 Welcome to the NPTEL course on Remote Sensing and GIS for Rural Development. This is week 11, lecture 1. While concluding week 10, we looked at a lot of indicators that can be used for crop statistics. Of the indicators, we looked in-depth on NDVI, the Normalized Difference Registration Index. And we looked at multiple platforms that can be used to share NDVI data as raw data. For example, Bhuvan Masar's data, you could download it and do calculations. However, we also showcased some platforms where the analysis has already been done and provided as an output. This is important because we do get to see the analyzed data quickly and look into the search aspects. So now slowly what is happening is we are getting high-speed computing facilities, internet and memory capacities as clouds. So these data can readily be downloaded, applied, calculated for indicators. So what is missing? What is missing is how do you apply these data? So now initially we had data issues but finding applications has become more and more difficult. So we are looking at Remote Sensing for Rural Development, datasets specifically. And as I said in the week 10, we started with an analysis of remote sensing need for crop statistics because there are a lot of issues in crop statistics in the current system. A lot of latency, latency is delay and transparency. Transparency means is it unbiased? Is it less prone to human and instrumentation errors, etc.? So there was less transparency in this aspect because we do not know when the statistics was taken readily I am saying for a public or research institute whereas if you use remote sensing based data along with the observation data, there is more transparency. You know exactly when the image is taken. You have to the minute when it was taken and downloaded. Then you also have the proven record of the use of the data through scientific literature and a lot of people have vouched for particular software has driven these remote sensing data. So now we will look into further analysis of these indicators. In week 10, we looked at crop indicators of crop type and crop yield mapping and how they are very, very important for multiple stakeholders. So I am bridging the summary kind of summary between week 10 and week 11 because due to time availability, we have been focusing on each week separately and there is a continued link between each week which we will discuss in week 12 which is the last week. So in week 10, we looked at the need for remote sensing data for crop statistics because there is delay, there is data issues and data gaps along with bias and transparency. Then we looked at remote sensing indicators for crop growth and health acreage and NDBI was found through literature review as a top crop indicator or vegetation indicator. If we remember that when we looked into each of these government portals, the Indian ones, the United States, the European unions, the Indian through Bhuvan, the United States through NASA and Giovanni Earth Explorer and also the European Union's Sentinel hub, we noticed that only two indicators came up as dominant and of that NDBI is the dominant across these platforms and research papers and that is why we spend more time on explaining the theory of NDBI and what data is needed, how do you calculate it? When we step into how do you calculate NDBI, we showed the equation of NIR minus red by NIR plus red and then we said the range is minus one to plus one, giving classes for the range as minus one is water, barren to plus one is the peak healthy vegetation. However, even though this can be done on QGI software, we discussed the possibility of using platforms and the platforms were given as Bhuvan, NASA, Giovanni, Sentinel hub. So in week 11, what we will do is we will build upon these exercise and then showcase that the NDBI has been improved and will stop with NDBI in lecture 10 itself because there are multiple other indicators that we should be looking at. We will have a hands-on quick indicator for water and also look at some other very, very important aspects. So remote sensing indicators database we will go through today in week 11 first lecture followed by remote sensing tools which are aided with crowdsourcing tools. I will revisit the synergized mapping schematic and show that how satellite data, government observation data and crowdsource data can be pulled together in a complete platform and used for rural development. So we will be using QGI extensively in week 11 and showcasing how these data can be used, plotted, etc. We'll go back and forth between Google Earth Pro. I've shown you how to install it, how to run it. So hopefully you could have Google Earth Pro installed and QGI is installed. Very, very important to install QGI and keep it ready for this week's exercise. And then we have, as I said, the rural infrastructures we'll dig into. This is not just a rural remote sensing and GIs course for water management and crop management, but also health care schools, roads are important. The only issue here is most of the population is dependent on agriculture. And for them that is the livelihood. That is where they want to see themselves for the full life. They don't want to come out into urban systems where education demands it. The basic