Please refer to the following links for some useful resource for tree canopy density mapping using remotely sensed data.

http://www.isprs.org/proceedings/XXXVII/congress/8_pdf/11_WG-VIII-11/21.pdf

http://landcover.usgs.gov/pdf/canopy_density.pdf

http://www.ijiset.com/v1s7/IJISET_V1_I7_53.pdf

https://www.itc.nl/library/papers_2003/msc/nrm/panta.pdf

As far as QGIS and ArcGIS are concerned, it really depends upon the scenario of your analysis. You may also consider using both of them in conjunction. Although, QGIS is open source and freely available while ArcGIS comes with different levels of user license.

Hope this helps!

Dear Cyrus,

It depends on the degree of precision you need, the approach (quantitative or qualitative) you choose and the images you got.

Papers quoted above work with the FCD model and Landsat ETM+ images. To perform such a model it is easier to use a RS software, like ERDAS or ENVI, because you will need to calibrate your images.

But with a qualitative approach, I think you can just perform a vegetation index and correlate it with your field data. For this kind of treatment, QGIS is enough.