ArcGIS Pro: Image Stack and Band Combinations

okay so now i have this uh sentinel 2 tile downloaded to my local machine and as you see it downloads as a compressed file so a dot zip file so before you do anything you're going to have to uncompress the data so to do that i want a windows machine so there's a built-in tool to do that so you can just do extract all mouse also 7-zip is a good option um it's it's free so you can download it um if you want i'm just going to click on extract all because that seems to work fine in this case we'll do an extract there okay so that's going to extract out the data to a folder and then we'll go in here and look at the the folder structure a bit before we get started with actually analyzing the data okay so here's the folder i'm gonna go into here and you can see there's a couple different subfolders so we have granule um image data those are all of your image bands and you can see there's some additional ancillary data layers here so what we're mainly going to work with is the image data so from this image data folder okay so i already have arc pro opened so um i'm just created a new blank project so before we do anything else let's go in and add a few of the bands so we can see what they look like so i'm going to go to my folder wherever it is there it is um and then we want to go to this folder and then that's our new oh actually do a refresh sometimes when you make changes to folders and they're not like after you've opened up arc it doesn't see the changes so i'm going to do a refresh so yep there we go so if you don't see what you're looking for that might be the issue okay so we go to granule and then image data there and those are our bands so before we add anything in we kind of want to know what these actually are so i pulled up a let's just open up this image here this is just on the uh landsat nasa web page here okay so this is just comparing the sentinel data to some of the other um for some of the landsat sensors so this is the landsat 8 sensors allowing tears and then the landsat 7 etm plus so bands 1 9 and 10 are kind of like ancillary data um if you i have that there we go so band 1 is this coastal aerosol band 9 is water vapor and then 10 is this cirrus cloud band we're not really going to work with those but they're good ancillary data if you need them 2 3 and four are your visible bands so uh blue green red five six and seven are red edge bands eight is a near ir band and we have this eight a which is like a thinner near our band and then out here we have 11 and 12 which are your short wave ira bands as you can see occur in atmospheric windows right so what we're going to mainly work with or are all the bands other than 1 9 and 10. so what we'd like to do now is actually bring some of those in to look at and then make a composite out of them which is a as a multi-band image and just as another note this just lists out all the bands here so you can see the band names the central wavelength and micrometers there and the resolution one thing that's interesting with sentinel data is that the bands are not actually all the same resolutions so for example these kind of ancillary bands one nine and ten they're at sixteen by sixty meter pixel um we are visible in our nearer ten and everything else is at twenty okay so let's add in two three four and eight so right now they're not stacked up as a multiband image they're just separate files so we're going to read in our um 10 meter data first okay so there we go so all the bands came in um zoom in here a little bit so we can see some detail there we go so this is a city here i think this is zurich um and this is showing eight so this is the near ir band and let's turn that off this is the red band 4 the green band 3 and the blue band 2. okay so let's say that we actually wanted to do something with that we wanted to actually stack this up so we have a single image that we can work with so an arc the way to do that is with the composite band tool so i'm going to open up the toolbox here we'll do composite composite bands and then data management and let's go into our folder and we'll just add in the bands that we want image data and then we want two three four five six seven eight eight a 11 and 12. and we hit okay and make sure the order is right from one 2 3 4 5 6 7 8 8 a 11 and 12. so one thing to note is in arc when you're doing a composite and the bands have different band of resolution spatial resolutions it's going to use the pixel size of the first band so this is a 10 meter band which means our output will be 10 meters which means it's going to kind of like down sample or i guess i guess time to up sample the 20 meter bands so i'm going to go into environments beforehand and make some changes so we're going to make sure it spits the output out in whatever projection the original data is rr so um so this is utm zone 32 north and then reference against the wds 84 datum okay um the extent we'll just set that to the same extent as the the um as the input data we really don't need to do that we'll have this snap to the data which again we probably don't need to do that either but i'll just make it a line with the original for building pyramids let's just change the method to bilinear for resample i'm going to change the resampling method to bilinear i generally don't like to use nearest for um for image data and that should be good enough so we'll go in here and then i want to change the name of this just go back to that data folder rs data and then videos i'm just going to say that here it opens we're going to call this zurich s2 for sentinel 2 and then i'm going to use the date so it'll be 10 12 2018. there we go okay so that looks good so let's do a save there and now let's run it didn't like that oh that's right because this is like if if i save it as a grid it's can't that's too long of a name so i'm going to put a dot we'll use an image so do dot img so erdas imagine format which doesn't have the uh file name length restrictions all right cool so now let's do a run and