简介
Cloud Score+ 是一种用于中高分辨率光学卫星图像的质量评估(QA)处理器。Cloud Score+ S2_HARMONIZED数据集是由统一的哨兵-2 L1C数据集制作的,Cloud Score+的输出可用于识别相对清晰的像素,并有效去除L1C(大气顶部)或L2A(表面反射率)图像中的云层和云影。
Cloud Score+ S2_HARMONIZED 数据集包括两个质量保证波段,即 cs 和 cs_cdf,这两个波段都根据表面能见度在 0 和 1 之间的连续刻度上对单个像素的可用性进行评分,其中 0 表示 "不清晰"(遮挡),1 表示 "清晰"(未遮挡)。cs 波段根据观测到的像素与(理论上的)清晰参考观测值之间的光谱距离对质量保证进行评分,而 cs_cdf 波段则根据给定位置随时间变化的分数累积分布估计值来表示观测到的像素清晰的可能性。换句话说,cs 可以被看作是一个更即时的大气相似度得分(即该像素与我们期望在一个完全清晰的参照物中看到的像素有多相似),而 cs_cdf 则捕捉了估计得分在一段时间内的期望值(即如果我们拥有该像素在一段时间内的所有得分,该得分会如何排名?)前言 – 人工智能教程
Cloud Score+ S2_HARMONIZED集合中的图像与制作这些图像的单个哨兵-2 L1C资产具有相同的id和system:index属性,因此可以根据共享的system:index将Cloud Score+波段链接到源图像。
整个哨兵-2 档案的 Cloud Score+ 回填工作目前正在进行中,随着新结果被添加到 Cloud Score+ 集合中,数据集可用性日期将定期更新。
Resolution
10 meters
Bands
Name | Units | Min | Max | Description |
---|---|---|---|---|
cs | Dimensionless | 0 | 1 | Pixel quality score based on spectral distance from a (theoretical) clear reference |
cs_cdf | Dimensionless | 0 | 1 | Value of the cumulative distribution function of possible |
Image Properties
Name | Type | Description |
---|---|---|
DATE_PRODUCT_GENERATED | STRING | Production date. |
MGRS_TILE | STRING | Sentinel-2 Military Grid Reference System ID. |
MODEL_VERSION | STRING | Cloud Score+ model version. |
NO_CONTEXT_FRACTION | DOUBLE | Fraction of subtiles processed with no temporal context. |
PROCESSING_SOFTWARE_VERSION | STRING | Cloud Score+ processing software version. |
SOURCE_ASSET_ID | STRING | Earth Engine Asset ID for source image. |
SOURCE_PRODUCT_ID | STRING | Sentinel-2 Product ID for source image. |
代码:
// Harmonized Sentinel-2 Level 2A collection.
var s2 = ee.ImageCollection('COPERNICUS/S2_SR_HARMONIZED');// Cloud Score+ image collection. Note Cloud Score+ is produced from Sentinel-2
// Level 1C data and can be applied to either L1C or L2A collections.
var csPlus = ee.ImageCollection('GOOGLE/CLOUD_SCORE_PLUS/V1/S2_HARMONIZED');// Region of interest.
var ROI = ee.Geometry.Point(-119.9087, 37.4159);// Use 'cs' or 'cs_cdf', depending on your use case; see docs for guidance.
var QA_BAND = 'cs_cdf';// The threshold for masking; values between 0.50 and 0.65 generally work well.
// Higher values will remove thin clouds, haze & cirrus shadows.
var CLEAR_THRESHOLD = 0.60;// Make a clear median composite.
var composite = s2.filterBounds(ROI).filterDate('2023-01-01', '2023-02-01').linkCollection(csPlus, [QA_BAND]).map(function(img) {return img.updateMask(img.select(QA_BAND).gte(CLEAR_THRESHOLD));}).median();// Sentinel-2 visualization parameters.
var s2Viz = {bands: ['B4', 'B3', 'B2'], min: 0, max: 2500};Map.addLayer(composite, s2Viz, 'median composite');
Map.centerObject(ROI, 11);
Citations:
-
Pasquarella, V. J., Brown, C. F., Czerwinski, W., & Rucklidge, W. J. (2023) Comprehensive Quality Assessment of Optical Satellite Imagery Using Weakly Supervised Video Learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 2124-2134). PDF