RUS Webinar: Land Monitoring with Sentinel-3 - LAND04

so hello everybody and welcome back to another goose webinar my name is Michael Costello and today I'm here back with another topic in this series of roast webinars so today we are using Sentinel 3 data to do land monitoring as you can see in the title of this webinar and before starting let me tell you the objectives of this session so you will learn two main things first how to do land monitoring analysis with sending up tree data and secondly word is the route service and how it can help you in your projects with Sentinel data so therefore for this exercise we will combine both routes and setting all three for our land monitoring analysis so just before starting be aware that this webinar is being recorded and that you will be able to repeat the exercise by yourself but don't worry about that I will explain you how to do it later in the in this session so let's move and let's get started by having a look to the outline of this webinar so we will start by checking our study area for today we see some details that are needed to understand what we are going to do and also to interpret the final result we will then cover the theory behind the use of satellite earth observation for these types of applications and how the data from space can be helpful for land monitoring we will then describe moisture route service and what are the capabilities and then we will use this service to perform the exercise of today and at the very end we will have some time for Q&A session so since we are quite a lot of people here today I would ask you please to send your questions as soon as possible there is a dedicated option in the GoToWebinar so please use it and send the questions as soon as you have them so that we are sure that we can reply to as many questions as as possible so the complete duration for this webinar is going to be more or less one hour three minutes depending on this Q&A so let's get started and as you can see let's go to our study area so this time we are traveling to Cyprus the third largest and third most populated island in the Mediterranean which there are things about this country but we want to focus on climate and vegetation to understand how we affect land surface dynamics so we talk about the geography of the island we talk about topography the physical belief of the island is dominated by two main mountainous areas but all those mountains in the West and the collegia range in the north and a cetera and a central plain known as miss area what the largest river of the island is located so in this image you can see those containers areas mainly this one here and then the one in the north and in the middle who have displayed about its climate Cyprus has a Mediterranean and semi-arid climate with very mild winters and warm to hot summers and as in many other countries of this region rain occurs mainly in winter with summers being generally dry or very dry actually so this is going to be today the area we will use for the land monitoring exercise so let's move now to the remote sensing background well reliable information on land dynamics is required to improve for example agricultural management to map our changing land and to face food security challenges but of course also to monitor the effects of climate change for that different methods can be used to gather this information but satellite Earth observation techniques offer a suitable approach based on the coverage and type of data that are provided the imagery data from The Sentinel satellites enables a new approach for land monitoring the combination of their temporal spatial and spectral resolutions together with relevant analysis can lead to improvements of the decision-making process and this is the context of how satellite data is used in this case or language or so for this exercise we will be using native spectral data provided by the ocean and land core instrument better known as all trip of The Sentinel 3 a satellite in case this is the first time you heard about them The Sentinel satellites are included in the space component of the Copernicus program of the European Commission and the European Space Agency three satellite is let's say the most complex of the central missions since we have a multi instrument platform carrying four sensors that work in synergy what we are using for this exercise only one let me introduce you very briefly the four of them so that you have a good overview of the satellite so first of all we have what's known as the sea and land surface temperature radiometer which is also known as SLS TR it measures the global sea and land surface temperature every day to an accuracy of better than 0.3 Kelvin so that's pretty accurate so for example in this image here you can see actually the first image that was released by this sensor and it's over the coast of Namibia so you can see different colors for different temperature ranges both over the land and also over the ocean so as you can imagine there are RF applications that can benefit from this data we then have the sending of three topographic package which is formed by the synthetic aperture radar altimetry instrument better known as SR al and the dual channel microwave radiometer so together they measured the height of the sea surface waves and surface wind spins over the ocean they provide also accurate topography measurements over sea ice ice sheets rivers and lakes and as an example in this image we can see the height of Antarctica as ice sheet derived from the data of the topography package of Sentinel 3 and then we arrive to the ocean and non-color instrument the ocean instrument which is actually an improved continuity of the MV sad mary's instrument in case you have been working with remote sensing data earlier you may be know this sensor so we are using the data from the sensor today and because of that let me give you some extra details about this sensor so we can see the altie sensor as the big brother of the new spectral imaging instrument of sensing a tube but of course with some difference so first of all we have a wider soft of one sorry of 1270 kilometers with 21 distinct bands in the 0.42 wine pot to 1.02 micrometers region of the spectrum the revisit time is less than two days at the equator once the constellation of Sentinel 3 is completed and the spatial resolution is 300 meters at full resolution and 1.2 kilometers at previous resolution so let me show you a comparison with Sentinel 2 so that you have an example of what we are talking about so here we have a sentinel image over Cyprus it's 