该数据集使用 Google 地球引擎上的 335，709 张 Landsat 图像，构建了 1985 年至 2022 年中国第一个源自 Landsat 的年度土地覆被产品 （CLCD）。我们通过结合从中国土地利用/覆盖数据集（CLUD）中提取的稳定样本和来自卫星时间序列数据、谷歌地球和谷歌地图的视觉解释样本来收集训练样本。通过所有可用的 Landsat 数据构建了多个时态指标，并将其馈送到随机森林分类器以获得分类结果。进一步提出了一种结合时空滤波和逻辑推理的后处理方法，以提高CLCD的时空一致性。
\n“*_albert.tif”是通过 proj4 字符串“+proj=aea +lat_1=25 +lat_2=47 +lat_0=0 +lon_0=105 +x_0=0 +y_0=0 +datum=WGS84 +units=m +no_defs”的投影文件。2022 年的 CLCD 现已推出。
\n鉴于 USGS 不再维护 Landsat 集合 1 数据，我们现在使用集合 2 SR 数据来更新 CLCD。
\n此版本中的所有文件均已导出为云优化 GeoTIFF，以便在云上更高效地处理。详情请点击https://www.cogeo.org/。
\n每个文件中都内置了内部概览和颜色表，以加快软件加载和渲染速度。
\n", "ds_source": "使用 Google 地球引擎上的 335，709 张 Landsat 图像，结合从中国土地利用/覆盖数据集（CLUD）中提取的稳定样本和来自卫星时间序列数据、谷歌地球和谷歌地图的视觉解释样本收集训练样本。
", "ds_process_way": "1、使用 Landsat 图像，构建 1985 年至 2022 年中国第一个源自 Landsat 的年度土地覆被产品 （CLCD）；\n
2、通过结合从中国土地利用/覆盖数据集（CLUD）中提取的稳定样本和来自卫星时间序列数据、谷歌地球和谷歌地图的视觉解释样本收集训练样本；\n
3、通过所有可用的 Landsat 数据构建多个时态指标，并将其馈送到随机森林分类器以获得分类结果。
", "ds_quality": "数据质量良好。
", "ds_acq_start_time": "1985-01-01 00:00:00", "ds_acq_end_time": "2022-12-31 00:00:00", "ds_acq_place": "中国", "ds_acq_lon_east": 135.05, "ds_acq_lat_south": 4.0, "ds_acq_lon_west": 73.66666666666667, "ds_acq_lat_north": 53.5, "ds_acq_alt_low": null, "ds_acq_alt_high": null, "ds_share_type": "login-access", "ds_total_size": 54790873801, "ds_files_count": 70, "ds_format": "jpg.tif.xlsx", "ds_space_res": "30m", "ds_time_res": "年", "ds_coordinate": "无", "ds_projection": "D_WGS_1984", "ds_thumbnail": "9de270f3-b5ad-4e19-afc0-2531f3977f2f.jpg", "ds_thumb_from": 0, "ds_ref_way": "", "paper_ref_way": "", "ds_ref_instruction": "", "ds_from_station": null, "organization_id": "0a4269e1-65f4-45f1-aeba-88ea3068eebf", "ds_serv_man": "敏玉芳", "ds_serv_phone": "0931-4967596", "ds_serv_mail": "ncdc@lzb.ac.cn", "doi_value": "", "subject_codes": [ "170.45" ], "quality_level": 3, "publish_time": "2023-08-25 15:20:42", "last_updated": "2026-01-14 10:51:39", "protected": false, "protected_to": null, "lang": "zh", "cstr": "11738.11.NCDC.ZENODO.DB3943.2023", "i18n": { "en": { "title": "China's 30 meter annual land cover dataset and its dynamic changes (1985-2022)", "ds_format": "jpg.tif.xlsx", "ds_source": "Collect training samples using 335709 Landsat images from Google Earth Engine, combined with stable samples extracted from the China Land Use/Cover Dataset (CLUD) and visual interpretation samples from satellite time series data, Google Earth, and Google Maps.
", "ds_quality": "The data quality is good.
", "ds_ref_way": "", "ds_abstract": "This dataset used 335709 Landsat images from Google Earth Engine to construct China's first annual land cover product (CLCD) derived from Landsat from 1985 to 2022. We collected training samples by combining stable samples extracted from the China Land Use/Cover Dataset (CLUD) with visual interpretation samples from satellite time series data, Google Earth, and Google Maps. Multiple temporal indicators were constructed using all available Landsat data and fed into a random forest classifier to obtain classification results. A post-processing method combining spatiotemporal filtering and logical reasoning was further proposed to improve the spatiotemporal consistency of CLCD.
\nThe projection file of \"* _albert. tif\" is created using the proj4 string \"+proj=aea+lat_1=25+lat_2=47+lat-0=0+lon-0=105+x_0=0+y_0=0+status=WGS84+units=m+no_defs\". The CLCD for 2022 has now been released.
\nGiven that USGS no longer maintains Landsat Set 1 data, we are now using Set 2 SR data to update CLCD.
\nAll files in this version have been exported as cloud optimized GeoTIFF for more efficient processing on the cloud. For details, please click https://www.cogeo.org/ 。
\nEach file comes with an internal overview and color chart built-in to speed up software loading and rendering.
", "ds_time_res": "年", "ds_acq_place": "China", "ds_space_res": "30m", "ds_projection": "D_WGS_1984", "ds_process_way": "1. Using Landsat images, construct China's first annual land cover product (CLCD) derived from Landsat from 1985 to 2022;\n
2. Collect training samples by combining stable samples extracted from the China Land Use/Cover Dataset (CLUD) with visual interpretation samples from satellite time series data, Google Earth, and Google Maps;\n
3. Construct multiple temporal indicators from all available Landsat data and feed them into a random forest classifier to obtain classification results.
", "ds_ref_instruction": "" } }, "submit_center_id": "ncdc", "data_level": 0, "license_type": "CC BY 4.0", "doi_reg_from": "reg_outside", "cstr_reg_from": "reg_outside", "doi_not_reg_reason": null, "cstr_not_reg_reason": null, "is_paper_in_submitting": false, "ds_topic_tags": [ "土地覆盖", "陆地卫星", "谷歌地球引擎" ], "ds_subject_tags": [ "地理学" ], "ds_class_tags": [], "ds_locus_tags": [ "中国" ], "ds_time_tags": [ 1985, 1986, 1987, 1988, 1989, 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019, 2020, 2021, 2022 ], "ds_contributors": [ { "true_name": "杨杰", "email": "yj_whu@whu.edu.cn", "work_for": "武汉大学测绘遥感信息工程国家重点实验室", "country": "中国" } ], "ds_meta_authors": [ { "true_name": "杨杰", "email": "yj_whu@whu.edu.cn", "work_for": "武汉大学测绘遥感信息工程国家重点实验室", "country": "中国" } ], "ds_managers": [ { "true_name": "杨杰", "email": "yj_whu@whu.edu.cn", "work_for": "武汉大学测绘遥感信息工程国家重点实验室", "country": "中国" } ], "category": "遥感及产品" }