application of pixel-based and object-based approaches for lulc mapping in jiroft region, s.e. iran

aims: due to population growth and the increase of demand for industrial and agricultural products, many tropical regions of iran have experienced landscape changes. satellite imagery and remote sensing (rs) are widely used to map these changes. the present study detects the land use/land cover (lulc) using some pixel-based and object-based classification approaches. method: this research was conducted in the jiroft area, kerman province, using landsat-8 satellite images and some pixel-based and object-based image analyzing methods known as the pbia and obia. to this end, the methodology was carried out in two different phases. at the first one, the lulc maps were extracted using some pbia techniques for september 2020. these techniques are including as mahalanobis distance (md), maximum likelihood (ml), neural network (nn), support vector machine (svm) as well as unsupervised technique of isodata. in the second phase, the lulc was produced using the obia approach, encompassing the multi-resolution method and decision tree (dt) technique for segmentation and classification, respectively. to this end, using a hybrid methodology, the high-resolution images of worldview-2 were firstly segmented. the segmented objects were later combined with the 7-month time series of ndvi, from october (2020) to april (2021), to find the necessary thresholds as the dt inputs. in this regard, the pre-processed landsat images were trained using ground control points (gcps), and their performances were finally evaluated. findings: results of the lulc maps demonstrated that the kappa coefficient and overall accuracy for isodata, md, ml, nn, and svm methods were calculated to be (51%, 66%), (81%, 86%), (88%, 91%), (90%, 93%) and (88% and 92%), respectively. the outcomes of the second phase for mapping the lulc showed that the obia achieved a high overall accuracy of about 96%. conclusion: results showed that among the pbia techniques, the nn and svm classifiers had slightly superior performance, but regarding both accuracy and execution time, the ml is known to be the best. although both pbia and obia approaches are highly applicable in mapping lulc, the obia significantly outperformed the pbia classifiers by higher overall accuracy and kappa statistics.