QGIS 4 Arch - Satellite Remote Sensing with Sentinel-2 (SCP, NDVI, Band Combos, Raster Calculator)

in this tutorial we explore satellite remote sensing for archaeology this begins with a brief introduction to multi-spectral imagery including the common types that are freely available then we install the semi-automatic classification plugin to help download data this is a powerful plugin that we will only scratch the surface of in regards to its functionality this is followed by a discussion of common band combinations used by archaeologists and other remote sensing professionals finally calculating ndvi and the importance of comparing various data round out the lesson as always time stamps are available in the description so you can jump from section to section okay let's get started for the most part when archaeologists discuss satellite remote sensing we are referring to imagery collected by multispectral satellites these satellites contain remote sensing equipment that captures image data within specific wavelengths which range across the electromagnetic spectrum we access this data as separate files each matching a different range within that spectrum archaeologists tend to use blue green red and near infrared bands which is accomplished by combining these bands in different ways of course other bands can be used for different analyses as well there are many different multi-spectral satellite systems orbiting the earth these include the various landsat missions managed by the us government which have been recording data since the 1970s the european space agency manages sentinel missions which have been running since the 2010s other examples include the aster and modis systems also of course satellite imagery by services like quick bird and iconis exist but these of course focus on visible wavelengths think of something like google satellite services this tutorial focuses on sentinel 2 imagery i'll discuss how to use different bands from this satellite and do note that bands of different satellites do not necessarily record the same wavelengths in other words the bands we use from sentinel 2 may be different for landsat 8 or landsat 7 and so forth so paying attention to the wavelengths associated with each band in this tutorial will allow you to complete similar analyses using data from other satellites refer to the description for links about comparing bands between these different satellites as you move through this tutorial but also if you go out download other data and want to do the same types of analysis we're going to use a plugin to download and work with data in qgis of course you can download satellite imagery from various websites and i've included links to some of the most common ones in the description now the plug-in we'll be using for this is the semi-automatic classification plug-in of course first we need to install it so come to the manage and install plugins and search for semi automatic classification plugin install the plugin and once it's installed you'll notice some new panels and toolbars have been added you can close the panel i won't be using it we won't be using it leave the toolbar because that's actually a useful way to access this plug-in quickly now be aware that this panel we just closed will reopen itself oftentimes so that's kind of up to you if you want to keep it open or close i will often just close it to make more space in my layers panel now we have to change a couple settings in the scp before we can use it and the first one deals with downloading our products so come to the login data tab and make sure you've created credentials or accounts at each of these three websites copy the user and password data into these fields and then finally for the sentinel data note that there may be a different address for the sentinel data portal and i've included a link to this new address in the description below finally go into the settings tab and check the temporary directory settings i'm going to reset this to what would normally be automatically loaded here and you can see that there's a space in my directory structure so i need to go ahead and change this to something that doesn't have a space if you have a space in that directory or folder structure this plugin likely won't work so i just create a new folder specifically for this plugin and select that in one of my main drives okay with that all set up we're now ready to download sentinel 2 imagery using the scp so now we'll go ahead and download some sentinel 2 imagery with the scp i've already added a base map to qgis and navigated to an area that i want to download data for so in the scp we'll use the download products and search tab to set up this area so i'll just drag this a little out of the way and use this button that's set area in the map and the way we do this is we click with the right hand mouse button first and then we set the other corner with the left mouse button and so now when we go back into the scp or the plug-in window we can set the product sentinel to you can see all sorts of other ones are available here and again remember you have to have all these correctly set up for data to download you can then also select a date and i'm going to change this to be a little more recent i'm going to set the results to 40 just to give myself a few more results and then click the find button this may take a few minutes this is accessing the sentinel download website i'll go ahead and pause the video and resume once it's finished searching okay and we're back now for me that took about two or three minutes to download one of the main attributes that we really want to look at here is cloud cover and as you can imagine the cloud cover this is basically a percentage of how much of the area has cloud cover so we want something with a low amount of cloud cover also pay attention to the acquisition date because these can be different times throughout the year and of course since one of the things that we're examining with multi-spectral satellite data is vegetation and then of course you can also change the amount of cloud cover so if i don't want anything above say 10 cloud cover it would remove everything above that and i would get results going back further in time i like the idea of something with almost no cloud cover so let's click on this and then we can see this previewed here so if we click on the display preview of highlighted images in map this will actually download a very low resolution but small file version of what we're looking at so let's go ahead and download this one and we want to go ahead and pay attention to some of these buttons down here i'm going to unclick the load bands in qgis all