******************************************************************
File format here was date (M_DD_YYYY) burn unit (BBB) and camera 
name (CCCC) resulting in a path of:
M_DD_YYYY_BBB_CCCC. Plot locations (lat/lon) need to be
referenced from the fort_stewart_firecam_locations.csv as they 
are not in the folder names.
******************************************************************

Files are for FIRE ONLY. Cameras run ~6 hrs. If you are interested in that data contact:
Steve Flanagan: steven@flanagan@usda.gov

Cameras were built by Dan Julio:
https://www.danjuliodesigns.com/products/firecam.html

Link provides access to his GitHub for detailed information.

Cameras export json files of the form:


{
	"metadata":	{
		"Camera":	"firecam-0B39",
		"Version":	"2.0",
		"Sequence Number":	7400,
		"Time":	"11:31:08",
		"Date":	"3/3/22",
		"Battery":	3.826174259185791,
		"Charge":	"OFF",
		"FPA Temp":	36.739982604980469,
		"AUX Temp":	36.980003356933594,
		"Lens Temp":	33.970500946044922,
		"Lepton Gain Mode":	"HIGH",
		"Lepton Resolution":	"0.01"
	},

	"jpeg":	….

  “radiometric”: …

  “telometry”: …



These files, for the fire, are in the “FireOnlyJSONfiles.7z” compressed file.

JSON files are grouped by 100 files, so subfolders beyond this identifier are just how 
the json files were stored.

The first script, “FireCameImage…” extracts the visual and IR images from the json files.
For the IR:
Camera sees itself so the first image (non-fire) is subtracted from all subsequent images.
To account for background temperature, every pixel has “Lens Temp” from the metadata added to it.
In the rare case that “Lens Temp “ did not record, 25degrees C is used.

High and low gain modes, VALUES FROM CAMERA will not make sense.

This is adjusted for and converted from C to K
A neutral density filter is used to prevent saturation (camera tops out at 180C in high gain and 
450C in low gain). THESE ARE HIGH GAIN. This adjustment is also made

Final image is each pixel in K, which is stored in a time array and as CSVS for processing.


Once adjustments are done, the “ExtractedImages” folder would typically contain subfolders 
for every camera. These were instead compressed and saved as “IndividualCameraData.7z”

Each Camera folder has:

“CSVfiles” folder - each IR image (120x160) with the timestamp as a header

“OriginalImages” folder - visual and adjusted IR images

“.npy” file - Python [y,x,t] array of the adjusted IR images

“errorlog” file - the visual and IR camera occasionally fail to record and these are saved here. 

ADDITIONAL IMAGES AND MOVIES CAN BE MADE WITH THIS SCRIPT. This wasn’t done for storage reasons.
Additonal images are useful if you are trying to clip a specific area (visual and IR are set to 
same scale, overlayed, grid applied). Movie is the IR and visual images side by side.



The second script, “GenerateOutputsFinal” does just that. Folder “time_csvs” has some energy 
metrics for each time step of each camera. “summary_stats” file contains generic information 
for each camera (max FRP, mean FRP, fire duration, FRE). 

“FinalImages” folder contains an image of the total temperature each pixel experienced during the 
burn, a histogram of the distribution of those values, and total energy release with respect to 
time for each camera.

WIth [y,x,t] arrays and CSVs for each time step, other outputs should be able to be generated 
depending on the research question. Or, email steven.flanagan@usda.gov to discuss potential
modifications to the code.