Multiple policy projects have changed land use and land cover (LULC) in China’s rural regions over the past years, resulting in two types of rural settlements: new-fashioned and old-fashioned. Precise extraction of and discrimination between these two settlement types are vital for sustainable land use development. It is difficult to identify these two types via remote sensing images due to their similarities in spectrum, texture, and geometry. This study attempts to discriminate different types of rural settlements by using a spatial contextual information extraction method based on Gaofen 2 (GF-2) images, which integrate hierarchical multi-scale segmentation and landscape analysis. A preliminary LULC map was derived by using only traditional spectral and geometrical features from a finer scale. Subsequently, a vertical connection was built between superobjects and subobjects, and landscape metrics were computed. The vertical connection was used for assigning landscape contextual information to subobjects. Finally, a classification phase was conducted, in which only multi-scale contextual information was adopted, to discriminate between new-fashioned and old-fashioned rural settlements. Compared with previous studies on multi-scale contextual information, this paper employs landscape metrics to quantify contextual characteristics, rather than traditional spectral, textural, and topological relationship information, from superobjects. Our findings indicate that this approach effectively identified and discriminated two types of rural settlements, with accuracies over 80% for both producers and users. A comparison with a conventional top-down hierarchical classification scheme showed that this novel approach improved accuracy, precision, and recall. Our results confirm that multi-scale contextual information with landscape metrics provides valuable spatial information for classification, and indicates the practicability, applicability, and effectiveness of this synthesized approach in distinguishing different types of rural settlements.

Detailed and precise information of land-use and land-cover (LULC) in rural area is essential for land-use planning, environment and energy management. The confusion in mapping residential and industrial areas brings problems in energy management, environmental management and sustainable land use development. However, they remain ambiguous in the former rural LULC mapping, and this insufficient supervision leads to inefficient land exploitation and a great waste of land resources. Hence, the extent and area of residential and industrial cover need to be revealed urgently. However, spectral and textural information is not sufficient for classification heterogeneity due to the similarity between different LULC types. Meanwhile, the contextual information about the relationship between a LULC feature and its surroundings still has potential in classification application. This paper attempts to discriminate settlement and industry area using landscape metrics. A feasible classification scheme integrating landscape metrics, chessboard segmentation and object-based image analysis (OBIA) is proposed. First LULC map is generated from GeoEye-1 image, which delineated distribution of different land-cover materials using traditional OBIA method with spectrum and texture information. Then, a chessboard segmentation of the whole LULC map is conducted to create landscape units in a uniform spatial area. Landscape characteristics in each square of chessboard are adopted in the classification algorithm subsequently. To analyze landscape unit scale effect, a variety of chessboard scales are tested, with overall accuracy ranging from 75% to 88%, and Kappa coefficient from 0.51 to 0.76. Optimal chessboard scale is obtained through accuracy assessment comparison. This classification scheme is then compared to two other approaches: a top-down hierarchical classification network using only spectral, textural and shape properties, and lacunarity based hierarchical classification. The distinction approach proposed is overwhelming by achieving the highest value in overall accuracy, Kappa coefficient and McNemar test. The results show that landscape properties from chessboard segment squares could provide valuable information in classification.