that would do is once we start downloading these bands it would automatically add them all to qgis i don't need them all added so i'm not going to do that i'm also going to unclick pre-process images we don't want to do any sort of pre-processing to these images although the scp has some very robust pre-processing tools included so we're going to turn off pre-process images we can leave only if preview and layers because this is previewing in layers if we uncheck this and and don't have or even do have something selected it's going to start downloading all of these so go ahead and only if preview is in layers make sure it is in the layers panel this is click then we can hit run and this is going to ask us to save this somewhere so i'm going to go ahead and create a folder within that larger temporary directory and select that as my download point we can see the status download up here we can see it downloading various information and if we open that folder on our hard drive we'll eventually see it populating with these different files sentinel 2 data is pretty big so this is going to take a while to download okay these have now finished downloading and we can look at these these are jp2 files and as we hover or click on each one we'll see that they're different sizes and we'll talk about these different bands you can see this labeled band 1 band 2 and so forth you're going to see major differences in the file sizes between some some will be similar band 2 band 3 band 4 are all going to be fairly large but if we go down to band 5 for instance we'll see a significant drop it's about a third of the size and so what we're seeing here actually is of course larger amounts of data and that's because each of these bands have different spatial resolution bands 2 3 and 4 have a spatial resolution of about 10 meters per pixel meaning we have pretty good coverage here much better than say landsat 8 or other multi-spectral satellites now if we go back to qgis we can actually exit out of our scp let's remove this preview layer and navigate to what we've just downloaded and we'll see here a number of different files have been added we're of course interested in the actual image file not the layer information file that qgis reads let's go ahead and click on band 2 and bring it in it won't make a lot of sense now but as we start to combine these different bands in the next section of the tutorial we're going to start to create some really stunning imagery that we can use for various forms of archaeological analysis okay now we're going to explore three common sentinel 2 band combinations the equivalence of these and other band combinations for landsat 8 are provided in the description so if you're working with that data set you can just scroll down and see what these values or these bands would be for that data set now we're going to go ahead and add bands 2 3 4 8 and 12 to qgis okay so these have all been added i'm going to go ahead and collapse these and quickly rename them so i'm just going to highlight them press f2 and that lets me type in a new name so i'm just going to name that band 12 use the down arrow key f2 band 8 and just quickly do this for all of these different bands and this is going to make it easier for me to read this as i'm using various tools okay so the first bang combination we're going to explore is the natural color band so let's go ahead and turn all of these off and we can just see our google roads layer down here if we turn that off in the google satellite we can see we're going to actually end up with an image that looks similar to this and the way we do this is we go into our processing toolbar and search for build virtual raster and so for natural color we're going to use bands 4 3 and 2. now take note that this of course puts everything in basically alphabetical order the actual band combination we would need for natural color would be four on top three in the middle and two on the bottom so keep that in mind it'll become more clear in a moment when we're symbolizing this but for now go ahead and click we're not too worried about any of these settings except for resolution we want the highest i'm going to go ahead and let it just save to a virtual temporary file click run close this as we can see this has created an image but the colors seem a little off and that's because the bands aren't properly lined up the red band should be band 3 the green band is band 2 and the blue band should be band 1. now keep in mind that band 1 2 and 3 do not actually obviously correspond to these bands it's given them basically virtual or false band numbers so we have to keep this in our mind when we're using that build virtual raster tool in this case band 1 was actually band 2 band 3 is band 2 band 4 is band 3. i know that sounds a little confusing but you can make a note and just keep that once you've worked with this a couple times it's going to be easy to sort of make sense of and we can see now this that this looks much more like a natural color image in fact if we turn on google satellite and scroll down we can see where they meet you can continue working with the layer styling for this i often find that reducing the brightness can make this look a little bit more natural but of course keep in mind that any differences between these two images is partly because we're looking at data in this case that was updated just a few days ago and so any differences here have to do with um likely with that so we're seeing here maybe the results of rainfall or lack of rainfall we don't really know what time of year the google satellite image is from we do know what time of year our sentinel 2 data is from okay let's go ahead and turn this off and do our next band combination which is a color infrared combination and in color infrared band combinations vegetation really pops out in red heavier vegetation being more vibrant it's easier to tell different types of vegetation apart in this color infrared band combination than in the natural color image so this is very commonly used band combination and remote sensing when looking at things like vegetation crops and wetlands and of course as archaeologists vegetation crops and how archaeological features may affect those is really important let's go ahead and turn off google satellite just turn on the google road again of course to create a new band combination we're going to build a new virtual raster and in this case we're going to use bands 8 4 and 3. so click ok again highest i'm going to leave this as a temporary file and click run okay let's go ahead and quickly rename these and call this one color infrared i'll rename the one below it to natural color of