10 meters pixel size but if we have a look to the same area in Sentinel 3 well this is how it looks like so yeah nothing very clear we can see more or less the urban area in the middle but nothing else so you can think that in that way Sentinel 3 not very useful if we compare to Sentinel true but of course this is because the applications are not the same for example if we zoom out a little bit we will see the coverage of a sentinel 3 image and if we now compare this to sentinel - well it's very small so as you can see depending on what you want to do you will find something 3 much better than something of - of course if you want to zoom in a lot and and do something very precise sentence is not the earth sensor but then you have some entry and the other way around so as you may know there is only one of the two centenary satellites and up in the space so according to the last news something of will be is coming next 25 of April so stay tuned for that now that there is a hype for rocket launches you can also see the center of a launch next month so let me introduce you the different data products that are delivered by the Aussie sensor because this is very important because once you want to download the data you have to understand the different products are available so we have the og sensor doing its job and they're bringing data back to earth so first we divide the data or yeah the data into late one bit and level two so in level one B we can find the top of the atmosphere volumetrically corrected measurements which are also calibrated and spectrally characterized the pixels are also geo-located so we have latitude and longitude coordinates and provide values in radians units so in level B this is the output you get when the sensor is known sorry when the sensor is in what is known as Earth observation mode there are other modes but I will not go through them since I want to keep this explanation as simple as possible so those top of the atmosphere radiances are delivered in two ways full resolution that is the pixel size has more or less 300 meters and then we use resolution then the pixel has the pixel size is one point two eight kilometers then we have the level two products which consists of geophysical quantities derived from the processing of measurements data provided in the level one products level two products specifically for marine and land applications are generated separately and why well because you do not want to derive the same geophysical parameters over land that over oceans so that's why we speed so for land applications we can get the on sheet global vegetation index the tourist terrestrial chlorophyll index and the integrated water vapor and again those are available in full resolution and reduced resolution for water applications there is also a long list of physical parameters that you can get but I will not go through them since it's quite long and today we are focusing on land just remember you can also get them in full resolution and reduce resolution and finally be aware that the products of level 2 he over land are managed by the European Space Agency and the ones for water are managed by elmeshad so depending on which one you want to get you will have to go to the coder website where the level truthful water are available if you go if you want to work with with line you have to go to either the operations half of Sentinel 3 or once it's in operational mode into the open hub as a sentinel - so if you are interested in knowing more about coda and level 2 we have a previous webinar last month where we used ulti data level 2 water products so you can check that but I will explain you what to find this information later so knowing setting of 3 and the different data products that are available let's now move to the route service so reus stands for research and user support for sentinel core products sorry yes there you go it is an initiative founded by the European Commission and managed by the European Space Agency with the objective to promote the uptake of Copernicus Sentinel data and support research and development activities the service provides a free and open scalable platform in a powerful computing environment hosting a suit of open-source tool boxes pre-installed on paper machines which allows you to handle and process the data derived from sentinel satellites so what does that mean well with large amount of data we have from the sentinel satellites and with more sentinel coming up in the next months and years the challenge is no longer data availability but rather storage and processing capacity so to solve that Bruce offers verbal machines so that you can have the appropriate computing environment to handle the data and in addition to that Bruce also provides specialized user helpdesk to support your remote sensing activities so that means that if you are working with Sentinel data and you are a ruse user you can send us your doubts about how to use something later and a team of remote sensing experts will come back to you with solutions so that you can keep moving with your projects and in addition to all that we also have a dedicated training program such as the webinar we are doing today and also face to face events so all the information of the route service can found induced in those two main websites so let me go very quickly to them so that you can get the complete picture of roots and there we go so this is the first one the ruse training dot-eu and here you can find all the information that is related to the training activities of the route service so something that I want to highlight in this website is the training tab here if we click we can either access the upcoming webinars or the past ones if you go to the past ones actually we see the complete list of webinars that we have organized in the past and for example if you are interested as I said before in all see data for water you can go to the webinar that took place last month and by clicking here you can access the recorded video and also the Q&A session document with some of the most important questions that will reply during that session and also some instructions to repeat the exercise but don't worry I will explain you later how to do it specifically this how to repeat the exercise you can also find in this ruse training portal the e-learning section so here we have different material that can help you to improve your knowledge about remote sensing concepts so we have started so far with a series of sir lectures so here you can find different courses to improve your knowledge about radar and SAR Earth