Search

Abstract: That the multi-source remote sensing information integrates knowledge-based geospatial constraints to develop efficient and practical Land cover classification algorithm has become one of the most important developing directions in the field of remote sensing ground object classification. Remote sensing classification is a strictly incompatible problem, but the spectra distribution of remote sensing data has compatible attributes especially in mountainous regions, and such contradiction is one of the main reasons leading to uncertainties in remote sensing classification. In this paper, the remote sensing spectra feature compatible information is transformed into the probability of the association degree firstly, and then the matter-element theory is introduced to establish models to achieve the integrated classification of multi-source data to fuse knowledge-based geographical constrained condition probability. Taking the grassland–wetland fragile ecosystem in Ruoergai plateau of China as a case study, this paper selected the multi-source data including images of Landsat TM and CBERS, ASTER-GDEM and MODIS-NDVI to construct a comprehensive classifier, in which the relationship between topography and land cover, and the prior knowledge on vegetation growth difference were taken as constraints to support the decision-making. The classification accuracy was evaluated by a field investigation and existing land cover map. The test result shows that, the overall accuracy (89.89%) and Kappa coefficient (0.8870) are better than those derived by the Maximum Likelihood method. It indicates that the proposed classification method is not subject to the dimensionality and form of data sources, and it can integrate the data source information to improve the classification accuracy, so that it is very useful to apply multi-source data and prior knowledge to land cover classification in mountainous regions. [Copyright &y& Elsevier]

Abstract: This paper reports the development of a decision tree algorithm to classify the surface soil freeze/thaw states. The algorithm uses SSM/I brightness temperatures recorded in the early morning. Three critical indices are used as classification criteria—the scattering index (SI), the 37 GHz vertical polarization brightness temperature (T 37V), and the 19 GHz polarization difference (PD19). The thresholds of these criteria were obtained from samples of frozen soil, thawed soil, desert, and snow. The algorithm is capable of distinguishing between frozen soil, thawed soil, desert, and precipitation. In-situ 4-cm deep soil temperatures on the Qinghai–Tibetan Plateau were used to validate the classification results, and the average classification accuracy was found to be 87%. Regarding the misclassified pixels, about 40% and 73% of them appeared when the surface soil temperature ranged from −0.5 °C to 0.5 °C and from −2.0 °C to 2.0 °C, respectively, which means that most misclassifications occurred near the soil freezing point. In addition, misclassifications mainly occurred from April to May and September to October, the transition periods between warm and cold seasons. A grid-to-grid Kappa analysis was also conducted to evaluate the consistency between the map of the actual number of frozen days obtained using the decision tree classification algorithm and the reference map of geocryological regionalization and classification in China. The overall classification accuracy was 91.7%, and the Kappa index was 80.5%. The boundary between the frozen and thawed soil was consistent with the southern limit of seasonally frozen ground from the reference map. The statistics show that the maximum area of frozen soil is about 6.82×106 km2 in late January, accounting for 69% of total Chinese land area. [Copyright &y& Elsevier]

Range and sill are two important parameters of a variogram. Their extraction usually Involves experimental fitting of variograms using models specified by the analyst and requires much use of trial and error. The objective of this paper is to desâgn an algorithm for extracting the range and sill of a variogram automatically without fitting a model Combined with the semivariance at the lag of one pixel (γ1), the extracted range and sift were applied to the textural classification of a panchromatic IKONOS image over Xichang, Sichuan Province, China. Results show that any of these three parameters can lead to the increase of the classification accuracy. When all three parameters were used with the raw image data, the average kappa statistic for five window sizes increased from 0.24 to 0.76, indicating promise of the range and sill in texture classification. [ABSTRACT FROM AUTHOR]