• Introduction • Instructions • Citation • Acknowledgment
Using cloudApp, the CDE team learned how to recognize the correct category from challenging image patches (difficulty: 5). First, we acquire all of the images taken throughout the same season with less than 5% cloud coverage. Secondly, we performed a visual comparison between the cloud-free images and the desired IP. Finally, we complete the labeling process collaboratively by conducting independent votes among all members.
App parameters
- run: Render graphics after click?. By default true.
- sensor: Sensor data to be analyzed. By default
Sentinel-2 SR. - lon: Longitude data. If run is true, it can be obtained by clicking on the map. By default -121.68804.
- lat: Latitude data. If run is true, it can be obtained by clicking on the map. By default 36.46517.
- rgb: Image composition of image thumbnails. By default
SWIR1-NIR-GREEN. - initYear: Year acquisition time of the image to analyze. By default 2018.
- initMonth: Month acquisition time of the image to analyze. By default 8.
- initDay: Day acquisition time** of the image to analyze. By default 12.
- cloud: Cloudy pixel percentage** threshold. By default 5
- chipwidth: Size of the chip in the image thumbnail section. By default 2.
- imgid: Image id of the image to be analyzed. By default
20190212T142031_20190212T143214_T19FDF. - llb1: Blue Hampel lower threshold. By default -1.
- ulb1: Blue Hampel upper thershold. By default 1.
- llndvi: NDVI Hampel lower thershold. By default -1.
- ulndvi: NDVI Hampel upper thershold. By default 1.
- llb11: SWIR1 Hampel lower threshold. By default -1.
- ulb11: SWIR1 Hampel upper threshold. By default 1.
Try it yourself here. If you prefer run the cloudsen12_app.js in the Earth Engine code editor.
COMMING SOON
cloudApp is based on the fantastic tool ee-rgb-timeseries created by Justin Braaten.
This project gratefully acknowledges:
for computing resources