Observation so we go now to the second website which is ruse - Copernicus dot-eu here you can find all the information about the service so something important to know is that it is here where you can purchase it for the service and become a root user as I said we before we offer go to machines that come with pre-installed software so we go to the routes offer and more in detail - computing environments you can actually see the list of software that is already installed in the virtual machines and which is ready to use you can also access the different types of virtual machines that we provide so depending on your project and depending on the requirements you need you will receive a different metal machine in terms of processing and storage capacity so here you can see the different specifications of the reus free virtual machines so knowing that let's see how you can become a loose user so the first thing you have to do is to register for that so we have here this registration tab in the upper right corner so we click here and we click on register so it's very easy we just have to fill in the information and then the account will be activated this process of activating the account can take 1-2 days so expect that it will not be completely automatic so once you have your account you can login on the website so we go again back to the login tab now we click login and if I access my account I will show you how it looks like so once you are in ruse in Toulouse you have this dedicated menu here and the important part is your dashboard so it is here where you will be able to access the viewed function so let's have a look so now what you can see here is actually the toolkit on machines that I have assigned and this is because I have already done the application for those Metro machines but if it's the first time you are joining ruse you will see that there is no table here displayed to you so what you have to do is actually apply for that virtual machine and for that we click on request a new user service so we'll click here it's now when you can tell us which is your project what are you going to do so that we can understand your needs so let me show you the procedures so that you get familiar with it so first thing we have to do is specify the years of experience we have if we have already handled Copernicus data and if we have already processed these data and now we have this box here where the application is asking us if we want to replay to repeat sorry a winner for example let's imagine you want to repeat each webinar from last month so here you have to specify a training code this running code will be provided to you at the end of this webinar if you want to repeat this webinar if you want to repeat all the webinars you will find them in the reus training portal so for example let's let's imagine we want to repeat the webinar which code was LAN zero one so you can very easily here and then specify if you want to use the verbal machine only to repeat the webinar or if you want also to perform additional analysis in your own projects so for example this and now we click next so the next time we have to specify some extra data for example the somatic area also the kind of operation we are we are going to perform and if we want the route service to download the data for us or not so this is actually a very nice feature of booze imagine you are working with a large data set of Sentinel a matrix let's say 8 is 10 team of 3 images and so you need to know these data to the vector machine so you can either do it by yourself or you can ask the route service to do it for you so once you receive your vector machine all the imagery will be there ready to use and you can skip that part and start with the analysis you want to do so in my case I want to do self downloading and now here I can put some extra information about the project think that the more information you give to us the better for us to understand your needs and of course we can also give a name to the project so then we click Next and then we are in the last step we need to specify the type of data we want so today for example we are using Sentinel 3 data we will be using the og sensor as I told you before we can also specify the region so in this case is Cyprus we can also specify the region of interest by uploading a save file containing the polygon of our study area so this is actually very convenient as well and then we can specify if we want to do a multi temporal analysis or not so for example if we say yes well we just need to specify the sensing period if we say no that's all and at the end some extra information again about the project something that you need to clarify about the data etc so once you are done with the application you can just check all the information you have provided and once you do that and agree with the terms of and conditions of the service you can either click here and then submit the request so once you submit a request it will be checked by our by our front desk and they will come back to you in few days that's a 2 to 3 days with a response to your application so they will check your needs they will think about which type of retro machine you need and they will send you all the information within two to three days so I'm not going to submit this request because I have already two baton machines so going back to the dashboard and show you the next step so once you receive the confirmation to your virtual machine this is what will happen you could see this table here in your profile and through this table you can interact with the vector machine so the most important thing that you can do here is it's obviously access the portal machine so you can click here in access my virtual machine and you will go to the virtual machine you can also for example get support I told you Roo's also offers support for Earth Observation doubts so if you don't know how to do some specific things with Sentinel data you can click here and write your doubts and then a team of remote sensing experts will be there to help you in your process so that's all for the introduction to ropes now let's start our exercise for today and for that let's go to the root spectrum shape so for that I you can click here I already have my account sorry my virtual machine open here so before starting because now we are we are starting the exercise just remember if you have doubts during the exercise don't wait until the end just send it to us as soon as as you have the doubt so here we are in the reus virtual machine whoa this is an environment that is based on