17,669 reputation
333112
bio website
location
age
visits member for 2 years, 2 months
seen 14 mins ago

I am first and foremost an ecologist working to promote the integrity, stability and beauty of the biotic community. I enjoy finding solutions to problems on this site that I otherwise would not have exposure to in my day-to-day work.

Specialties: Remote sensing, GIS, spatial statistics, big data, wildlife conservation, and landscape ecology. My work focuses mainly on multispectral and Lidar remote sensing for assessing vegetation structure and composition.

"The feed-forward back-propagation multi-layer perceptron (MLP) is the type of neural network most commonly encountered in remote sensing..." Atkinson and Tatnall, 2010 International Journal of Remote Sensing


Aug
17
comment Random-Forest Classification of 10cm Imagery for species-distribution in R (no point-shapes)
The methods you are describing are common to packages like Erdas where you can grow regions of similar pixel values into a polygon, or define specific polygons by hand that make up a sample in a class. Usually the MLC algorithm takes the mean value of these polygon regions of similar pixels. In this RF case, the coordinates mark individual pixels, which are then used in the algorithm to define what makes up a particular class. This is why it is important to have a large sample size and make sure to use "pure" pixels that make up that class.
Aug
17
comment Random-Forest Classification of 10cm Imagery for species-distribution in R (no point-shapes)
Could you please elaborate a bit @user2999399?
Aug
17
revised Random-Forest Classification of 10cm Imagery for species-distribution in R (no point-shapes)
added 266 characters in body
Aug
17
answered Random-Forest Classification of 10cm Imagery for species-distribution in R (no point-shapes)
Aug
17
revised Random-Forest Classification of 10cm Imagery for species-distribution in R (no point-shapes)
Added tags
Aug
17
comment ArcGIS model builder: How to reclassify 16,000+ rasters and do weighted sums on these rasters per year?
Is it critical to use a weighted sum, or would cell statistics with a sum statistic suffice?
Aug
17
reviewed Leave Open Multiple entries in attribute table for multiple buffers using ArcGIS for Desktop?
Aug
17
reviewed Leave Open Download high res NDVI imagery
Aug
16
comment ENVI to map composition
I would avoid any cartography whatsoever in ENVI and opt for a GIS such as ArcGIS or QGIS.
Aug
16
reviewed No Action Needed Adding Facebook and Twitter sharing to CartoDB infowindows
Aug
16
reviewed Approve suggested edit on How not to repeat text in linesimbolizer?
Aug
16
reviewed Approve suggested edit on Finding coordinates of the image corner
Aug
16
reviewed Leave Open Cause of ExecuteError: ERROR 010067 from arcpy.sa.CostPath?
Aug
16
reviewed No Action Needed Using arcpy.StackProfile_3d within arcpy addin
Aug
15
revised get count of like features types based on field attributes
added 122 characters in body
Aug
15
revised get count of like features types based on field attributes
added 511 characters in body
Aug
15
answered get count of like features types based on field attributes
Aug
15
comment get count of like features types based on field attributes
You are looking for something along the lines of X = 5, Y = 5, Z = 2? How do you want to use the random number generator? What would the output from that look like?
Aug
15
accepted How to create a transparent box in the map layout?
Aug
15
revised Max cell value from a stacked raster
Fixed discrepancy in text and script example.