样本数据是开展湿地制图的研究基础之一，对于数据生产和精度验证具有重要作用。针对湿地生态系统类型多样，大区域的全湿地类型样本生产困难的问题，本研究提出了一种准确、高效的大区域密集湿地样本生产框架。该框架主要包括两部分：首先，基于已有的湿地数据产品，使用规则筛选的方法直接生产稳定的湿地样本点，能够得到河流、湖泊、水库、滨海木本沼泽（红树林）、滩涂的5种湿地类型样本点；其次，基于多源专题数据进行规则筛选，生产潜在湿地样本点，并利用Google Earth Engine大数据云平台、Google Earth平台和Collect Earth平台进行目视解译，以确定潜在湿地样本点的类型属性。本文开展大洲尺度的全湿地类型样本生产，结果表明：本研究共生产了150 688个湿地样本点，其中内陆湿地样本点为121 412个，滨海湿地样本点为11 563个，人工湿地样本点为17 693个。13种湿地类型中，湖泊样本点占比最大（39.22%），潟湖样本点占比最小（0.19%）。本文生产了稳定、高质量的湿地样本点，样本数量充足，空间分布合理，能够为湿地分类器的训练和分类结果精度验证提供可靠的数据基础。

Sample collection is one of key research foundation for wetland mapping, which play an important role in classifier training and accuracy validation. Generally, wetland samples are produced by visual interpretation based on high spatial resolution images or automatic generation based on multi-source existing dataset. The visual interpretation is time and labor consuming, which cannot meet the demand for large-scale wetland classification. The automatic sample generation method is not suitable to detailed-type wetland mapping, due to the diversity of wetlands and classification scheme inconsistency of existing wetland datasets. Thus, the efficient and accurate sampling method is in demand for large-scale and detailed-type wetland mapping. In our study, we collected a series of auxiliary datasets, and developed an efficient solution for continental-scale wetland sample generation by combining automatic sampling method and visual interpretation. In the first part, the samples of 5 wetland types could be automatically generated by rules filtering based on multi-source existing datasets. The river, lake and reservoir samples were created by using JRC Global Surface Water (JRC-GSW), Global River Widths from Landsat (GRWL) and HydroLAKES datasets. The coastal swamp (mangrove) samples were produced by using Global Mangrove Watch (GMW) dataset. The tidal flat samples were generated by using the Global Intertidal Change dataset. In the second part, combining time series of MODIS NDVI images and existing auxiliary datasets, we first produced potential wetland samples for coarse wetland types (i.e. vegetated wetland samples and inundated wetland samples). Then, we identified them by visual interpretation based on Google Earth Engine platform, Google Earth software and Collect Earth software. We applied our sample method in our study area, and produced continental-scale and detailed-type wetland samples. The results indicated that the total wetland samples in our study area were 150 688, of which 141 412 points were the inland wetland samples, 11 563 points were coastal wetland samples, and 17 693 points were human-made wetland samples. Among the 13 wetland sub-categories, the lake accounted for the largest proportion (39.22%) and mainly distributed the northern and central of study area, while the lagoon accounted for the smallest proportion (0.19%), mostly scattered in coastal region of the study area. Meanwhile, the samples of river, reservoir, inland swamp and inland marsh also shared considerable amount, accounting for 16.93%, 9.86%, 7.16% and 11.12% of total wetland samples, respectively. The river samples mainly distributed in northern and southern of study area, the reservoir samples mainly scattered in southern of study area, while the inland swamp and inland marsh samples mostly distributed northwest and southern of study area. This study successfully produced stable and high-quality wetland samples at continental scale. The generated samples shared sufficient quantities and reasonable spatial distribution, which could lay good foundation for classifier training and accuracy validation. Meanwhile, through combining the multi-source thematic datasets and multiple platform, our designed sample solution could not only make full use of the existing database and greatly reduce manual workload, but also created high-quality samples for complex wetland types, such marsh, swamp and floodplain. Overall, the designed sample method in our study was efficient and reliable, which was of significance for large-scale wetland mapping.