Linux so if you are familiar with Linux already you you will not have any problem to the app to it but anyways it's very easy this is like a regular computer we have our dedicated internet browser here which in this case it's Mozilla we also have the some icons that we some of the software that's already pre-installed on the widow machine so you can see here for example snap are etc and you also have a dedicated file manager to store all your data so something important to tell about the virtual machine we got a lot of questions about that is how to interact with your own computer so let's imagine you have done an analysis in snap with sensory data and you arrive to your final output and you want to download this output to your computer so you do that by pressing the keys ctrl alt shift and then this side menu appears so now here in device you can click here and you will access all the of the system of the vector machine you can navigate to a specific path but you have for example your your your output and you can download it how to download it's very easy just need to double click on the file you want to download and then a regular download process will start if you want to upload files to the people machine it's again the same you go to the path where you want to save the data and you click in upload file and then just select the file and that's all so let's stop the exercise now and let me give you some extra details about what we are going to do today so in this exercise we will do as standard lindell in the title land monitoring for that we will use the cloud feed Sentinel tree images available for 3 consecutive months May June and July 2017 we will process the data and we will derive a mean NDVI image for each month and then we will analyze the temporal evolution of this parameter to get an idea of how the land surface dynamics are and how it evolved over time so for that the first thing we have to do is to download the data so today we are using if you remember the diagram that I showed before level 1 product so level one products can be downloaded actually from Israel or from elements at so today I'm using the either side and let me show you how to do it so we open the Internet browser and we need to go to the Sentinel 3 pre operations hub so something our three steals is still in pre operations once everything is set for operational mode the the access point Luiza will be the open hub but so far we have these Center 3 preparation so we click here and we access a very similar interval the same interface as the operational so the first thing to do is to login in the system or if you don't have the account to create one so just log in with my credentials so what do you know how to get all the data well the first thing is to navigate to our study area so for that we use the ban option and we go to Cyprus there we go and next we need to define our study area and for that we use the box or the polygon option so today I'm using the box like this so this is going to be our study area the complete island and then we need to specify the data we want so for that we have on the side menu here and the first thing is to define the sensing period so for example I want to don't work not for example but today we are using a sensing period from the 1st of May until the end of July of the same year okay once we have the sensing period we specify the mission in this case only Center of 3 and then we can put some parameters to access the products we won't remember the diagram I was showing before so it's here where you can trigger the database so that you get the appropriate result so we want to use all Chi data we want to use level 1 and we want to have it in full solution so all those acronyms that you see here are the ones that I was showing before so here we have all for all T 1 for and then this is fr4 solution or reduce resolution here we have the level two products so all chi level two land full resolution or land produce resolution the other things you see here are the products from the other sensors onboard Sentinel three so forget about them today we are only using all three level one full resolution okay and once we have this we can just press the search icon and get the results of the query there it goes so you can see actually we have a lot of products available so what to do next how to download the data well the first thing I would say you have to do is to check the image check if it's the one you want to download and for that you can access the details of the image in this high icon here so the press here you can see for example some metadata of the image of the product instrument except of some parameters that might be interesting for you and then we have here this quick look image so this is very convenient because if you are looking for cloud free images you can visually see if your area is covered or not by clouds you know let's say in an easy and fast way so what we are going to be this exercise today is to download all the cloud free images for May June and July 2017 so once you have identified this image you can just click here in the arrow and the process will start so you can see that the image is being downloaded to the virtual machine and remember we are working the virtual machine you are not crashing our computer with tons of data so since we are working actually with a lot of images let me just show you the dates we are using today so we will be downloading seven images in May so here you can see the dates more or less all over the month we are also downloading eight images of June and then we are using book so we are also downloading cloud free images for July so in total we have 22 Sentinel 3 or t level 1 full resolution products that the complete data set we are using as you can imagine I'm not downloading the data right now I've done that already so we can start our analysis straightforward so so the next thing we have to do is to start our analysis in the software we want to use the processor data and for that we open snap so yeah okay so let's open snap there it goes I had it open already so if it's the first time you are sing snap let me give you a very brief introduction to the server it's very easy we have here a product expert where we have the products that we that are opening snap and that we are working with we also have some quick menus here in the lower left corner that help us to work with the data for example the work view to see the location the footprint of the images we are using we have this main gray area here we'll what the display of the images will take place once we