Object-based crop pattern detection from IKONOS satellite images in agricultural areas

Nowadays, with the development of remote sensing technologies and image evaluation methods, remote sensing methods have become frequently preferred in studies to determine the crop pattern in agricultural areas and to monitor the change depending in time. In this study, it was aimed to detection the crop pattern of the high resolution IKONOS satellite image using object-based classification methods. In this study, it is aimed to detection the crop pattern in agricultural areas with high accuracy by using object-based classification technique from high spatial resolution IKONOS satellite images. The study area is located on the South-west of the Karacabey district of Bursa and covers an area of 18 km × 13.5 km. The ESP-2 (Estimation of Scale Parameter) tool was used to accelerate and automate the determination of the scale parameter in the segmentation step of the object-based classification. Various combinations have been tried for shape and compactness parameters in order to find the optimal segmentation parameters. In order to increase classification accuracy, six GLCM texture measurement methods have been identified, including homogeneity, contrast, dissimilarity, mean, variance, and entropy, which are frequently used in the literature. These methods were applied to the original bands of the IKONOS satellite image and an additional data set of 24 bands was obtained. Using a total of 29 bands together with the original bands and the additional data set of 24 bands, the image classification process was performed using the object-based nearest neighbor classification technique in the eCognition software. The obtained classification results were tested on parcel basis using 2212 ground truth data. In the accuracy assessment performed, it was seen that classification results and ground truth data were coherent with 87.48% (total accuracy).