education is done. So as I would say, normally in villages you will see a girl child stopping at 10 and boys also stopping at 9 and 10, 11, 12 maybe, because they enter into farming. So only some have the opportunity to go out and study. So the schools are placed in the villages, but the higher education is outside the villages. You will have to travel and come back. So depending on a lot of social and economic limitations and challenges, you are allowed to go for education. So my father, as I said, came from a village, studied a PhD in the U.S. Through these systems, village school until 10th and then PUC in the main town and then college in Chennai and then PhD in the U.S. Whereas my mother did not have that option, she was stopped at 10th standard. So here's where a live example of rural limitations are there. Slowly this is changing, which is good. Only when it changes, everyone has access to quality education. So we will go through week 11, specifically mentioning these schools, healthcare systems and how they have to be updated by the government using the data from remote sensing and crowdsourcing sources. Then we will also look into some government databases like Mandrega and IWMP and showcase how these could be evaluated further. Used further for bridging the gap between the available data. Then there is a data gap of errors and latency. The most important error in the gap or data gap is or data issue, I would say is latency. How can you prepare for the next 10 years if you have data only from the past 20 years data? Let's say if I need to plan from 2023 to 2050 or 2023 to 2033, I do need data from 2011 to 22. However, it stops at 2011. For example, the census data we have is 2011. The next census data should come out soon, but due to COVID it was kind of stopped. So please think on these terms that for rural development to happen in a very sustainable way, we need to have current data. And sometimes the data is limited due to challenges faced by the government. It's very, very expensive to send capacity to collect data in the villages, rural areas. Cities are more easier because people commute within the city. They get their data, but villages are very, very difficult. And some areas where tribals are there, it is very inaccessible also, forest and livelihood options they have. So that data is very difficult to get. And for those regions, we can use remote sensing data. So that is what we will be using in this course of lecture. We will go through a particular beautiful software, community initiative, volunteer initiative, open source that mixes remote sensing data and open source data. And those who are taking this course, I hope you can also contribute to the community by using the data, contributing data, and also checking if the data is correct or not. So with this, we will start this week's lecture. And what we have noticed that NDVI has been used widely, but also there has been updations of NDVI or developments of NDVI. And that is multiple reasons. One is for site-specific regions and site-specific conditions. Maybe the NDVI did not work well. So they used an enhanced NDVI, E-NDVI is there. And some researchers would put their names in front saying, let's say, dependent NDVI, P-NDVI. So these kind of NDVIs are also there in literature. The base is the same, the ideology is the same, which means that the basic ideology is that a healthy plant will reflect more green and infrared, whereas a non-healthy plant will observe the green color, so you'll see a difference in color reflectance. And that is the basic principle that is being used, but E-NDVI would use hyperspectral, multispectral images rather than maybe red, they would use a different color. And then the principle is the same, so they would use it. Or it could be also crop-specific. So for example, your green plants will always stay green, but then when it grows and starts to yield, it turns to brown, like paddy and rice, I say. Usually people remember wheat as brown, right? But when you go to the field, when it's growing, it is always green. And then when it's maturing, it becomes brown or golden brown color. Husk, husk is the wheat, paddy husk is also brown in color. So these colors are reflected in a different way. So just because green is not there doesn't mean that the plant is healthy. It is maturing, and then it's ready for harvest. So these are captured as different indicators. You will see some of them now. There are multiple, multiple remote sensing derived indicators for the same reason that costly to measure on the ground, observation data is limited. There is a big latency as in gap between the data collected and the data distributed or published. So for example, an observation data, the government will take the data, then bring it back. It takes a couple of weeks to travel from the rural to the cities where officers are there. Then they start working on the data, cleaning the data, etc., etc., takes time. So that is why you see a latency of almost a year, okay? So a year or six months between observation two, putting it on the webpage. Whereas the remote sensing data is there, all you need is a model, which is NDVI is a model already existing. You just apply the model to the observation data and then see how the results