The superpixel-based multiple kernels (MK) model uses the average value of all pixels within superpixel as the spatial feature, which results in inaccurate extraction of edge pixels. To solve this problem, a local binary patterns and superpixel-based multiple kernels (LBP-SMK) method is proposed for hyperspectral image (HSI) classification. Firstly, the original hyperspectral image is segmented into multiple superpixels by using the entropy rate superpixel (ERS) segmentation algorithm. On the HSI with superpixel index, the spectral kernel is secondly obtained by combining the spectral feature map with the radial basis kernel (RBF); By introducing local binary pattern (LBP) and weighted average filtering into RBF, the spatial kernels are obtained within and among superpixels. Finally, the combined kernel containing the above three kernels is inputted into the support vector machine (SVM) classifier to generate a classification map. The experimental procedures in this paper used LBP to extract the information in superpixels, which prevents the loss of edge features in superpixels. The experimental results show that the proposed method is superior to the state-of-the-art classifiers for HSI classification.

Comprehensive land-use planning (CLUP) at the county level in China must include land-use zoning. This is specifically stipulated by the China Land Management Law and aims to achieve strict control on the usages of land. The land-use zoning problem is treated as a multi-objective optimization problem (MOOP) in this article, which is different from the traditional treatment. A particle swarm optimization (PSO) based model is applied to the problem and is developed to maximize the attribute differences between land-use zones, the spatial compactness, the degree of spatial harmony and the ecological benefits of the land-use zones. This is subject to some constraints such as: the quantity limitations for varying land-use zones, regulations assigning land units to a certain land-use zone, and the stipulation of a minimum parcel area in a land-use zoning map. In addition, a crossover and mutation operator from a genetic algorithm is adopted to avoid the prematurity of PSO. The results obtained for Yicheng, a county in central China, using different objective weighting schemes, are compared and suggest that: (1) the fundamental demand for attribute difference between land-use zones leads to a mass of fragmentary land-use zones; (2) the spatial pattern of land-use zones is remarkably optimized when a weight is given to the sub-objectives of spatial compactness and the degree of spatial harmony, simultaneously, with a reduction of attribute difference between land-use zones; (3) when a weight is given to the sub-objective of ecological benefits of the land-use zones, the ecological benefits get a slight increase also at the expense of a reduction in attribute difference between land-use zones; (4) the pursuit of spatial harmony or spatial compactness may have a negative effect on each other; (5) an increase in the ecological benefits may improve the spatial compactness and spatial harmony of the land-use zones; (6) adjusting the weights assigned to each sub-objective can generate a corresponding optimal solution, with a different quantity structure and spatial pattern to satisfy the preference of the different decision makers; (7) the model proposed in this paper is capable of handling the land-use zoning problem, and the crossover and mutation operator can improve the performance of the model, but, nevertheless, leads to increased time consumption.

The quality of cultivated land determines the production capacity of cultivated land and the level of regional development, and also directly affects the food security and ecological safety of the country. This paper starts from the perspective of spatial pattern of cultivated land quality and uses spatial autocorrelation analysis to study the spatial aggregation characteristics and differences of cultivated land quality in Henan Province at the county level scale, and also uses bivariate spatial autocorrelation to analyze the influence of neighboring influences on the quality of cultivated land in the target area. The spatial autoregressive model was used to further analyze the driving factors affecting the quality of cultivated land, and the influence of cultivated land area index was coupled in the process of rating analysis, which was finally used as a basis to propose more precise measures for the protection of cultivated land zoning. The results show that: (1) The quality of cultivated land in Henan Province has a strong spatial correlation (global Moran’s I≈0.710) and shows an obvious aggregation pattern in spatial distribution; positive correlation types (high-high and low-low) are concentrated in north-central and western mountainous areas of Henan Province, respectively; negative correlation types are discrete. The negative correlation types are distributed in a discrete manner. (2) The bivariate spatial autocorrelation results show that Slope (Moran’s I≈-0.505), Irrigation guarantee rate (IGR, 0.354), Urbanization rate (-0.255), Total agricultural machinery power (TAMP, 0.331) and Pesticide use (0.214) are the main influencing factors. (3) According to the absolute values of the regression coefficients, it can be seen that the magnitude of the influence of different factors on the quality of cultivated land is: Slope (0.089) >IGR (0.025) > Urbanization rate (0.002) > TAMP (0.001) > Pesticide use (1.96e-006). (4) Based on the spatial pattern presented by the spatial autocorrelation results, we proposed corresponding protection zoning measures to provide more scientific reference decisions and technical support for the implementation of refined cultivated land management in Henan Province.

In this paper, we propose a new architecture of densely connected convolutional networks for pan-sharpening (DCCNP). Since the traditional convolution neural network (CNN) has difficulty handling the lack of a training sample set in the field of remote sensing image fusion, it easily leads to overfitting and the vanishing gradient problem. Therefore, we employed an effective two-dense-block architecture to solve these problems. Meanwhile, to reduce the network architecture complexity, the batch normalization (BN) layer was removed in the design architecture of DenseNet. A new architecture of DenseNet for pan-sharpening, called DCCNP, is proposed, which uses a bottleneck layer and compression factors to narrow the network and reduce the network parameters, effectively suppressing overfitting. The experimental results show that the proposed method can yield a higher performance compared with other state-of-the-art pan-sharpening methods. The proposed method not only improves the spatial resolution of multi-spectral images, but also maintains the spectral information well.

Rural residential area is the core component of rural land, and the process of its transfer exists to guide the change of rural land use. Its evolution is closely related to population migration, so exploring the spatial and temporal evolution of the coupling type of human-land relationship and its driving mechanism is the core scientific basis for the optimal allocation of rural land resources. From the perspective of human-land relationship, this paper introduces spatio-temporal big data on the basis of 3S technology and constructs an elasticity coefficient model. The composite index method is also used to further explore the coupling types of rural human-land relations in the Luoyang region. At the same time, spatial autocorrelation and spatial autoregressive models were used to reveal the driving mechanism of the spatio-temporal evolution of rural human-land relationship types in the Luoyang region. The results showed that: (1) Against the background of the continuous reduction of the rural population in Luoyang, the scale of rural residential area first increases and then decreases, and spatially they continue to gather around the central city (2) At the township scale, the conflict between people and land tends to be moderated. The elasticity coefficient shows a zonal distribution pattern in the southwest of Luoyang region and the townships around the central city. (3) There is a non-negligible spatial correlation between rural human-land relations in Luoyang region. The initial scale of rural residentials, arable land area, rural resident population and industrial output value are positively correlated with the coordination of people-land relationship, and the slope, elevation and distance from the central city have an inverse effect on the harmony of the relationship between people and land.