You might want to start by looking at Sentinel 2 imagery at your location (it's free, 10m resolution, multispectral so you have a near-infrared band to use NDVI for vegetation analysis).
If you have example areas for the crops you want to differentiate, you can do an supervised classification to train a model on the examples to look for similar crops. If you don't have examples, you could use imagery from different times and knowledge of the growth pattern / phenology of the crop to identify the crops you'd like to use as examples.
A time series of imagery could be used to monitor growth. Sentinel 2 has a 5 day revisit.
Not sure about the soil question. If you're in the US, SSURGO (Web Soil Survey ) is available for soil mapping; elsewhere local governments generally do that work and make it available.
Supervised Classification | Google Earth Engine API | Google Developers
