Please suggest me any latest technique for crop mapping using SAR data (Sentinel 1A) in combination with multi spectral data.
There are various approaches: In any case I would suggest using time-series of S1 data covering a whole phenological cycle (one year). Maybe you use one image per month to fully cover the variations in preparation, growth and harvesting of the different plots.
1. Each plot with homogenous growing and harvesting conditions will reveal a distinct backscatter curve during the year (= temporal SAR signature). This will help you to identify plots with similar growing conditions. If you can identify the crop types for selected plots (e.g. by multisplectral analysis) you can relate their temporal signatures to other plots with simlar signatures.
Combination with optical data could be achieved to extract the plot boundaries as polygons from optical data first and then use mean values of SAR backscatter (and other parameters such as texture) per polygon. This should help you overcome specle induced variations in homogenous fields.
These two studies rely on this principle and also show limitations:
Article Interpreting ERS SAR signatures of agricultural crops in Fle...
Article SAR signature investigation of rice crop using RADARSAT data
2. Another idea would be to calculate interferometric coherence for your study area. If bare areas do not change between two images, coherence should be high. But if there is tillage or any other preparation on the field, coherence should decrease significantly. This study shows an example on how to make use of it:
Article Potentials of polarimetric SAR interferometry for agricultur...
3. If you have access to polarimetric data (https://www.researchgate.net/post/Which_SAR_missions_and_instruments_have_the_following_properties) you can also distinguish different crop types based on their backscatter mechanism. Wishart classification (decomposition based on Alpha and Entropy values of each pixel) could help you to discriminate between volume scattering (= field crops) and bare areas (surface scattering), such as demonstrated in this study:
Conference Paper Crop classification by polarimetric SAR
The newly established EO college provides excellent free materials on many SAR topics, agricultural applications amongst others. The materials are based on the same multi-temporal approach as demonstrated under point 1 above: https://eo-college.org/articles/resource/agriculture_tutorial/
yesterday an excellent tutorial on crop mapping was published in the EO college:
https://eo-college.org/topic/agriculture-tutorial/
- Badges
- Science topic
- Similar topics
- Geoscience
- Cartography
- Mapping