let this thing stack up this can take a little bit of time i mean this is actually a fairly large image it's 10 by 10 meter pixels um just as a side some side notes why this is running the the band the names are actually pretty useful here so for example we've got the year the month and the date kind of as part of the band name and also like the tile that is that is part of in like the unique id and then the band number so they're really good about how they name their data it's actually if you know how to read this you can figure out you know where it is and what the the collection date and whatnot was okay so i'm going to go ahead and cut it here while this is running there's no point in just watching this run and then when i come back we'll um we'll look at just visualizing the data all right so now we have an output of which is a multi-band or or multi-spectral image and again we accomplished that by just simply compositing the bands that we wanted so before we move on let's just look at some of the um data or metadata about this band stack that we made so i go into properties for this new layer and then source and have a look at the information so we see where it's located on my disk the data type the file name vertical units um this is the number of rows and columns of the data the the dimensions so there's 10 10 by 10 and there's 10 bands and we can see it's an 8 bit unsigned short data type and it has pyramids generated and then we have metadata for the bands may not have actually calculated those i guess yet and we have statistics for the bands so you can see the the range of values minimum to maximum the mean the standard deviation on a band by band basis the spatial extent relative to the um projection and then the spatial reference information so we've got it's bcn utm 32 north and again you can see the units are in meters and then yeah so that's just about the projection and then that's the the datum wgs 84 datum okay so that all looks reasonable um let's zoom in here so we can kind of see the detail so this is a fairly detailed image i mean it's not really high resolution it's kind of your high end moderate resolution it's a little higher res than landsat which is near 30 by 30 meters but looks pretty good so let's now play around with changing the band combinations and that'll be the main focus of this video alright so to do the band change band combinations you want to be clicked on the layer you're interested in so we're clicked on our multiband image there and then you navigate to the appearance tab a bunch of options will pop up so what we're going to look at is band combinations okay so there's some default what we're going to do is make some custom ones i'm going to click on custom so when you make a band composite you can change the name so i'm going to just call this s2 uh we'll just call this s2 true color and right now it's kind of reversed so the red is showing actually no mine it is correct so um this is basically a true color composite so red is red green is green blue is blue so i'm just going to add that just so i have it there we go so that looks good right so we have standard kind of true color composite there kind of looks like how we would perceive the landscape let's do a cut another custom one and we'll call this s2 standard fc for false color so now we want to do red is near infrared so remember this is blue green red then we have our three um our three red edge so that means that seven should be our near infrared band and then our green should be mapped to red which is layer three and our blue should be mapped to green which is layer two so let's do add there there we go so it looks like your kind of standard false color composite right so our vegetated area show up is red there because they're reflecting a lot of the near infrared yeah let's zoom out a little bit so you can kind of see the detail there okay so that's your standard false collar composite and let's just add one more we'll do um s2 and then we'll say well it's called false collar too i know a good name for this so another common um symbology is to do the shortwave to red so would you choose 10. and then the green will show the near infrared which again was seven and then blue will show green which will move to two so let's add that there we go so this is a really common composite you see for sentinel data and then also landsat data so it really does a good job of you know highlighting the water versus the developed and or barren areas that look kind of this purple collar and then obviously the green areas which are generally your vegetation right um so let's for a second let's think about why this looks the way that it does so why in this case is vegetation showing up as green well so remember we have our mapping our band mapping here is red is shortwave green is near infrared and blue is green so vegetation is going to tend to absorb a good bit of the shortwave and the and the green so that means it's mainly going to be dominated by the green collar which is showing the near ir reflectance right um in contrast these like developed areas they're going to be reflecting a lot more of the shortwave infrared radiation which gives them this kind of pink color um yeah so anyway that's kind of the idea there does a good job of differentiating surface materials okay and then lastly if we want to switch between these band combinations then now that we've created them is really simple you can just click on this drop down here and just click on it and it'll switch it back and we can switch between our different band combinations okay so there we go that's our standard um those are some um band combinations here in in arc pro so we're gonna end this one here and then in the next video we're gonna look at some other just visualization enhancements

Channel: Aaron Maxwell

Views: 431

Rating: 5 out of 5

Keywords: West Virginia View, AmericaView, Remote Sensing, ArcGIS Pro, Composite Bands, Band Combinations

Id: sf_MAi5pmw8

Length: 16min 11sec (971 seconds)

Published: Sat Dec 26 2020