Hemp used to be farmed across the United States, but thanks to its association with the psychoactive form of cannabis, the government banned the crop from commercial and university fields for most of the 20th century. Now, hemp could once again become an American staple. For that to happen, researchers like Post—employees of land grant universities, which are located in every state and are federally mandated to help American farmers succeed—can fill in the knowledge gaps that have appeared and widened over decades. “We get tens and tens of questions each week that we can’t answer,” says Post.
These gaps include how best to plant hemp, what varieties to use, which insects and weeds are most likely to cause problems, and, most important of all, how farmers can turn a profit.
We need to find the field filling for University of Kentuck. They did a full section in 2017 the required paperwork but be field by April 20th😬.
I hear the numbers very interesting.
I try to find the USDA request for grow. That will give us the location to use https://earthengine.google.com/timelapse/ to compare the field from plow to harvest. For biomass you need to compare images the far-reds. One of the options is spectrum.
We may even be able to tell a bit about the absorption spectrum for the crop at each picture. @GrowFlux and @eyal? Any know a bit about satalite spectacular analysis? I only have a bit from a satalite mapping class in 1987. But, with the goggle product and a photo program, we can compare 2 pictures for changes in far-red by percent. We heve the evaporation index for cannabis! We can calculate the crops biomass.
We have been doing it since 1984 in the big 4 crops for futures pricing. The model on how to do it is published work. You just need really good evapotraparation indexes. We have quit a good one published in 2017 for the same latitude as Kentucky.
You don’t need to create the full index. You only need to take the math provided in the paper, convert it to Reverse Polish notation (RPN) and use Lagrange format for the polynomials. now you have a linear formulaic calculation. What do computers do really well? Yup:grimacing:. Now you say the data is really big. That’s what RDMS databases are really good at.
Take a photograph tool pic any that allow for line image pixel export. I used the resolution of the image to give you the size of each graphical line. I picked 1080p. And loaded the line of data as a raw type in access. In a big dB I would pick a graphical analytic type. This will create the bass for your pixel map of the image. You can take the color value of your far red filter. 710 to 850. And store the HEX light value for every pixel in your data set. You now have 1920x1080 data points (2,073,600 for an exact number)
plus you need the solar intensity index for the location date and time. Generally it shows up on any ag star pic for you or if in the US take the USDA value for your ag district for the time date of the photo. This is your light correction value. It’s a constent for normalizing intestitity we us this number plus the far red number to get a normalized value.
You will have at least two photos to compare. You have to know the exact location of the photo.
Because we just create an ordered ordinal set of far-red for each photo normalized. You can tell the percent difference between the two sets. This gives you all four data inputs that you need for the evapotranspration index calculations.
Now plug each number in to the calculation. And what pops out. The biomass of a field.
On my first try with a baby access database I got a result on my new laptop in a little more than five minutes for 4M rows to compare. Took me longer to convert the math. And I am not very good at it. I usually have a mathematician do it for me. They can do in five minutes what took me 2 hours of hard work.
I got within 2% on a hectare of land used in the Italian study. To move the decimal point two.2% error I only need to double the resolution. most voxel models a logarithmic increase in dataset to improve stitistical reliance. This method only requires an hard stop at resolution. The raw pixel count of the native image.
That not much data. You don’t need that level of persision. If you use two photo taken of plants that are covered by three cameras at 1080p the graphics people say I can get a percent allocation for a 3d voxel. I am comparing two 3d object. But, I can create the value using a modification to Bresenham’s line algorithm instead of doing getting the fill quotdinents. I am give the fill values and want back the difference
Have fun and see if you can do it.
From the voices in my head
From my limited knowledge on this subject. I do know that they dont know much. Hell we as growers really dont know much. Im one of those who knows what, not really how. if that males any sense.
We know a lot. 50k pages of data from WW2 Hemp for victory.
We know that Chicago board of trade thinks we have enough knowledge to grow Hemp. Hemp futures stating this summer.
We know a lot about cannabis. I would not bet against the board of trade.
From the voices in my head
They make cheap thermal cameras for phones, bet you could attach to a drone and get a high resolution image
You can use my iPhone 7 camera with to tools. Procamera https://itunes.apple.com/us/app/procamera/id694647259?pt=949580&ct=recom&mt8&uo=4&at=11laV7
And a tool called polarr. I think it under $100 for the pc or Mac edition. You can also do it in adobe’s light room.
Polorr has a line strip edit function. There are lots of photo programs that support line striping or line posting function. I the basis for photo stegacyphers. You want to ship a lot of secret data do it in a photo. US government
orange book on security 3rd edition.
And do what I said above.
I will help anyone except with linarasation of the math. I do that part really badly.
But, one person on the forum should be able to this.
The line striping is what IBM now uses for pathology reading of slides. The boost of 98.5 % hit rate. They us it everywhere but the USA.
They published a paper on method for simple photograph pattern match with a learning coefficient for continual improvement. Spring last year.
From the voices in my head