come out. On this note, I would like to say that just not NDVI, there are multiple, multiple indicators, a lot of people have done research. And this website I have found is really, really impressive. It gives you almost all the indicators that you would like to assess. And the curators of this website have done really well. It's a database from Germany, a database for remote sensing indices, as it says on this part of the webpage. I'll be happy to explain this in this current lecture in over the next 20 minutes. So let's go through this link. I will click the link and share now. So we are opening the webpage. So I hope my screen is visible. Yes, it is. Okay, so when you go to the start page, normally the database.de, you will come to this. So this is the logos they have. You can look at it as a database for remote sensing indices, indicators. And today, different registration devices exist. But they have not been put in a common document, which is very, very important. Because you should not be redoing what others do. There's a lot of indicators. Just look at the postures and negatives. Look at the literature review and then use it. But where to use it? What is the formula, et cetera? So this person has put in very, very well. And a lot of information is there. We'll just see how this is going to be. So you can give feedback. We go from back. You can put your name, numbers, and give a feedback. Contact is Verena Hendrige and Pratheena Brutcher of University of Bonn. Instead of crop science and then credits, who they give credit, conceptualized and realized by Verena Hendrige, both across Christian Bortzei and Christopher Sandow, and in the data angle Sentinel Hub. Because it's not only Sentinel data, it's a lot of data that has been put up. It's old. They started in 2012, almost 10 years old. But still it gets updated, which is really fascinating. And then how to use it is there are some tutorials, indicators, click, et cetera, so which we'll do now. So what is IDB as further development? They're adding more references. So you don't have to do the literature review. They are doing it for you. And if you have any mistakes, any new data that can be added, which is missing from this database, you can put it in the contact feedback section and then give it to them. Okay, so here it is. You can also do a search, NDVI, et cetera. And then see if you could search for a particular database. Then what you could do is you can actually look into the different sections in this database that can be created. And you can see that how people have been used using and citating. So it's also good to cite it. It's not needed as they didn't put any disclaimers, but it's always good that you can tell your friends and how I'm saying. I've immensely been helped by this website. A lot of my students, I hear this on day one, the PhD students to go through it because you don't have to reproduce what others are doing already. So here it is. You can start from here. You can search the database for a particular sensor, the satellite sensor. So if someone has said in a talk, let's say that list three is used. So if you could click this and then say that what sensor it is, and then you can go down to the particular sensor and then see if it is available. So Landsat eight, for example, is available. IRS is available. And then all these registration indexes, all sensors. It's not only for vegetation. As it says in the writing, it says vegetation, but it's not only for vegetation. That is water, land management. Every other thing that can be used is there. So we can definitely use this for multiple, multiple users. So this is the Indian satellite. And it's saying, like, do you have any indicators specifically built for the Indian satellite? No, it's not. So you can come back and then say, okay, Landsat. So what indicators have been made on Landsat? And you can see that it's tremendously, 114 indicators have been done. So as indicated, we'll just look at NDVI. So I'm just gonna click search here, control F. And if I say NDVI, 26 versions of NDVI are there. It's not just the NDVI, but there is a composite NDVI. I'll show you how it is different. So it is not called the NDVI. It is called Corrector Transform Vegetation Index. So C, TVI, but the NDVI is used here. So somewhere, as I said, NDVI becomes the base and then it gets updated and or regularly improved for a particular region or something. And here they could see, you could see that they're given the formula. So red minus green, red plus green is the NDVI part they have used the visible red or visible green you could use if you don't have NIR. So red is given in the front minus green by red plus green. And then you can see here, the citations automatic calculator automatic. You can go to more info and then it'll give you a beautifully, it'll give you the formula, specific calculated. What are the sensors that have been used? The sensors launch date kind of metadata for it. Very important on the spectrum, spatial resolution, 15 to 100 meters, inclination. And sometimes you also get the temporal resolution. So these are the sensors that have been used are the colors in the sensors that have been used. So we'll go back here, click back the NDVI and, so before populated, I have done