If I wanted to detect a small (up to 3 m in height) invasive plant species growing within an open canopy pine forest (around 50% canopy closure), and differentiate it from other plant species (small tree species, for example), which of the available remotely sensed imagery products would you recommend? Phenology of the plant suggests that it emerges in April, flowers in August, then goes dormant in November. I may have LiDAR, but it is from one point in time - not sure the month yet. We may collect field measurement within sample plots.
I think by measuring the area of the frequency of light reflected from the native wood surface, and subtracting the frequency of the introduced species.
Hi
It depends on your target size, in this case. If your target ( the amount of land covered by the specious) is big, then you can use some satellites like Landsat (spatial resolution 30 m and free) but if your target is small and spare so you would need finer resolutions like SPOT (10m) or Rapid Eye (5m) or WorldView (0.3m) but not free. or try some aerial photos of the study area that should be available in any country.
Also it matters if you want to detect the specious in individual base, i mean to see like one by one or to see where they are located and general map producing.
Good that you have given the phonological information, it is also possible to use multi-temporal images like Landsat.
Moreover LiDAR data also can be used to validate or check the accuracy of the other above mentioned imagery, plus using some geo-statistical analysis.
You may see some of the available images in this link: http://earthexplorer.usgs.gov/
In short, more information is need but using Landsat and some LiDAR data as validation for some plots is advisable.
The Sentinel 2A Images downloadable for free from the European Space Agency provide a best basement bands for vegetation mapping. the images includes 5 NIR Bands and 1 red band. this high radiometric resolution allow you to best map the vegetation. the satellite takes images from a large surface each 5 days, i think you can do a multitemporel observations. note that the spatial resolution of the sentinel images is 10m for visible and 1 band in the NIR the others are 20m resolution.
Good luck
Thank you for all of the great suggestions - although we have considered some of these issues you have provided a number of new ideas for us to consider.
Dear Pete,
Before you get involved with a whole range of satellite data, some of which are going to be expensive, you may want to go to the field and acquire spectral (preferably hyperspectral) data from the pine forest, the soil and the invasive species. Those spectra should then be ingested into a suitable canopy reflectance model to establish whether there is any objective basis to distinguish between those species on that basis.
And once you are on the field, you may also want to acquire information on the structure of the plant canopies (tree height, crown diameter, DBH, etc.) While both types of vegetation may look "green" and have similar spectral profile, their geometric structure may be sufficiently different that their anisotropic response may be different, and perhaps a better lead to distinguishing between them. Here again, a good plant canopy model should be able to represent the BRDF of these canopies, and indicate a priori whether this might be a worthwhile approach to pursue.
Good luck. Michel.
To add at the very good explanations above, if you need to detect any object in the land, you need previously to know some properties near the object that you are detected at the space imagery at the same time; these properties are known such spectral firm. The spectral firm is the pattern reflectance curve, and it needs a convolution calculus (between land & space spectrum) before its characterization and spatialization of pattern. You would find better information at this link: http://dx.doi.org/10.4995/raet.2014.3230.
What I miss is the seasonality of the dominant indigenous forest floor/under-storey species versus to the invasive. If different,, as is often the case, high temporal resolution imagery would be promising.
Thank you for the responses - you gave us some great ideas to consider moving foward.
- Badges
- Science topic
- Similar topics
- Geoscience
- Remote Sensing