Images of the morphology of a polygonal network of ice-wedge casts are a valuable aid to paleoclimatological reconstructions. Usually such images are obtained by aerial photography showing polygonal crop marks refl ecting textural diff erences between wedge fi lling and host material. Our objective was to investigate an alternative method by measuring the soil apparent electrical conductivity (EC a ) with an electromagnetic induction (EMI) sensor. Based on an aerial photograph showing polygonal crop marks in an agricultural fi eld in Belgium, a test area of 0.63 ha was selected. A small part of the test area (6 by 6 m) was excavated revealing a clear pattern of ice-wedge casts. Th e wedges penetrated clay-rich Tertiary marine sediments, covered by a 0.6-m layer of eolian sandy sediments, and were associated with the permafrost during the last glacial period. We took 94 subsoil (0.6-0.8 m) samples distrib- uted over the test area and analyzed their texture. Th e results showed a clear diff erence between the Eocene host material (on average 21% clay) and the Quaternary wedge fi lling (on average 6% clay). Th e test area was surveyed with an EMI sensor (we used an EM38DD) which resulted in an accurate image of the polygonal network. We concluded that an EMI survey is an appropriate technique to image the morphology of a polygonal network of subsoil ice-wedge casts. A fi nal perspective comprises the strong heterogeneity of the subsoil, since nearly half of the subsoil consists of ice-wedge material. Th is might open perspectives for precision agriculture in such landscapes. Abbreviations: DOE, depth of exploration; EC a , apparent electrical conductivity; EMI, electromagnetic induction.