The intensive use of geophysics as a tool for archaeological science demands new ways of fast and conclusive data interpretation. With the increasing size and complexity of geomagnetic data, manual delineation and classification of magnetic anomalies can be a large time consuming process. In this respect, our study introduces a new approach to automate this laborious procedure, which is being implemented as a ready-to-use tool for the eCognition® software. The presented methodology starts with a multiresolution segmentation algorithm operation performed on a single layer containing magnetic values. For reasons of objectivity, the ESP (Estimation of Scale Parameters) tool is used to automatically set the scale parameter. The segmentation delineates magnetic anomalies at three levels of scale. To automatically select and combine relevant objects, a proper classification scheme was designed, which is based on only two parameters: magnetic values and their standard deviation (SD). Objects of interest resulted from the segmentation are assigned to classes by using thresholds given by the two above mentioned parameters. The desired impact of standard deviation on the classification is adjustable, users having the possibility to choose between two values: 0.5 and 1 SD. The current experiments were undertaken on the geomagnetic images of a lost medieval village in southwestern Romania. The data has been acquired along parallel profiles covering six squares, each having the side of 100 m. Hence, to avoid potential errors caused by boundary lines, analyses were performed on individual squares. The tested scenes indicated accurate results, displaying positive- and negative-valued magnetic anomalies with levels of detail almost similar to manually delineated anomalies. Our methodology is fully automated and simple to apply. Being implemented as a customized process for the eCognition® software, the tool has a significant potential to interpret any image obtained through geophysical measurements and we consider it to be an aid for the large-scale surveys.