The extraction and classification of crops is the core issue of agricultural remote sensing. The precise classification of crop types is of great significance to the monitoring and evaluation of crops planting area, growth, and yield. Based on the Google Earth Engine and Google Colab cloud platform, this study takes the typical agricultural oasis area of Xiangride Town, Qinghai Province, as an example. It compares traditional machine learning (random forest, RF), object-oriented classification (object-oriented, OO), and deep neural networks (DNN), which proposes a random forest combined with deep neural network (RF+DNN) classification framework. In this study, the spatial characteristics of band information, vegetation index, and polarization of main crops in the study area were constructed using Sentinel-1 and Sentinel-2 data. The temporal characteristics of crops phenology and growth state were analyzed using the curve curvature method, and the data were screened in time and space. By comparing and analyzing the accuracy of the four classification methods, the advantages of RF+DNN model and its application value in crops classification were illustrated. The results showed that for the crops in the study area during the period of good growth and development, a better crop classification result could be obtained using RF+DNN classification method, whose model accuracy, training, and predict time spent were better than that of using DNN alone. The overall accuracy and Kappa coefficient of classification were 0.98 and 0.97, respectively. It is also higher than the classification accuracy of random forest (OA = 0.87, Kappa = 0.82), object oriented (OA = 0.78, Kappa = 0.70) and deep neural network (OA = 0.93, Kappa = 0.90). The scalable and simple classification method proposed in this paper gives full play to the advantages of cloud platform in data and operation, and the traditional machine learning combined with deep learning can effectively improve the classification accuracy. Timely and accurate extraction of crop types at different spatial and temporal scales is of great significance for crops pattern change, crops yield estimation, and crops safety warning.

Small reservoirs play an important role in mining, industries, and agriculture, but storage levels or stage changes are very dynamic. Accurate and up-to-date maps of surface water storage and distribution are invaluable for informing decisions relating to water security, flood monitoring, and water resources management. Satellite remote sensing is an effective way of monitoring the dynamics of surface waterbodies over large areas. The European Space Agency (ESA) has recently launched constellations of Sentinel-1 (S1) and Sentinel-2 (S2) satellites carrying C-band synthetic aperture radar (SAR) and a multispectral imaging radiometer, respectively. The constellations improve global coverage of remotely sensed imagery and enable the development of near real-time operational products. This unprecedented data availability leads to an urgent need for the application of fully automatic, feasible, and accurate retrieval methods for mapping and monitoring waterbodies. The mapping of waterbodies can take advantage of the synthesis of SAR and multispectral remote sensing data in order to increase classification accuracy. This study compares automatic thresholding to machine learning, when applied to delineate waterbodies with diverse spectral and spatial characteristics. Automatic thresholding was applied to near-concurrent normalized difference water index (NDWI) (generated from S2 optical imagery) and VH backscatter features (generated from S1 SAR data). Machine learning was applied to a comprehensive set of features derived from S1 and S2 data. During our field surveys, we observed that the waterbodies visited had different sizes and varying levels of turbidity, sedimentation, and eutrophication. Five machine learning algorithms (MLAs), namely decision tree (DT), k-nearest neighbour (k-NN), random forest (RF), and two implementations of the support vector machine (SVM) were considered. Several experiments were carried out to better understand the complexities involved in mapping spectrally and spatially complex waterbodies. It was found that the combination of multispectral indices with SAR data is highly beneficial for classifying complex waterbodies and that the proposed thresholding approach classified waterbodies with an overall classification accuracy of 89.3%. However, the varying concentrations of suspended sediments (turbidity), dissolved particles, and aquatic plants negatively affected the classification accuracies of the proposed method, whereas the MLAs (SVM in particular) were less sensitive to such variations. The main disadvantage of using MLAs for operational waterbody mapping is the requirement for suitable training samples, representing both water and non-water land covers. The dynamic nature of reservoirs (many reservoirs are depleted at least once a year) makes the re-use of training data unfeasible. The study found that aggregating (combining) the thresholding results of two SAR and multispectral features, namely the S1 VH polarisation and the S2 NDWI, respectively, provided better overall accuracies than when thresholding was applied to any of the individual features considered. The accuracies of this dual thresholding technique were comparable to those of machine learning and may thus offer a viable solution for automatic mapping of waterbodies.