it. So let me just say NDVI is from the six. So let's look at another one. So here we have the green normalized difference vegetation index where instead of NIR infrared minus red, it is NIR infrared minus green and that is why G comes in front. So someone has done that G NDVI, green blue NDVI. So instead of NIR infrared minus red, it is green plus blue. And so you'll have to add green plus blue first and then do it. So normally you don't see all these indicators on the bigger dashboards because these are updated or developed further and yet it doesn't have that much of literature review or people using it. So it is not yet as popular as the other indicators. So infrared percentage vegetation index is IPVI. So near infrared by, near infrared plus red by two kind of averaging it. So average, so NIR divided by the average of NIR and red and then NDVI is added to it, kind of multiplied. And then we have V NDVI which is normalized difference blue, near blue. So instead of red, you are using blue and then green, normalized difference green NDVI instead of red, you're using green. And then NDVI C, so vegetation index C so it's a lot of multiple higher, higher end updation cycle. Red blue instead of red, it's just red blue. And then we have the NDVI. So it's not actually 26 because there's double calibrating. Somewhere I would say around 10 to 15 even if you divided by two, you will have around 15 indicators. So there'll be more, there'll be added on to this as an NDVI. So this also actually, for example, this also you can say is NDVI into red because the near infrared minus red, the infrared minus red is your NDVI. So it's kind of 0.1 times your NDVI which is your wide dynamic range vegetation index. So people have used NDVI and from there they've built further NDVI structures. So this is a by sensor. So let's go to the start again and do one by one. So if you can do the sensor, it will first give you all the sensors available here. So satellite is one and then there's a sensor. So satellite is the payload. So first steps are this, the rocket is there. The rocket has the satellite in the nose part or somewhere in the body. It gets launched into space and then the satellite is put into orbit. Inside the satellite, there are sensors. There are cameras and those are different sensors. So here we have different sensors. The mission is different. The sensor is different but one mission can have multiple satellites and multiple satellites can have multiple sensors. So one satellite need not have only one sensor. It can have multiple. Okay, so these are the exact sensors. So you have Sentinel-2A which is very, very famous. And if you can click on the indicators, you can say much, much, much more than 144 that we saw earlier is 250 and counting because of the high, high spatial resolution. 15, 16 days is really good but more important here is the size, the spatial resolution. So 10 meters to 30 meters is a very good resolution especially for developing nations like India where the average land holding size is very small. So think about your average land holding size. So at least you can have 10 pixels into your dominant land holding size in India which is very good to take out crop signatures and very specific crop dynamics for development. So you have 144 in the Natsar 1.1 approximately. Here we have 249 and it is still getting updated. And let's also look at the NDVI's in this one. You don't know how many NDVI's are there, there's 32. And even if you divide by that it's 18 or something, 16 but plus two I'm saying just in case. So at least some higher than the previous version. Why VI is coming to the United States NDVI. So registration tax is really, really high. And you can see the modified M is there which is different from what was there in the Natsar. This is because not only the resolution but the sensors that they use is much different. The sensors that it could be a multi-spectral sensor or a hyperspectral sensor or infrared band is added. So we go back, that is what sensor is. The second and third are really good. So second is very important for us, application. What application do we want to use? Here you will see not only agricultural aspect of rural development, we'll see multiple aspects. Let's say water management for domestic, industrial, water use in agricultural areas, water use efficiency for vegetation. All these are agricultural. Oil availability or how do you sense oil from various indicators. So you can just take it from a band. There's no indicator, you just use a band. That's all it says. So when you click this band, you can also see which satellites are giving this band. Okay, so for example, when I click the band for oil, it's a single band 1040 wavelength. So nanometers is wavelength. And then it says that all these sensors can give. So now what you do, go back to this sensor and then extract that particular band for oil. So indicator is a multiple bands you put in an algorithm, you get an output. So NIR minus red, NIR plus red, NDBI. But if one band is enough, so you don't need all the bands or an indicator, just use the band. And that is what this article is saying. 