course again we have to make sure that these bands are properly lined up so band one should be band three band two and then band one and so unsurprisingly we're not seeing a lot of vegetation out here but we are seeing some red along the river banks which makes sense that's where we're likely gonna have most of our vegetation out here okay finally we're going to go ahead and use one of my favorite band combinations just because it looks really neat particularly if you're working in an area that does have a lot of vegetation and this is the natural with atmospheric removal and here we're going to see that vegetation shows up in more vibrant shades of green so let's go ahead and turn this one off and create that band combination we're going to build another virtual raster and in this case we're going to use bands 12 8 and 3. click ok highest again and run click close let's rename this to i just usually call this natural with atmo removed again we're going to have to make sure that the bands are set up properly so we're just going to basically flip the first and third and here we're going to see i might darken this a bit maybe even more and we can see here again vegetation popping up in green so if we turned on the natural with atmospheric removed and the color infrared and clicked between them we can see the red and the green lining up and so we're seeing vegetation again now we're already starting to see some features on the landscape pop out if we turn this off and we turn on our google satellite we can zoom in and see that this is actually an archaeological feature here and so it's already showing up in some of these sentinel-2 band combinations our final form of remote sensing analysis with sentinel 2 imagery is ndvi this stands for the normalized difference vegetation index ndvi is a simple graphical indicator used to analyze whether or not the target being observed contains live green vegetation more vegetation or greater vegetation vigor may be associated with better general plant health or with buried archaeological features that affect vegetation health atop it of course keep in mind that vegetation values fluctuate depending on the time of the year it's one of the great things about these data sets is you can download them at different times of the year stretching back years and in the case of some landsat data even decades so calculating ndvi is very straightforward and involves a simple equation it uses the near-infrared and red bands so this equation is basically the red band is subtracted from the near infrared band and this is divided by the near infrared band plus the red band for sentinel 2 this requires band 8 and band 4 which we've already loaded here so of course to do this we're going to use the raster calculator open up that window and of course we're working with these bands down here so what we want to start with is we want to begin our equation band 8 or near infrared minus band 4 or our red band close that divide and start a new area band 8 plus band 4 close this we do have to specify an output layer here so let's go ahead and maybe head to that directory we've been working in so far and just name this ndvi click ok that'll process for a few moments and then we'll have our ndvi raster so the resulting raster will have values between negative one and one the closer the value is to one the higher the overall vegetation value while the closer the value is to zero the closer the landscape will be to baron values between 0.2 and 0.5 are associated with scrubs shrubs and or grasslands values between 0.6 and 0.9 are associated with denser vegetation we now have our ndvi label loaded archaeological features of course affect vegetation in a variety of ways in some cases completely different vegetation types may grow atop archaeological sites while in other cases the general vegetation signatures will be stronger buried walls or decomposing organic matter may retain higher amounts of moisture and increase vegetation density in these areas now qgis comes with some nice symbology pre-loaded or available for ndvi we change to single band pseudo color and we're going to choose create new color ramp go to opt city or sip city okay and then we're going to select the qgis set of values scroll to the bottom and here we have an ndvi so go ahead and click that i often change the number of classes here that's probably just a personal preference reclassify this of course you can tweak the symbology as as needed i will often force these values back to negative one and one and that will often times actually produce a much more clear ndvi data set so here we can see a lot of dry and barren areas but also some vegetation again that's of course corresponding to waterways creeks and so forth of course we've only scratched the surface in regards to the types of archaeological analyses that are possible with satellite remote sensing data for instance you could create two different ndvi data sets perhaps from different seasons of the same year and use the raster calculator to calculate the difference for instance subtract one from the other experiment which one to choose and symbolize them as reds where the deeper the red the greater the difference also combine these different band combinations with your ndvi lower the opacity to help bring out features between them you could even do this by just simply having your ndvi on top of an aerial image and lowering the opacity to help visualize or draw out features on the landscape here in this area we can already see several of the archaeological features become highlighted when we combine techniques like this so as you explore various types of analyses that are possible with satellite data let me know in the comments if you run across something that you'd like to see a specific tutorial on for instance read in a journal or a book about a type of satellite remote sensing and perhaps the analysis is difficult to learn with the tools we've already explored here i'll go ahead and keep making some other tutorials about this for instance exploring all all of these other types of functions in the scp that would be useful for archaeology in upcoming videos so as always links to location of data are in the description make sure to hit that like button and subscribe to get future updates until next time keep mapping the past [Music] you

Channel: AnthroYeti

Views: 2,461

Rating: 5 out of 5

Keywords: QGIS, GIS, archaeology, tutorial, tutorials, Sentinel-2, SCP, Semi-Automatic Classification Plugin, Satellite Imagery, Remote Sensing, NDVI, Raster Calculator, Multispectral Satellite Imagery

Id: A99pMv4CFWg

Length: 21min 37sec (1297 seconds)

Published: Wed Jul 15 2020