open a band and we also have here the tools we need to analyze the data in this main toolbar there are also organized by menus so for example raster tools optical tools or write-up tools depending on the sentinel sensor you are using so let's get started and let's open our true color RGB composition so I have the product already open it's not here but just to show you how to open a project in snap assenting of your product while you click here in the open product icon you navigate to the path where you have stole your image and once you have done download that the center of your image and once you have unzipped the product that you receive from the server you can open the folder and select the XML file this is the file you want to select to open the Center 3 image in snap so I've done that already and here is my image now we can see actually a lot of information only in the product name so we see for example this is a sentence we image from the altie instrument level 1 full resolution and the sense of time so just by the by the name you can already get a lot of information but you can also expand the product access much more for example metadata about the image etc the relevant folder here is the bands folder it's here where you have the measurements the bands plus all the products that that are together so for example in the irradiance for folder we can access the 21 spectral bands of this sensor so to visualize the data we are going to open a true color RGB composition and for that we right click on the product and we select open RGB here we can define the profile for the red green blue channel so I'm selecting a different band combination to show you the true color composition so in this case for Sentinel 3 is 8 6 4 4 the red red green blue channels it would click OK and now we have here the true color composition there it goes as you can see it's very dark and this is because the coils are stretched according to the maximum and minimum values of the histogram of the image so how to change that what we have here the color manipulation menu and what we need to do is to move a little bit we need to move the slider so that we stretch the colors in a different way and we improve the view so I'm just moving it more or less like this and if we do the same for all the the channels the red and blue you will see that we end up having actually a nice much to reddish so we can just change this there it goes well you can turn this a little bit so that you get let's say what we are all used to see in a as a true color composition so we can now zoom into our study area which is Cypress that we have Cypress and we can also move around the image by using the pan option here so you can just use the hand to move around as many other GIS and remote sensing software so now that we have this image we can start our methodology so remember we have 22 Center 3 images in total for the analysis of today so I'm to show you how to do the analysis only for one of the images and then we will use what it's called batch processing to run the same analysis in the remaining images at the same time so let's go through the procedure step by step the first thing we need to do is to create a cloud mask so the the the the objective of the first step of this methodology will aim to remove cloudy pixels from the image for that we will use the I depicts processor that is available on snap which provides a pixel classification into properties such as clear or cloudy land water snow eyes etc the products we have downloaded already contain a cloud mask but can be improved using this tool and in addition to that it allow us to convert the pixel values from radians to reflectance at the same time so remember radiance is the variable directly measured by the mode sensing instruments it is the amount of light seen by the instrument from a surface of an object reflectance on the other hand is the ratio or percentage of the amount of light leaving a target to the amount of light arriving to the target so it has no units so let's create our cloud mask and convert the pixel values from radians to reflectance for that we go to optical pre-processing masking I depicts Sentinel 3 you can see this processor is available for many of the satellites but today we are working with Sentinel 303 so let's select this option so here we have the common snap interface for the tools that are available in the software we have first an input/output tab will define the input of the tool and then we have the output area where we define the path but we want to save the product the output of this tool and also the name that will be assigned to this output so as you can see the default name is the input name plus a suffix that represents the tool that has been used so actually this is very convenient when you are combining a lot of steps because in that way you can keep track of what you are doing so we live here everything as default and then in the processing parameters we can specify which bands do we want in our product do we want to create the mask that I depicts produced together with radiance values in the pixels or in a reflectance so as I said I'm selecting the factor or reflectances today remember top of the atmosphere so here we have the selection and once this is done we just need to press run so I'm not I'm not pressing one because the process takes some minutes and I want to save time for this Wagner so I will show you directly the output which I have already here open so we open this product you can see the name at the end contains the suffix of the tool we have used and now if we expand the product we can see that in the Bands folder we have reflectance instead of radiance remember before it was radiance here so now we have reflectance and the important part also here is the mask folder so here we have all the output of this tool we have the cloud mask snow ice land etc well you can see the information in the snap help so let's have a look actually to this cloud mask and when we double click in ID picks cloud and I'm also going to combine both views and for that I go to windows tile horizontally okay so let's have a look to this mask something I didn't mention is that the mask that is produced by the eye depicts processor is a binary mask so it's either true or false for clouds so what is true and black is false so let's go to our study area to see how well it works so just in and there we go so I would say it works pretty good we have here Cypress we have this area here there's this white area here all those clouds are selected even the clouds here in the middle of