1040 is enough. And who gives 1040, all these sensors give 1040. So 24 indicators kind of say, you can see. So that is oil, in terms of oil. And then you have the metal, heavy metal contamination, metal iron. So these also can help in associating the land quality and the health quality. Because if it is too much iron in the ground and if it leaks into the water, then people when they drink it, they get really bad health issues. Especially you'll see that in rural areas with a lot of iron oxides present in the soil. And then we have hyperspectral remote sensing and multi-spectral remote sensing. The geology, the ground type, the rock type is there. And then you have the forest. What type of forest cover? This is also linked to the tribals, livelihood options and biodiversity conservation, like animals, birds, plants, herbs, medicinal plants, et cetera. So then you have alpine, and then you have agriculture. So position crop management, crop yield, crop irrigation. Let's look at the crop parameters, what indicators we have. We have the green leaf index, crop water stress. Yes, for sure, we have the NDVI, calibrated and other NDVIs. And then the soil adjusted vegetation index. So you have a couple of indexes here. Vegetation index is there, but then there's a soil adjusted also. When you click it, all the people who have done it will come. Since we have done a lot of crop parameters, let's go to crop yield. And then if you display the indicators, as we have used, NDVI is used a lot. You can use hyper NDVI. So normal NDVI is the normalized difference. Is there, then the pigment specified is there. And then the normalized difference, NDVI minus CDVI is also there. So you can use this to get at the area because if it is green in the healthy growing season, we assume that it is crops, not only for us. So for example, in the Songli region, most of the land is covered by sugarcane. I will not expect a forest to grow there unless and otherwise it is a conservation area. Okay, so that is the agricultural crop yield. Irrigation, land management, precision crop management. What indicators can be used is very, very specific, the chlorophyll content because a healthy plant has high chlorophyll and then the crop water, NDVI, et cetera, et cetera will come. Okay, so I'm going to go back to the start and then say that it is sensor and application. So how do you combine these two? Let's say I want the very later. So Sentinel, I want Sentinel to A and then I'll say what are the applications for that in crop parameters and then display in this size. So you have green leaf, normal. So now we can club as a search, as a query. I want both of them, I only want Sentinel to A and that, where is it used for agriculture? So instead of going to all the crop parameters, I can go through this and then say, find what agricultural crop parameters in Sentinel are available. So you can see I can click the indicator, it has a formula, all the sensors that are being used and then the applications where it can be applied and references. Here's where I feel they have to do more justice is to update these references. Still it's at really 12 date, but again, these are done on voluntary time and stuff. Still what they have as data, as links to the data is really, really impressive. So I'm going back to Sentinel to A. You can say, what are the channels? What are the bands? So there are 12 bands given as starting wavelength and middle wavelength, ending wavelength. So wavelength is a range. So if you say green color, it is a range of colors and then these are the indicators that have been made using the Sentinel to A. We saw that to be 249, which is the same here. Applications, references, et cetera. So as I said, it stops in 2012. They're made to add it. So these are the different spectrums, the colors that are available in Sentinel to A, which we can see here. So all the different types are here. So I'll go back to now this part, show sensor for selected index. So you can select an index and see what the sensor is available. We'll go to NDVI, Hyper NDVI. So as I said, we already know this because we went to NDVI, just NDVI and then we selected the sensor. So these are the sensors that give this particular 1080,000 to 60,000 NDVI. Then we have show sensors for selected application and then show bands for certain. So here it is indicators. What are the indicators for the applications? Here, what is the sensor? I'm not gonna talk about the indicator itself. Let's say, what are the sensors that are giving these data? So these are the sensors. So these are the basic base sensors that are collecting the data and then giving it to you. You do the indicators and then assess the benefits. The operators are here. You have NASA, for example, CSIR or the Australian company. And then you have the ESA, the European Space Agency and then GoldEye, GeoEyes, ROE, et cetera, are all private and Mormon partnerships, et cetera. So you have the CentroNationality, it is specialist, CNES, British Survey. These are part of the ESA also and the USA for sure, the NASA. Okay, so then we go to bands for selected sensor. What is the bands that are available? As we don't know, sometimes we have to search. So for example, we did search for Sentinel in the previous lecture just to make sure that we are in the correct domain. And here are the bands. So it has around 13 bands, eight, eight, eight, et