The evaluation of community livability quantifies the demands of human settlement at the micro scale, supporting urban governance decision-making at the macro scale. Big data generated by the urban management of government agencies can provide an accurate, real-time, and rich data set for livability evaluation. However, these data are intertwined by overlapping geographical management boundaries of different government agencies. It causes the difficulty of data integration and utilization when evaluating community livability. To address this problem, this paper proposes a scheme of partitioning basic geographical space into grids by optimally integrating various geographical management boundaries relevant to enterprise-level big data. Furthermore, the system of indexes on community livability is created, and the evaluation model of community livability is constructed. Taking Wuhan as an example, the effectiveness of the model is verified. After the evaluation, the experimental results show that the livability evaluation with reference to our basic geographic grids can effectively make use of governmental big data to spatially identify the multi-dimensional characteristics of a community, including management, environment, facility services, safety, and health. Our technical solution to evaluate community livability using gridded basic urban geographical data is of large potential in producing thematic data of community, constructing a 15-min community living circle of Wuhan, and enhancing the ability of the community to resist risks.

The identification of overpass structures in road networks has great significance for multi-scale modeling of roads, congestion analysis, and vehicle navigation. The traditional vector-based methods identify overpasses by the methodologies coming from computational geometry and graph theory, and they overly rely on the artificially designed features and have poor adaptability to complex scenes. This paper presents a novel method of identifying overpasses based on a target detection model (Faster-RCNN). This method utilizes raster representation of vector data and convolutional neural networks (CNNs) to learn task adaptive features from raster data, then identifies the location of an overpass by a Region Proposal network (RPN). The contribution of this paper is: (1) An overpass labelling geodatabase (OLGDB) for the OpenStreetMap (OSM) road network data of six typical cities in China is established; (2) Three different CNNs (ZF-net, VGG-16, Inception-ResNet V2) are integrated into Faster-RCNN and evaluated by accuracy performance; (3) The optimal combination of learning rate and batchsize is determined by fine-tuning; and (4) Five geometric metrics (perimeter, area, squareness, circularity, and W/L) are synthetized into image bands to enhance the training data, and their contribution to the overpass identification task is determined. The experimental results have shown that the proposed method has good accuracy performance (around 90%), and could be improved with the expansion of OLGDB and switching to more sophisticated target detection models. The deep learning target detection model has great application potential in large-scale road network pattern recognition, it can task-adaptively learn road structure features and easily extend to other road network patterns.

Agricultural drought is a condition of insufficient soil moisture caused by a deficit in precipitation over some time period. Soil moisture drops to a certain extent, adverse to the crop yield, and then reduces the production of crops. Soybean is one of the most important sources of oil and protein in the world. It is vulnerable to recurrent drought condition in the U.S. state of Iowa. This study was conducted to identify agricultural drought indicator that strongly correlated with soybean crop yield. Detail crop data (e.g., soybean crop yield, etc.) were collected from the USDA's National Agricultural Statistics Service (NASS). A region of interests is defined based on the MODIS 16-days 250m resolution vegetation index synthetic products (MOD13Q1) anddaily land surface Temperature/Emissivity 1km resolution products (MOD11A1) from 2000 to 2013 in Iowa, which were used to compare three kinds of remote sensing derived agricultural drought monitoring indicator of crop water demand status: (i) Crop morphological indices (e.g., NDVI/VCI); (ii) Crop physiological indices (canopy temperature, e.g., TCI); and (iii) Crop comprehensive indices (e.g., VSWI). Drought cumulative effects were considered according to the specific soybean crop growth stages including from planted to emerged, vegetative period (from emerged to blooming), reproductive period (from blooming to setting pods), and growing season (from emerged to dropping leaves). The impacts of drought duration on the soybean crop yield by both of indices were analyzed. These results imply that physiological indices and comprehensive index were more correlated to assess the effect of drought on soybean yield. Drought indices accumulated in reproductive period (from blooming to setting pods) are highly superior to other accumulated for impacting on soybean yield, while over the growth season (from emerged to dropping leaves) is also highly correlated with the total yield assessment. The results can aid on evaluating the effects of drought on soybean yield in different growth stage. This could be very useful to providing auxiliary decision-making information for drought relief, agricultural manager and grain merchant to plan solutions and prepare for potential drought in advance.

Imbalanced learning is a methodological challenge in remote sensing communities, especially in complex areas where the spectral similarity exists between land covers. Obtaining high-confidence classification results for imbalanced class issues is highly important in practice. In this paper, extreme gradient boosting (XGB), a novel tree-based ensemble system, is employed to classify the land cover types in Very-high resolution (VHR) images with imbalanced training data. We introduce an extended margin criterion and disagreement performance to evaluate the efficiency of XGB in imbalanced learning situations and examine the effect of minority class spectral separability on model performance. The results suggest that the uncertainty of XGB associated with correct classification is stable. The average probability-based margin of correct classification provided by XGB is 0.82, which is about 46.30% higher than that by random forest (RF) method (0.56). Moreover, the performance uncertainty of XGB is insensitive to spectral separability after the sample imbalance reached a certain level (minority:majority > 10:100). The impact of sample imbalance on the minority class is also related to its spectral separability, and XGB performs better than RF in terms of user accuracy for the minority class with imperfect separability. The disagreement components of XGB are better and more stable than RF with imbalanced samples, especially for complex areas with more types. In addition, appropriate sample imbalance helps to improve the trade-off between the recognition accuracy of XGB and the sample cost. According to our analysis, this margin-based uncertainty assessment and disagreement performance can help users identify the confidence level and error component in similar classification performance (overall, producer, and user accuracies).

Understanding crop rotation on satellite remote sensing derived vegetation indices is very necessary, because it helps us develop more scientific methods or indices for revealing the mechanism of agricultural drought in the species-level. In this paper, the impacts of crop rotation on vegetation condition index (VCI) was explored. First, we tried to justify that whether crop rotation is a typical agricultural practice in the study area, and counted the proportion of crop planting changes over any pixel in multi-year; and second, a neighbor-average based VCI index was developed for species-level cases, and the comparison with traditional VCI index had been conducted. The experimental study was conducted in state of Iowa, the primary cornproducing state in the Corn Belt of the United States. Moderate resolution imaging spectroradiometer (MODIS) normalized difference vegetation index (NDVI) time series and NASS's cropland data layer (CDL) among years 2002-2013 were utilizedfor data analysis. The results shown that crop rotation limited impacts the VCI index on corn drought monitoring across the study area. Even so, the research inspires a more accurate and valuable mean in the future for examining the mechanisms and processes of species-level drought monitoring.