冰山识别对于海洋环境监测和船只安全运行等具有重要的意义，是北极航道开通和北极开发过程中的重要内容。采用合成孔径雷达（SAR）影像进行冰山识别具有独特的优势，多种机器学习算法均可用于SAR影像的冰山识别中。为了最大限度地发挥机器学习算法的性能，有必要对不同机器学习算法及其搭配使用的特征与特征标准化方法进行评估，从而进行最优冰山识别方法的选择。因此，本文基于Sentinel-1A SAR影像，采用多种机器学习方法、多种特征组合及多种特征标准化方法进行冰山识别，并比较各流程方法的识别性能差异。采用的机器学习算法包括贝叶斯分类器（Bayes）、反向神经网络（BPNN）、线性判别分析（LDA）、随机森林（RF）以及支持向量机（SVM）；特征标准化方法包括Min-max标准化、Z-score标准化及log函数标准化；数据集是含有12个SAR影像特征的969个冰山与非冰山样本，样本主要位于格陵兰岛东海岸。分类效果采用接收者操作特性（ROC）曲线下的面积（AUC）进行衡量。结果显示，最佳搭配下的RF的AUC值最高，达到了0.945，比最差的Bayes高出0.09。从识别率上来看，RF在冰山查全率为80%的情况下非冰山查全率达到92.6%，效果最好，比第2位的BPNN高出1.4%，比最差的Bayes高出2.6%；BPNN在冰山查全率为90%的情况下非冰山查全率达到87.4%，比第2位的RF高出0.8%，比最差的Bayes高出2.7%。上述结果表明，对冰山识别而言，选择最优的机器学习算法和最佳的特征与特征标准化方法都是十分重要的。

珊瑚礁生态系统是全球初级生产力最高的生态系统之一, 在维持海洋生物多样性、防浪固滩、资源供给等方面发挥着巨大作用。珊瑚礁遥感地貌分类体系是珊瑚礁保护、管理及可持续发展的必要基础, 但目前还未有结合珊瑚覆盖度进行分类的体系。本文基于WorldView-2高分辨率遥感影像, 以中国南海西沙群岛七连屿北部赵述岛和西沙洲所在礁盘作为研究区, 结合珊瑚覆盖度、区域地貌成分和水动力条件等指标, 建立了既适用于遥感监测又与珊瑚生存状况相关联的珊瑚礁地貌单元分类体系。同时, 利用面向对象的支持向量机(Support Vector Machine, SVM)和随机森林(Random Forest, RF)分类方法进行珊瑚礁地貌单元的信息提取, 并对分类结果进行精度评价。结果表明, SVM和RF两种分类方法均能较好地提取出珊瑚礁地貌单元, 分类精度分别为87.59%和79.81%。针对分类过程中出现的错分、漏分问题, 结合珊瑚礁成因和分布规律对分类结果进行修正, 修正后分类提取的精度达到91.3%, Kappa系数为0.9041, 表明本文构建的珊瑚礁地貌单元分类体系在一定程度上能满足当前珊瑚岛礁信息提取的需要。

本文以1983-2011 年的多源遥感数据为基础, 利用GIS 空间分析和地图代数功能, 提取杭州湾海岸线, 分析了其变迁的位置、长度, 以及增加和减少的陆地面积。结果表明:受到自然和人为因素的影响, 杭州湾南北两岸的海岸线变迁规律不同。杭州湾北岸1983-1993 年共有60.2km的海岸线向陆迁移, 最大迁移距离0.6km, 减少的陆地面积共为23.5km²。1993-2011 年由于围垦和工业填海北岸向海迁移。其中,1993-2002 年最大迁移距离3.6km, 新增陆地面积42.5km², 2002-2011 年最大迁移距离2.9km, 新增陆地面积61.0km²。由于淤积和围垦, 杭州湾南岸海岸线不断向海迁移, 1983-1993 年、1993-2002 年和2002-2011 年向海迁移最大距离分别是1.8km、2.7km和5.1km, 新增陆地面积分别为34.3km²、230.2km²及331.7km²。海岸线向海迁移的速度越来越快, 规模越来越大。研究成果对于地图制图、滨海湿地生态资源管理, 以及海岸线保护具有十分重要的意义。

