The sorting of laundry items, in particular dirty laundry items, has hitherto been carried out manually in most cases, which is costly in terms of personnel and time. High sorting capacities are therefore only possible with the corresponding personnel expenses. The invention provides that, at the start of sorting, specifically during their separation, and prior to their further transport to the individual sorting sites, the dirty laundry items are scanned by imaging techniques and, as a result of the appropriate image analysis, it is possible to determine the areas of the dirty laundry items to be preferably gripped and/or at least a number of sorting criteria, such as size and color. These measures result in a largely automated sorting procedure. These measures can be extended so as to achieve a fully automatic sorting procedure.