Obtaining + Loading Landsat Imagery in QGIS 3.10 (lab 2- V1)

hello my friends welcome to my video on obtaining and loading landsat imagery in qgis so i'm going to show you how to download some landsat data and get it into qgis why even bother downloading the data at all it's pretty old school for you young people who are used to streaming everything your music your video your data files us old school folks like to download the data get it on our hard drive couple reasons for that one is that it usually makes computation much much faster if you're not streaming over the internet and also you are able to interact more easily with the different bands and make the necessary calculations and things like that so why landsat landsat imagery is a great kind of workhorse imagery set this imagery has been being collected for 30 plus years 40 plus years eight different missions with a ninth scheduled to launch in 2021 it has a great combination of temporal coverage uh reasonable spatial resolution reasonable spectral resolution um and it's freely available so it's pretty good so many of you are familiar landsat uh like many satellite imagery is collected in multiple bands these are essentially replicate images of the same scene that are collected in different wavelengths of light so you can see here landsat 8 actually has 11 different spectral bands here they are one two three four so two three and four are in the visible five is in the near infrared and then six seven ten and eleven go further out into the infrared interestingly these have different pixel sizes a lot of these visibles are in i believe 30 meters still this pan chromatic band is in 15 meters this collects light across the entire visible and then these infrareds are 60 or maybe 90 meters so we're going to have 11 different files when we get this landsat downloaded and we're also going to see when we download it that at least usgs has divided landsat into two tiers based on quality tier one is imagery that has been really well radiometrically calibrated so the sensor has been really well calibrated and it's also been really nicely orthorectified and spatially registered so everything's perfect tier two has really good radiometric corrections but lesser quality spatial registration essentially so the exact location of the pixels isn't known quite as well which can be a limitation if you're doing for example a lot of time series analysis or change detection okay so where are we going to get data wow these two websites have great summaries of all of the open data portals i won't go through them but google it or check these sites you'll find lots of places to get data in this video and in this class if you're a student we're going to use usgs earth explorer a lot which has landsat imagery but it's also going to have a lot of other data that's that's handy for us so let's go right into that earth explorer window here you can google it or type in the address you are going to need to make an account and log in it'll let you search but it won't let you download until you're logged in so you'll need to make a an earth explorer account and that's free so once you're logged in we can start our search you can set your search criteria over here lots of different things you can put in there what i usually do is i just usually use the map window to limit my spatial search and in this case we're going to look for an image of lake champlain which is in vermont one trick for landsat is you you usually want to maybe set your search area kind of small because a lot of the tiles overlap and so on and so forth so i'm just going to zoom in on any old part of lake champlain basically here i'm going to go pretty far in though and then i'm going to set over here i'm going to say use map to define my coordinates you can also add them or use a predefined area you can upload a kml or a shape file so there's lots of ways you can specify your your search area you can also set a limit on cloud cover i'm not going to do that quite yet so next i'm going to go into data sets up top here and i'm going to choose landsat 8 and notice this is level it's level one which means it's collected um also means it's collection one which i believe means it is that it's that top tier data it's that tier one data uh so we're going to search that and we'll go down to the bottom okay and we'll look for our results it'll take just a second okay so it pops up cool thing earth explorer gives you these little previews over here and so it only came up with four images i didn't i also did not set a time constraint which you can do but landsat 8 hasn't been around too long so i'm not too worried i can see in the preview this one looks really nice it is already has very few clouds looks like it's from july 6 2020 so it's a nice summer image should show us the vegetation one cool thing if you if you have a lot of stuff you can click this footprint button and it will show you the actual footprint of that image and sometimes you'll have a lot of images so that's pretty helpful okay so to download it we're just going to click on this little button here again you need to be logged in and notice it's giving us lots of options um a couple of these a lot of these options are essentially just to get kind of like jpegs small files that have kind of like a cool you know photo quality export of the landsat image but if you want the real data you got to go to this meaty one down here notice it's 878 megabytes it's almost a gigabyte it's gigantic all these other things are much smaller so look for the biggest one it's in geotiff format which we like to work in now if you don't have a gigabyte of space on your hard drive then you better go clear better go uh delete all your taylor swift albums make a little room so i click download my computer should hopefully prompt me for where i want to download it and actually truthfully getting ready for the video i already did download it so i'm just going to skip this step and you just have to magically pretend that you watched me download it onto my desktop i put it into my gl 322 folder labs i created a folder for lab 2 tutorial and raster and so what i've got down here now is a doubly zipped is a zipped tarball so this file has been compressed twice which means we're actually going to have to extract it twice so if i right click on it windows should give me an option to actually this is interesting windows does not have a built-in option to extract this so you may need to install the free software 7-zip so you may need to take a break now from the video go install 7-zip and then when you go to right-click on this it should show you the options to extract the files so extract the files i'm going to again