Kalamas Delta belongs to Natura 2000 network. Most part of delta consists of agricultural land. Remote sensing and GIS are important tools to map and identify agricultural crops among others. The aim of our study is to map permanent agricultural crops in the plateau of Kalamas Delta using IKONOS satellite data by considering the spatial arrangement of the radiometric values of the pixels. The spatial pattern was estimated using the Moran's I spatial autocorrelation index and semivariance, which is the one-half the average squared difference between data values separated by the lag distance. The satellite data used in the analysis were geometrically corrected and enhanced by applying various multivariate data analysis methods. Additionally we studied the effect of the size and shape of the sampling area used to estimate the autocorrelation statistic as well as the orientation of the parcels when trying to detect and characterize the spatial patterns of the crops. The classification of satellite data was implemented using object-based classification techniques. The segmentation algorithm applied (segments) was based originally on the spectral and spatial characteristics of the IKONOS image. Afterwards, the classification of the segments was based, not only on their spectral and spatial characteristics, but also on the spatial arrangement of the radiometric values of the pixels of the image expressed by the autocorrelation and semivariance metrics. The various agricultural parcels, except the unique spectral characteristics, are also characterized by their spatial characteristics determined from the spatial patterns of the crops. This information when considered in the classification process together with their spectral and other spatial characteristics increased the final accuracy of the classification results from 60 to 73%. Concluding, it seems that permanent agricultural crops are distinct areas in high spatial resolution satellite images with particular spatial characteristics that are determined by the type of crop in relation to the spatial resolution of the satellite image.