Dear Step forum,
I found something interesting,
So I research the effect of incidence angle close range and far range to Palm, interestingly after all normalization, I found that backscatter intensity in far range is slightly higher than close range. This is opposite from majority papers that I read. Have someone found a similar case? and give me suggestion papers that can support my finding?
I wonder where can you see it (could you give an example?). Maybe dependent on the topography of the particular scene?
Having higher intensity values in the far range of a SAR image might depends on the following:
Sloppy areas located at the far end of the SAR image. If the slope faces the incidence angle of the radar, this will produce a stronger return
Rough surfaces: The land-type cover across the SAR image might differ quite a lot. If the land-cover type located at the near range is relatively smoother compared to the land-cover type (e.g different crops that produce high backscatter energy, maybe houses) located at the far range, then the SAR image values will be higher at the far range
So here is the example, the red part is the far range, the blue one is in the close range. The RGB one is the slice area, where close and far are overlapping each other. I established polygon for three palms in the slice area. Interestingly, the backscatter value of the polygons on far range are almost always higher than in the close range.
Thanks john, but this number actually from the same polygon in the slice area where close and far are overlapping. According to theory, the near range should have higher scattering back due to the distance is closer, so stronger signal. But, in my case the far range is slighlty higher.
What is that you are looking at for your research? is it palm trees? or different kind of crops?
The backscatter energy can also depend on the orientation of the leaves of a tree or crops.
According to weather condition, the orientation of leaves changes, this might increases the contribution of the backscatter energy
yes I only focus on palm trees (sago palm, oil palm, coconut palm). So, for one palm, I made polygons in the slice area, at the end, one polygon will have two backscatter numbers, one for near one for far. The result so far, oil palm has quite stable backscatter value throughout near and far. But sago and coconut are usually having higher backscatter in far rather than near.
I also think that leave orientation or surface roughness could be the reason, since sago and coconut are planted in more open space, so maybe the signal from near range scattered the bare soil more than the palm itself, compared to oil palm (denser).
But it is very rare finding the paper found the same.
Do you have any suggestion papers?
I do not really came across with an paper concerning mapping palm or coconut trees.
In order to avoid this effect of incidence angle, you can apply
incidence angle normalization.
Most scientists recommend to normalize the incidence angle to 30 or 35 degrees. This is as if the SAR sensor had a fixed incidence angle to 30 degrees.
In your case, you need to do a bit of research and see what incidence angle is the most effective in mapping palm or coconut trees.