delete this outer file folder and just have it extract right into the folder and we'll pause the video as it does so okay so it did that extraction so still highlighted here is the original file that i downloaded that was 900 megabytes now that i've extracted it the first round it's 1.8 gigabytes this thing is monstrous and i need to actually unpack it one more time so i'm going to right click again i guess use 7-zip you can also use this shortcut extract here and it won't prompt you to build another folder or anything like that okay there it is all of these uh crazy meaty files so now that i've extracted uh this here's what we see notice we see the root name of the landsat image i don't know what all of these things mean but lc-08 is basically landsat 8 this is level 1tp that's the quality of the data product that you've got the level it's been pre-processed to o14 is the path the satellite path zero two nine is the row the date is 2020 july 6 and i think this might be the date it was processed and released so that's it and then this b one right here that stands for band one as we go down the line it's band 2 band 3 and so on all the way to band 11. a couple quick so you have all the bands here couple quick things to note notice that band 8 is much larger it is 470 megabytes reason is that is the 15 meter pan chromatic band it has a 15 meter pixel size whereas the other bands like band 1 and 2 have a 30 meter pixel size so as a result they're about one quarter the file size also take a look here this is actually the metadata file this is a text file you can open didn't open very well there but it has all the information you'll need to display and interpret a landsat image and quite many gis softwares like qgis will read this automatically but if you ever need that information it's going to be in that metadata text file it's a very important file so one other thing before we go further rather than have two and a half gigs of stuff sitting on our desktop i'm going to highlight both of these zipped files that we just used to extract i'm going to right click and just delete those and then you want to make sure to empty your trash and actually expunge that two and a half gigabytes off of your computer so it doesn't slow you down okay so now let's get these bands into qgis so we can work with them so i'm going to go to qgis we're going to use a tool that is called build a virtual raster so that's going to be under raster off the menu here miscellaneous build virtual raster and that opens up so it wants to know our input layers and essentially what we're going to do here is stack all these bands into a single kind of virtual file so that we have access to them pretty easily and that they don't like each show up as an individual band in our layer list okay so i'm going to go to input layers one cool thing is instead of adding files you can actually add a whole directory which is pretty cool so here i am in my lab2 folder it's not showing me the files in there but i know they're in there so i'm just going to highlight the raster folder where i just where we just unzipped everything so wherever you just unzipped all those files that's where you want to be that it's that folder you want to pick select the folder it shows me all the files that that i could stack together i'm going to actually uncheck the metadata and this bqa so any any of the this text file this bqa and this this angular information so i'm going to uncheck those three then i'm going to actually i don't know if this is critical but i'm going to actually put them in order so i'm going to put b10 down there and b11 down here so now i have band 1 2 3 4 5 7 8 9 10 and 11 all in order they're all checked over here and i've unchecked the other kind of metadata or quality files and now i'm going to hit ok i'm going to set the resolution to be highest just to make sure that we don't lose any do any down sampling or anything like that place each input file into a separate band definitely want to do that don't need to worry about projection differences and i don't think we'll be resampling or anything like that and in terms of the file name here i'm just going to save it to a temporary file for now because if you save it as a permanent file you may end up again with lots of stuff clogging up your hard drive alright let's hit run and see how it does so it finished very quickly so i'll hit close and here is the landsat image here's lake champlain looks like we are in business okay so a couple things you might notice one is we can't see the image very well it's kind of washed out and we're going to try to correct that with a contrast stretch and there'll be more about that in a future video the other is that the image is currently false color everything that you see that is blue here or purple should probably be green right those are probably forests in new england so let's see if we can fix that to do that we're going to adjust the properties so we're going to go right click on the virtual layer go to properties and what you can see here is that this so we're going to be on the symbology tab the render type is multi-band color and it has allowed us to assign a band to to red green and blue and but what you'll notice and if you if you know about landsat or you look back at the slide band 1 for landsat actually is in the blue band 2 is in the green band or green wavelengths and band 3 is in the red so qgis has switched it up a little bit so what we're going to do is we're going to change blue to band 1. so we're going to assign the band that was actually collected with blue light and be displayed as blue for red we're going to do the same we're going to assign the band that was collected in red to actually be red and we're going to hit apply and hopefully have something closer to a true color image looks like we do and then the other thing we're going to want to do is try a contrast stretch so right now we're doing a linear contrast stretch to the min max and so let's try stretch and clip to min max and then so we're saying we do want to do some kind of clipped contrast stretch and if we open this up no change there let's let's try a standard deviation i like how that did make it a little bit darker so let's hit ok and accept that for now zoom in have a look at how we're doing not bad i'm not sure why it still looks a little washed out but we'll figure that out for the next video so thanks so much for listening in the next video we'll show you how to clip a landsat image and maybe a bit more about contrast stretch take care

Channel: Middlebury Remote Sensing

Views: 2,210

Rating: 4.9069767 out of 5

Keywords:

Id: mBk2VIMawRE

Length: 16min 52sec (1012 seconds)

Published: Mon Aug 24 2020