricultural land cover in western and eastern Slovakia was selected from the CORINE land cover map for classification. 7 suspected rondel sites were found, and 2 were determined to likely be rondels, based on their circular ditch-like appearance in 4 sets of multispectral images and in LiDAR elevation data. Results indicate that exact rondel layouts can be delineated with high-resolution orthophotomasics, however identifying circular elevation patterns of ditches proves to be challenging without using additional LiDAR data.

Traditionally, archaeological investigations, especially archaeological remains detection, mostly depend on human observation. In order to find the objects in large areas, a lot of fieldwork has to be done and it takes a long time for archaeologists to travel around. Nowadays, the development of LIDAR provides accurate 3D geometric information, which can be used for computer-based detailed terrain study. The application of deployment of computer vision methods also provides a new idea for the automatic object detection approach.

In this study, the neural network architecture "ResNet18" was applied to airborne LiDAR data from the Western regions of Slovakia for the automated detection of undiscovered Neolithic Circular Enclosures (also called rondel in the thesis). NCEs are mysterious stone hedge-like rings scattered through Central/Eastern Europe. The LiDAR data was processed into digital rater data and realized data enhancement by the visualization technique -- Simple Local Relief Model (SLRM). Since the positive samples were limited, expanding the training dataset was crucial and was realized by data augmentation methods based on the positive samples of rondels. The augmented roundels were created by cropping the real roundels and pasting them on the new empty areas after slight modification. After that, the positive image samples and the same number of negative image samples constructed the whole data set and it was divided into two parts -- training data and test data. After the training process of ResNet18, the performances of deep learning models with different combinations of parameters were evaluated, and the selected model was applied to a large area (44276 × 29984 m2), the spatial distribution of the probabilities could be observed and 32 possible new rondel areas were chosen for further validation.