the island are identify the house cloudy pixels and all over here well you can see it works pretty well that's why this is a step in our methodology we want to derive the NDVI index the normalized difference vegetation index for this image so I'm going to close this and I will show you how to do it there are several ways to do it in snap but I'm using the bad math option so we right click on the product and we select ban math so we are going to define the expression that will calculate NDVI and we will create a new band containing the NDVI value so I'm calling this band NDVI and I'm making sure that I'm not selecting virtual because I want to save this band as a real file not as a virtual band that will be stored in the memory of the software and once this is specified we can define our expression so we go here and we specify the expression which is this one here so but what we are saying in the with this expression is it's actually very easy we are saying if the pixel is classified as a cloudy so if the eye depicts is true or if the pixel is not in the area classified as land then we will put this pixel the value 0 and if not we will derive NDVI so NDVI incentive three is as always derived using the near-infrared and the red bands so this is the expression so remember we are deriving NDVI for those three cloud-free pixels and those pixels are of course over land not over water it doesn't make sense to drive NDVI over the ocean so this is our expression and once we have it and we see here this confirmation that there is no syntax errors in the expression we can click OK and then ok again so now the NDVI is derived for the complete image and we can have a look so let me show you let me show it to you so there we go that's the NDVI of doesn't look very familiar to us because of the colors so let's change the color go here to the color manipulation tab again and we select basic and now we can change the color ramp so I'm selecting the manage vegetation index and here we see something that is a little bit more familiar for vegetation indices so we zoom in to our area Cypress we see actually that we do not have the complete island with NDVI values this will this has to look okay so we do not have all the area with NDVI values because this area alone remember was classified as cloudy so that makes total sense we can even change for example the color of those pixels that were assigned as zero by going to the table and selecting here and a different color I'm pulling blue because it's water so for example this color here and there we go so in that way we can improve a little bit of visualization of the property so once we have the the complete NDVI hydrograph for for the Sentinel 3 image the next thing we need to do is a subset as you can see we are still working with a complete extent of the image and that's that's it's very large and because we just want to focus on Cypress we haven't convenient to do a subset for that we go to raster sorry we select the image and we go to raster subset and once we are here we can define the new extent in different ways we can use geo coordinates so that alone coordinate we can also use pixel coordinates or even we can draw the new extent in this column here so if we move the mouse we can define this rectangle and we can move it over the study area so I'm using this option today just to show it to you once we have this we can just press ok and the new product will appear here in the project Explorer so you see now we have subset as a keyword to this product so let's open the result which is the NDVI the new subset of the NDVI so we see now that the extent of this image of course it's more that's what we want okay so now we are ready to do the last step of this part of the methodology which is to reproduct so remember the centum - the images are auto j located so we have latitude and longitude coordinates but since we have a small subset we actually it's more convenient but in this case it's more convenient to change the common reference system to map projection and more in detail I'm using today the UTM which is actually the same projection as using something or two so to have similar things how to do that in snap it's very easy we just go to raster geometric projection and again we have the same interface so we make sure we are selecting the appropriate input which is a number 3 with this index here and we make sure we are saving the product in the appropriate path then for the settings of this tool we can define the new coordinate reference system in three different ways you can see here but today I'm using a very convenient one which is the custom coordinate reference system and for the projection I'm selecting the option but if you go down UTM automatic so this is actually very convenient because the software will locate the product in the correct UTM zone and it will adjust the parameters to do the EMU projection so once we have done this we can just press run and the process will start again I'm not running the process to save some time and I will show you directly the result of this tool so let me open the output that I have already processed before there it goes so here we have the final product after the rejection so you can see we have subset as keyword and if we go up to the end of the further we have IDPs and we project it so you know in that way we can keep track of our steps now if we open the same the same file and DVI we can see how it has changed due to the other projection so we see now it has moved a little bit to match the UTM zone which i think is 36 north if I'm not wrong so once we have done we are done with the analysis of one of the images but of course we are working with 22 so we have to repeat this step and it doesn't make sense to do it step by step as I've shown you this is extremely time-consuming but fortunately snap has some features that can help us so and this is called batch processing so - the batch processing the first thing we need to do is to define a graph that contains all the steps we want to do so for that we go to create that graph we go to tools grass builder and here we have this window so we have a read and a write so let's say input output and in between we need to put all the tools that we want to use to process the data to add a tool it's very easy we just right click on the white space and we navigate to the appropriate path where we have the tools so I have the graph already created for you let me show you so this is the graph it's actually the same that I've done before so we start with the read the input then we apply the ID