Coastal wetlands provide essential ecosystem services, while usually experiencing land transformation or degradation mainly due to intense anthropogenic activities and climate changes. Understanding the changes in wetlands ecosystem services is essential to decision makers for generating sound coastal planning. Hangzhou Bay is rich in wetland resources, and the urbanization of Hangzhou Bay in the past three decades has caused fundamental changes in the wetlands in the region. Based on the remote sensing images of the Hangzhou Bay area from 1990 to 2020, this paper analyzes the land use situation of the Hangzhou Bay area in seven periods. This paper calculates the area transfer matrix of various types of wetlands. It uses the InVEST model to evaluate the changes in the function of wetland ecosystem services in the Hangzhou Bay area. Hangzhou Bay wetlands show a trend of transferring natural wetlands to artificial and non-wetlands from 1990 to 2020. Carbon stocks fell by 14.24%. The annual water production decreased by 33.93% and then returned to the original level. The area of habitat degradation increased by 79.94%. The main influencing factors are paddy field degradation, increase in non-wetland area, and decrease in sea area. This paper proposes that the development and construction of farmland in the “red line” area and established wetland reserves are prohibited, and to strengthen the training of wetland management personnel, establish a sound decision-making consultation mechanism, and increase the scientific research expenditure on wetlands in the region.

围填海影响下海湾形态变化能够深刻反映人类活动对海湾自然环境的影响程度,分析海湾形态变化对合理高效地利用与保护海湾资源具有重要意义。研究以东海区12个主要海湾（包括陆域与水域）为研究区,基于20世纪90年代以来6个时期的Landsat TM/OLI遥感影像数据,通过海湾岸线与湾面形态分析东海区主要海湾的变化特征,探讨围填海强度与海湾形态变化之间的相关性。主要结论为：① 1990—2015年,东海区主要海湾岸线总长度共波动增长66.65 km,2005—2010年间海湾开发最活跃,阶段内岸线增长量达38 km。岸线长度三沙湾最大（439 km）,泉州湾最小（105 km）;兴化湾增长最多（54.53 km）,罗源湾缩短最多（25.75 km）。自然岸线与人工岸线长度此消彼长,岸线人工化程度不断加强,东海北部海湾岸线总长度大于南部海湾。② 1990—2015年,东海区海湾岸线共向海推进26.93 km,合1.08 km/a,在1995—2000年及2005—2010年间推进最多,分别达7.10 km和 6.00 km,在1990—1995年间推进量最小,为2.97 km。杭州湾（4.93 km）和兴化湾（4.15 km）岸线向海推进距离最长,厦门湾推进（0.55 km）最短;东海南部海湾岸线迁移量平缓,北部海湾则更为剧烈,是东海区岸线迁移变化的主体。③ 1990—2015年间东海区主要海湾水域总面积由初期的13.85 km2减少至12.29 km2,累积减少11.23 %,海湾形状不断向复杂化演变。其中杭州湾海湾水域面积减少量最多,达到0.726 km2,占研究区的46.69 %。空间上,北部海湾水域面积减少量更大,而南部海湾水域面积减小速率更快。④ 1990年以来,东海区主要海湾人工化指数平均值和岸线开发强度指数均有所上涨,21世纪以来的开发利用度显著提高。南部（闽）海湾的开发利用程度较北部（浙沪）更为深入,北部海湾开发强度的年际波动差异更大。海湾开发强度与海湾岸线长度、人工岸线长度、海湾形态指数呈正相关关系,与自然岸线长度、海湾水域面积呈负相关关系。当海湾开发强度增加时,同时段内海湾围填海活动的强度也显著增加。

