I am a beginner to Sen-1 analyses and remote sensing in general and am doing some pre-research to figure out if my idea for a Master’s project is possible given my data and knowledge constraints. I would like to create a backscatter time-series graph for a few grassland plots. This ESA tutorial has been helping me through it. However, they mention only images of the same relative orbit can be combined into a multi-temporal stack.
For my objective, the more dates I have for the graph, the better. Yet filtering by a single orbit would reduce the total of ca. 50 images of my roi and period to 14. Is there a way to include other relative orbits when creating time-series graphs?
Please excuse any poor terminologies; as mentioned before, I am a beginner. I am at your disposal for further clarifications!
Creating a multi-temporal stack with different relative orbits in Sentinel-1 analyses can be challenging, as the ESA tutorial you mentioned indicates that images of the same relative orbit are typically combined. However, if you aim to include more dates in your time-series graph, you may explore the option of applying orbit corrections to the images.
This process involves converting images from different relative orbits to a common reference orbit, allowing you to expand your dataset. While this might introduce some complexity and potential inaccuracies, it could be a workaround to include a broader range of data in your analysis. Further guidance and clarifications can be sought from experts or additional tutorials to ensure the best approach for your specific project.
You can process the different orbits separately, apply radiometric terrain flattening and other radiometric
operations in radar geometry, and then terrain correct them onto the same map projection and pixel size. After this they should be easily stackable.