Hello, I have a project about deforestation detection using Google Earth Engine as follows:
1)Polygon definition for the protected area of "Yosemite" and "Sequoia" in California, for a period of time(2000_2023), according to the following article: doi: https://doi.org/10.1101/2021.09.18.460897
2) Using NDVI MODIS, in this step non-forest areas are masked and only forest areas are considered.
3) Time series will be shown and deforestation will be considered
4) In the next step, High-quality images will be extracted for NDVI for each study site
5) A deep learning model should be predict DEFORESTATION
I selected an area and masked the non-forest images and considered only the forest images, the problem is that firstly, I could not select the areas of the mentioned article, secondly, in obtaining the deforestation time series and building the Deep Learning Model, I faced with a problem. I need your guidance and help to solve the mentioned issues.
Thank you
Hosein
I believe you understand purpose of NDVI from MODIS - in my undertsanding is good for Global Mapping. But for conservation area or a project i.e. less 500,00 ha etc., then you need Landsat or other low resolution radar or optical image products to conduct your assessment. Also check Global Forest Watch to see already processed deforestation or landcover change products from landsat.
For Deforestation is it better that we define based on study objectives. Simple definition would be Forest converted other land use. If it is at your sub-national or project level then consider this;
1. Define landcover and land use classes your properly in your initial year image. If you have or planning to do this then good.
2. Define likely human or anthropogenic activities that are expected in the 2nd year.
3. Practically if above and the point or wall-to-wall methods is done properly then you will have a good measurement of newly deforested areas or expanded deforested areas.
Be mindful of forest degradation. Please check this point sampling paper for more insight on a national level approach on deforestation and forest degradation assessment for Papua New Guinea: Article The Forest Extent in 2015 and the Drivers of Forest Change B...
- Badges
- Science topic
- Similar topics
- Data Mining
- Classification