pigs we derive NDVI we do our subset we reproject and we write the output so this is the graph containing all the steps now once we have this graph we can do batch processing so if you create the graph don't forget to save it and once it's saved you do not change any of the parameters here in those tabs because we will do that later so now we close the graph and we can open the batch processing option if we go to tools batch processing so the first thing you have to do is to specify which are the images you want to process so you can use this icon here to navigate to the path of your sending of the images for example I have here the images for me and I can one by one add the I want to batch process so for example I mean let me just have one so that you see it or more that's for example okay so once you have two product what you can do also is to update the metadata if you do so you will see here different information such as the acquisition time the type of product etc so that's actually very convenient next you need to unclick the the option keep source product name and this is very relevant because when you have batch processing if you do not I select this every input sorry every output will have the same name and if that happens you will be over writing your products so we change this and now we are ready to load the graph we have created before so we go to load graph we navigate to the path where we have saved the graph and ok and now we see again the same tabs that I show you before in the birth so the only thing you need to do them is to go through each of the tabs and put the same parameters as the ones I've specified before when I was doing the procedure step by step and once this is done you just need to click run and the process will start so of course I'm not processing all the 22 images right now because this will take some time so I'm going to show you directly what's the output and I will show you the output for the month of May so remember we are well I'm going to close all those products because we don't need them anymore and let me open the output of this batch processing for the month of May so here we have the output of this batch processing step for every of the input images for the month of May so we can see here the date and NDVI just the name I've put two so that we can remember what we have done so we now expand for example this product and we access the banns we can see here the NDVI so this is the NDVI for this image again the color we can change it and oh sorry here yes so we can change the color we can open all the NDVI images etc well you can do well let me just open another one so that you see that they are different so this is actually the NDVI for the 10th of May and this is the one for the 13th and when you have different images displaying at the same time you want to make sure that you are using the same color stretch that you can compare visually the images and for that you select for example this image as reference and we click on this option here apply to all two other bands if we do so and we select all now we are and then we select no here now we are making sure that both images have the same color stretch and we can check that on the histogram so that visually they they are comparable so once we have one here in this once we are in this step the next thing we need to do is to co-locate our products so as you can see i have seven different products each one containing the NDVI of each image what i want to do is to create a product containing all the NDVI defines together and you will see why this is required it's helpful for the for the last step of our methodology so this task is done using the Kincaid tool of snap and in this tool we define a master and a slave product co-locating to products implies that the pixel values of one product the slave are we sample into the geographical raster of the oven the master and unfortunately this is a process that has to be done two by two so if you can't co-locate seven images at the same time so let me show you just two and so that you know how to do it we go to Western geometric collocation here we can find the master and the slave product so for example in this case the slave product which is the NDVI image of the 10th of May will be resample into the raster grid of the master product and they will be combined so as I told you this is a process that you have to do 2 by 2 so you have to combine this and then the output of this operation has to be co-located with the third product and so on and so on so I've done this process already and I will show you the output directly so for that I'm going to open the D product so here we have co-locate me so if we expand this product now and we go to the bands folder we can see now that we have here nd BIA 1 which is the NDVI from the first image we also have NDVI tool and we go down and the VI 3 so as you can see we have all the NDVI products in the same sorry the NDVI files in the same product but as you can see we have many many other bands start together so actually this is not very convenient because it takes a lot of time to locate the different files so the next step is to do a band subset to remove all those bands that we are not interested on and for that we go to raster subset but now instead of doing spatial subset we are doing band subset so here we have all the bands of the co-located product we can just remove select none and then one by one we select NDVI one and keep going and vi2 and so on and so on if you go down you will find NDVI 3 4 5 to 7 so I've done this already let me just open the product so if we do the subset this is what you get the same name with the keyboard subset and now you go to the bands folder we can see that we have nd VI 1 and VI 2 3 and we do not have all those bands that were that before we do have those collocation flags which refers to the quality of the collocation process we than before but okay the important thing is that we have much less vans and it's easier to find them so now we are almost done with the analysis the next thing we want to do is to derive a mean NDVI value for each of the pixels of the image and how to do that it's again very easy we are using the again banned math option which is actually very powerful tool so we right click on the product and we select Panama and now we are going to specify a new name for this new profile that we are going to create and it will be called min NDVI again we unselect the virtual option to save the product and we are ready to define our expression so we click on edit expression and we specify the