cetera. So eight is around the LAI region and it gives you the colors of these bands where the bands are coming, which is visible plus the near infrared or VNIR. We do have some sensors for that. Okay, so we have all these and then we show applications for selected index or show applications for selected sensor. The index we can do, we can say, as again, NDVI modified, NDVI is there and just the normalized vegetation difference index NDVI. We have the NDVI difference water index and then all the NDVI's are here. And NDVI P and NDVI optimized. So you can just choose a particular NDVI and then see if it actually works along. So let's say I'm gonna choose this one, display the applications, where can it be applied? No, do a random search for another one. We do know that NDVI, NDVI, we know the applications of vegetation and water. So it picks up again, I don't see it getting updated from 2012, which is okay, at least this part, you can get it from literature, surveys and stuff. Okay, so all these NDVI's are getting really good applications, the NDVI's for vegetation and chlorophyll. You can get that information also. So these are the show applications by index, show application by sensor, a list of data available, indices available, references. The indices is the list that we had and you can look at how many sensors indicators are there. All these are driven by remote sensing indicators. So you have 300, I would say 300 plus, let's see how big it is. Around 519 indicators. And all of them, all of them driven by remote sensing. So if you want to use these indicators, if you know someone has told about an indicator for rural development, can directly come here, look at the formula, click on the indicator, it will take you to the references, go back and then get it. So the most important is the formula, how to do it, we have already done it in class using the faster calculator. So these are the lists and then the list of the other indicators, indices, sensors, applications. List of references, what they have been using as references, as I said, most of it is, or all of it is a year. Let's see the latest they have, the earliest is 1960s. The latest is 2011 because the website was done in 2012. They have not updated the references, but the indicators are getting updated because you could see that they're updating all the indicators. And then visualization of sensor bands. You can see how the bands are there for a particular satellite sensor. And for example, Sentinel. So for Sentinel, these are the bands that are in the sensors, not all of it is covered. So the wavelength goes in the bottom, in the X axis. Again, it's just still populating. All these sensors are taken here. The thing about how many datasets they would have used to get it. So if you look at this world view, it has all these big, big colors, big, big bands, and that is why it is expensive. It is not free. You will have to pay for these world view kind of sensors. And then visualization of required index wavelengths. You can visualize that for a particular indicator and stuff where what are the wavelengths that you need? What are the bands that you need? This is a visible, but you need some in the near the visible range also, as indicated here. Most of it you can cover by using your normal, available indicators. There's a lot of information, a lot of data available here. So normally it depends in the green is this one 57539. You can see how it's getting popular. And as you would expect, there's also a red region number. So with this, you can also use the web services. And then you can ask them how to use it directly. So JSON API can be used and that can actually quicken your aspect. This is mostly for learning and seeing what data is available. Then from here, you can go to the database that I initially showed. So as I said, doing this would take a lot of time for a PhD student, for a master student or a master student to read papers, read the applications, who's the owner, what is the formula, et cetera. Everything is given here. Then I expect you to use it directly into your analysis part or your part where you have your values and regions. So this is not region specific. There is no where you can put India, global, Malaysia, Australia, et cetera, or US, but you will have to put it later in your links. So this can give you links to the different sensors and applications, et cetera. It'll actually tell you where you can get different data sets, different indicators you can get. And then let's see if you click on one of these, let's say normalize difference, you can get on the indicators, the sensors that they do, the references for that. You can click on this, it goes to the paper. Hopefully, yes, it goes to the paper. And then this paper, what are the other indicators they have discussed about? All these things are given. Very well done, I would say. And then the applications part also. So I hope you will use this for your analysis and see where it can be used, the sensors that can be used. Applications are a lot in agriculture, lot of references, et cetera. So I'll stop with this view of the database for finding the indicators, the formulas, the wavelengths, the bands, and also for sensors. I'll see you in the next class. Thank you.