The spatial pattern changes of bays under the influence of reclamation can profoundly reflect how human activities affect the natural environments, which is important to effectively protect and utilize bay resources. Based on 6 Landsat TM/OLI remote sensing images during 1990-2015, this study analyzed the variations of major bays from the coastline and bay surface morphology and explored the correlation between the reclamation intensity and spatial pattern changes for the 12 major bays in the East China Sea (ECS). The main conclusions include that: (1) the length of the main bay coastline in the East China Sea, from 1990 to 2015, increased by 66.65 km. The extensive coastline growth was found during 2005-2010 and the growth reached 38 km. Sansha Bay has the longest coastline (439 km) and the shortest (105 km) was found in Luoyuan Bay; Xinghua Bay experienced the largest coastline growth (54.53 km) in the past decades, and the least was in Luoyuan Bay (25.75 km). In general, the artificial coastline continued to increase and the degree of artificialization had been continuously strengthened. (2) The coastline of the bay continuously moved to the sea, with a distance of 26.93 km (1.08 km/a). The most significant seaward expansions were found in 1995-2000 and 2005-2010, reaching 7.10 km and 6.00 km, respectively. Hangzhou (4.93 km) and Xinghua bays (4.15 km) experienced the largest seaward expansion of coastline, while Xiamen Bay had the shortest (0.55 km). (3) The total area of the major bay waters decreased from 13.85 km2 in 1990 to 12.29 km2 in 2015 in the East China Sea, down by 11.23%. Additionally, the morphological indices of the bays showed a continuous rise trend, which indicates that spatial patterns were transformed to be more complicated. The largest reduction with water area was observed in the Hangzhou Bay (0.726 km2), accounting for 46.69% of the research area. (4) The indexes of artificiality and development intensity showed a continuous rise trend. The utilization degree in the southern part of the study area is higher than that of the northern part, and the interannual fluctuation of the development intensity in the north is much varied. In addition, the bay development is positively correlated with the length of the coastline, the length of the artificial coastline and the shape index of the bay, and negatively correlated with the length of the natural coastline and the area of the waters. As the development intensity increased, the intensity of reclamation activities increased significantly.

Dynamics of surface water is of great significance to understand the impacts of global changes and human activities on water resources. Remote sensing provides many advantages in monitoring surface water; however, in large scale, the efficiency of traditional remote sensing methods is extremely low because these methods consume a high amount of manpower, storage, and computing resources. In this paper, we propose a new method for quickly determining what the annual maximal and minimal surface water extent is. The maximal and minimal water extent in the year of 1990, 2000, 2010 and 2017 in the Middle Yangtze River Basin in China were calculated on the Google Earth Engine platform. This approach takes full advantage of the data and computing advantages of the Google Earth Engine’s cloud platform, processed 2343 scenes of Landsat images. Firstly, based on the estimated value of cloud cover for each pixel, the high cloud covered pixels were removed to eliminate the cloud interference and improve the calculation efficiency. Secondly, the annual greenest and wettest images were mosaiced based on vegetation index and surface water index, then the minimum and maximum surface water extents were obtained by the Random Forest Classification. Results showed that (1) the yearly minimal surface water extents were 14,751.23 km2, 14,403.48 km2, 13,601.48 km2, and 15,697.42 km2, in the year of 1990, 2000, 2010, and 2017, respectively. (2) The yearly maximal surface water extents were 18,174.76 km2, 20,671.83 km2, 19,097.73 km2, and 18,235.95 km2, in the year of 1990, 2000, 2010, and 2017, respectively. (3) The accuracies of surface water classification ranged from 86% to 93%. Additionally, the causes of these changes were analyzed. The accuracy evaluation and comparison with other research results show that this method is reliable, novel, and fast in terms of calculating the maximal and minimal surface water extent. In addition, the proposed method can easily be implemented in other regions worldwide.

Classification procedures are some of the most widely used statistical methods in ecology. Random forests (RF) is a new and powerful statistical classifier that is well established in other disciplines but is relatively unknown in ecology. Advantages of RF compared to other statistical classifiers include (1) very high classification accuracy; (2) a novel method of determining variable importance; (3) ability to model complex interactions among predictor variables; (4) flexibility to perform several types of statistical data analysis, including regression, classification, survival analysis, and unsupervised learning; and (5) an algorithm for imputing missing values. We compared the accuracies of RF and four other commonly used statistical classifiers using data on invasive plant species presence in Lava Beds National Monument, California, USA, rare lichen species presence in the Pacific Northwest, USA, and nest sites for cavity nesting birds in the Uinta Mountains, Utah, USA. We observed high classification accuracy in all applications as measured by cross-validation and, in the case of the lichen data, by independent test data, when comparing RF to other common classification methods. We also observed that the variables that RF identified as most important for classifying invasive plant species coincided with expectations based on the literature.