main calculation so this is actually basic math we say NDVI 1 plus and dbi 2 3 4 7 divided by 7 just a regular average and we say ok and once this is done we have our output which is located in the bands folder you can see it here now we have min and DVI so let me change the colors again using the same procedure as I did before and I'm going to put the blue color for the node data which is 0 in this case yes see ok so this is the mean NDVI pixel values for the month of May so let me just close all those products that I don't need anymore and now I'm going to show you the same product that has been processed already for the month of June and July remember we are processing three consecutive months May June and July so I'm going to show you the equivalent product for June and July so let me just open it very quickly this is June and this is July so again let's open the NTV is the mean and DBAs of those months so here we have June and let's open now July so again make sure once if you are comparing visually different images make sure you are using the same stretch of the course or that you select for example this image this reference and we say ok cell at all and ok and now here we say no to make sure we are using the same stretch ok there we go so we can combine all the views in Windows tile horizontally and let me just move this a little bit and now if we move around we can see the evolution of the mink NDVI value over the month of May in this area here the month of June and the month of July so this is one way to perform this land monitoring let's say in a visual way but there is a great capability in snap which allow us to analyze time series so that we need to open the time series tool of snap which is located in this icon here in the main toolbar so we click here if we click here and now let me just close those views we don't need them anymore I'm just leaving this one for reference and playing this a little bit bigger ok so we have this time series tune and now we are going to create a graph that will show us the evolution of this parameter the main NDVI over the three images we have so for that we first select the images we want to include in this analysis so we sorry I didn't show it we we click here in this icon and we select okay I forgot to tell you something when you'd arrived when you do some type of bad math make sure you save the product so that everything is stored and everything works properly so just right-click safe product okay so I was saying we go to the taxes analysis tool and we open the settings and we specify that we want to include in this analysis the three images we have here opening SAP so we select the icon add opened and we refresh so you can see we have all the metadata we can also change the color of the line of the graph that will appear for example to red we can also change the name is now the graph is called graph one so that doesn't make sense so we can just change it to something you you like so for example NDVI evolution and we say apply so now we can just close this and now make sure you specify which is the common band that you have in the three three products that you want to use to compare the values so we want to compare the pixel values in the mean and the VI pan and we say okay and now if we move around the if we move our cursor over the image we can see the evolution of the mean NDVI values in our time series so you can see in the X direction you have the time so from the first of May till the 15 of July of course our images are those points here very visible but okay and here we have the values in the Y of NDVI from 0 to 0.75 in this case so if we move around we can just see how these changes so let me just zoom a little bit for example here it's kind of stable in other areas it increases etc well that's up to you how how you want to trade and the good thing is that the ones for example you're doing a specific analysis you want to export this graph actually you can so by using those buttons here you can export the file as a ESV to be open for example in next though or you can also export the graph as an image to be used for example in a report or a project etc so you can you can export it as PNG etc so well so that's all for the exercise and just now I just want to now highlight a couple of take-home messages that I think are important based on what we have learned today so let me go back to my presentation okay so if we take into account all the things I have explained today with the new Sentinel satellites the challenge in satellite remote sensing is no longer data availability but rather how to store and process all the information in addition to all that it is necessary to explain how the data can be used and support users in their applications the route service is here to solve those problems by providing virtual machines to store and process the data and by offering a dedicated help desk supported by a team of motorizing experts to help you in your projects with centimeter so die so for this webinar before moving to the Q&A session let me tell you again how you can repeat this exercise by your own if you want to practice the same analysis at dawn or if you want to adapt it to your study area etc so the thing you have to do is to go to the routes - copernica dot EU website register as a loose your service and then request a new user support service and during your application you have to specify as I showed you before the training code so the training code for this webinar is land 0 4 so when your arranged application put land 0 4 and in that way we know the type of little machine you need and we will give you all the data I have used today so that you can replicate the same analysis and you will also receive a step-by-step guide containing all the information I have explained so dice all from my side you can see here training code our websites my contact information and of course we are in social media Twitter LinkedIn Facebook and YouTube so please follow us we post all the news about Bruce there if you want to stay updated with webinars face to face events and everything we do you can just follow us and you will receive all the all the news okay so as I said thank you very much for joining this webinar it has been a pleasure for me to guide you through this topic I hope you have learned something today if you wish to repeat the exercise please feel free to register for us it's free and I'll see you in the next one ciao

Channel: RUS Copernicus Training

Views: 5,428

Keywords:

Id: KtJTj2gZRPc

Length: 62min 19sec (3739 seconds)

Published: Wed Mar 14 2018