The increased feature space available in object-based classification environments (e.g., extended spectral feature sets per object, shape properties, or textural features) has a high potential of improving classifications. However, the availability of a large number of derived features per segmented object can also lead to a time-consuming and subjective process of optimizing the feature subset. The objectives of this study are to evaluate the effect of the advanced feature selection methods of popular supervised classifiers (Support Vector Machines (SVM) and Random Forest (RF)) for the example of object-based mapping of an agricultural area using Unmanned Aerial Vehicle (UAV) imagery, in order to optimize their usage for object-based agriculture pattern recognition tasks. In this study, several advanced feature selection methods were divided into both types of classifiers (SVM and RF) to conduct further evaluations using five feature-importance-evaluation methods and three feature-subset-evaluation methods. A visualization method was used to measure the change pattern of mean classification accuracy with the increase of features used, and a two-tailed t-test was used to determine the difference between two population means for both repeated ten classification accuracies. This study mainly contribute to the uncertainty analysis of feature selection for object-based classification instead of the per-pixel method. The results highlight that the RF classifier is relatively insensitive to the number of input features, even for a small training set size, whereby a negative impact of feature set size on the classification accuracy of the SVM classifier was observed. Overall, the SVM Recursive Feature Elimination (SVM-RFE) seems to be an appropriate method for both groups of classifiers, while the Correlation-based Feature Selection (CFS) is the best feature-subset-evaluation method. Most importantly, this study verified that feature selection for both classifiers is crucial for the evolving field of Object-based Image Analysis (OBIA): It is highly advisable for feature selection to be performed before object-based classification, even though an adverse impact could sometimes be observed from the wrapper methods.

Object-based change detection (OBCD) has recently been receiving increasing attention as a result of rapid improvements in the resolution of remote sensing data. However, some OBCD issues relating to the segmentation of high-resolution images remain to be explored. For example, segmentation units derived using different segmentation strategies, segmentation scales, feature space, and change detection methods have rarely been assessed. In this study, we have tested four common unsupervised change detection methods using different segmentation strategies and a series of segmentation scale parameters on two WorldView-2 images of urban areas. We have also evaluated the effect of adding extra textural and Normalized Difference Vegetation Index (NDVI) information instead of using only spectral information. Our results indicated that change detection methods performed better at a medium scale than at a fine scale where close to the pixel size. Multivariate Alteration Detection (MAD) always outperformed the other methods tested, at the same confidence level. The overall accuracy appeared to benefit from using a two-date segmentation strategy rather than single-date segmentation. Adding textural and NDVI information appeared to reduce detection accuracy, but the magnitude of this reduction was not consistent across the different unsupervised methods and segmentation strategies. We conclude that a two-date segmentation strategy is useful for change detection in high-resolution imagery, but that the optimization of thresholds is critical for unsupervised change detection methods. Advanced methods need be explored that can take advantage of additional textural or other parameters.

Monitoring water quality is indispensable for the identification of threats to water environment and later management of water resources. Accurate monitoring and assessment of water quality have been long-term challenges. In this study, multivariate statistical techniques (MST) and water quality identification index (WQII) were applied to analyze spatiotemporal variation in water quality and determine the major pollution sources in the Qinhuai River, East China. A rotated principal component analysis (PCA) identified three potential pollution sources during the wet season (mixed pollution, physicochemical, and nonpoint sources of nutrients) and the dry season (nutrient, primary environmental, and organic sources) and they explained 81.14% of the total variances in the wet season and 78.42% of total variances in the dry season. The result of redundancy analysis (RDA) showed that population density, urbanization, and wastewater discharge are the main sources of organic pollution, while agricultural fertilizer consumption and industrial wastewater discharge are the main sources of nutrients such as nitrogen and phosphorus. The water quality of the Qinhuai River basin was determined to be mainly Class III (slightly polluted) and Class IV (moderately polluted) based on WQII. Temporally, the change trend of WQII showed that water quality gradually deteriorated between 1990 and 2005, improved between 2006 and 2010, and then deteriorated again. Spatially, the WQII distribution map showed that areas with more developed urbanization were relatively more polluted. Our results show that MST and WQII are useful tools to help the public and decision makers to evaluate the water quality of aquatic environment.

Excessive nutrient input to surface water, including nitrate, exacerbates water eutrophication. Clarifying the proportions of different nitrate sources in the aquatic environment is critical for improving the polluted water. However, nitrate sources in river basins are very complex and not clearly understood. In this study, nitrogen concentrations and nitrate isotopic compositions were determined to estimate the spatiotemporal variation in nitrate sources in the Yuntaishan River basin, Nanjing, East China, from March 2019 to January 2020. The results showed that the concentrations of total nitrogen (TN), ammonium (NH₄⁺-N), and nitrate (NO₃⁻-N) changed in the ranges of 0.53–18.0 mg/L, 0.01–15.4 mg/L, and 0.06–9.3 mg/L, respectively, wherein NO₃⁻-N was the main nitrogen form. Higher nitrogen concentrations appeared in winter and in the downstream parts of the river. In the entire river basin, the NO₃⁻-N mainly originated from sewage (67%) and soil (26%), with clear spatial variations. NO₃⁻-N in the Yunba sub-watershed was mainly derived from sewage (78%), which was higher than that in other tributaries, i.e., Shengli River (44%) and Yangshan River (49%). This was due to the fact that that Shengli and Yangshan sub-watersheds were covered by urban areas and were equipped with a complete sewage treatment system. In addition, the contributions of sewage to NO₃⁻-N rose from 60% upstream to 86% downstream, suggesting the increasing influence of the point source of sewage. The results showed that 53% of NO₃⁻-N in the basin outlet originated from the point source of sewage near the M4 site. Sewage contributed 75% of NO₃⁻-N in the rainy season and 67% of NO₃⁻-N in the dry season, suggesting the weakly temporal variation. Our results highlight the spatiotemporal variations in sources of NO₃⁻-N. These results will aid in the development of measures needed to control nitrogen pollution in river basins.