 Hello everyone, welcome to the next lecture in the course remote sensing principles and applications. We are discussing about the concept of LIDAR and in the last two lectures we discussed about what LIDAR is, how LIDAR system operates from terrestrial platform, aerial platform which can be extended to space-borne platform also. Then we discussed about like discrete waveform LIDAR, full waveform LIDAR and so on. In this lecture we will continue ahead with the topic. So, in the last lecture I described about having like a discrete return or a full waveform return. Discrete return can store maybe 1, 2, 3, some discrete number of returns, it will not store the entire waveform or the signal that is coming back after getting back scattered from the ground elements whereas a full waveform LIDAR will be able to store the maximum amount of information that comes back from the ground. Hence for certain applications especially dealing with like vegetation monitoring and all, the full waveform LIDAR is often used because of its ability to capture the all the most of the features present underneath the canopy. The next important concept that we are going to see is the effect of footprint of LIDAR. So, footprint we have already discussed, a footprint is like kind of the production of the laser beam onto the ground, how much ground area the laser beam covers. So, essentially the returns within each footprint will be going back towards the target and it will be saved at least some of them will be saved. So, each footprint provides us like valuable information about the terrain or whatever features present on the terrain. So, the footprint size is going to influence what features we record and what we store. So, in the initial lecture about the topic I told you that in order to properly capture the terrain we need to have kind of like a large number of footprints. So, like because each footprint will be giving us like 1x y and maybe like multiple z values associated with it. So, if we have like a large number of footprint we may get proper information about the topography. We may also think that having a large number of small footprints may be of real help to us. Yes, that is of real help to us. That is if you have like a large area, if you have like kind of like a topography, this is like x dimension and y dimension, there are like lot of features present here. So, if you have many number of small small small footprints we are going to get we will not naturally think that we will get accurate information about the x y z feature present within a terrain. Say the footprint sizes say in order of like 1 meter or 2 meter. So, even if there is like a large tree that will be captured properly within that like 1 meter footprint like say the foot tree may be standing like this. Say a footprint may be falling within a tree here and the next footprint may be falling like this and the next footprint may be falling like this and so on. So, the single tree is captured in like multiple footprints which may help us to capture the terrain actually like properly that is what naturally we will think. That is also true in most of the cases for like topographic surveying purposes people naturally go for like small footprint LiDAR in order to have like a small footprint plus a dense point cloud kind of like LiDAR especially for topographic applications. But sometimes it may give us some false information to us, false information in the sense say if you are like collecting data over like a forest. So, this kind of like small footprint LiDAR say in the order of like a meter or less than that may actually miss the tree tops properly say an example is kind of given here say this is like a small footprint LiDAR. Say let us say the small footprint LiDAR is seeing here and the next footprint is seeing here. So, what exactly are we capturing? First thing we are missing the tree top. Second thing is this small footprint is going to capture this leaf this stem another stem and so on. So, maybe some return from here. Let us say this is like a discrete return LiDAR system it is going to store only the first three returns if that is the case this will be stored this will be stored and this will be stored. Actually the ground data is going to be missed. So, because of the small footprint we are actually missing like the tree top because the first footprint is here and the second footprint is here missing the tree top and also being like discrete return we are also missing the ground. So, all these things being like a discrete system or having like a small footprint is going to change the way in which we are going to collect information about the ground. But let us think in another form let us say we have like a large footprint system say in the order of tens of meters say 20 meter system 20 meter footprint. If that is the case then the entire tree may be covered properly or even like a group of trees may be covered the example is given in this figure B. So, the group of trees is covered and it is kind of we are getting we will be getting information from the tree top because the entire thing is covered in one footprint. So, definitely the tree top would have been captured this is there this information will be there and this information will be there because on a whole if you combine like what to say the entire thing the maximum returns comes from this. So, having like a larger footprint will essentially capture the overall canopy structure it may not miss the tree tops and all. So, it is always for certain applications it is beneficial to have like a large footprint ladder. But if you want like to perform like a highly accurate and precise topographic surveying then we may actually go for small footprint, but for certain applications large footprint also may give us good results. And also as I told here like having like a discrete and full waveform return is further going to influence the information we are going to collect. So, the effect of footprint may actually give us or may remove some information that we are actually that we want to get. Say another example for like the footprint of ladder is given in this slide. Say here it is like you are having a bunch of trees with like a small lidar system is looking at it. Say let us say we have like many small footprints say here there is one footprint here there is one here there is one and so on. So, for this particular tree essentially if we want to capture the entire tree then we may need to get this information like the tree top and maybe like one or two side elements maybe here or here somewhere. So, that we will be able to characterize this is how the tree looks. Sorry if you look from the top you may have like three returns this is from the top with high elevation relatively higher elevation these two from the sides. So, this is in the top view. But a small footprint lidar system may actually miss the top of the tree may take measurements here or in certain cases for certain large trees it may take multiple measurements over like the same tree. So, either we can miss the tree tops which may actually force us or give us kind of like a sense that the elevation of the tree can appear slightly reduced. Say you have like hundreds of trees if you miss like the tree tops in at least like say 50 or 60 trees within that 100 we are naturally going to get like a false information or we are going to miss the actual height of the canopy. Say this is how it may look you have many number of points say actually this is like the points connecting all the tree tops let us assume this particular line what I have drawn here. But let us say our lidar system is missing at least 50 60 percent of tree tops then what it may give us is it may give us some line like this say the interpolated points may be lying here that is some of the tree tops we are actually missing without even collecting the data which may give us the topography that the canopy structure is highly undulating with lot of like large trees and small trees interposed with among each other such kind of picture we may get. So, but if we have like a large footprint lidar then essentially we are going to get like a collective information that is one advantage of like having like a large footprint lidar for certain applications. And also certain as I told certain large trees we may count like more than one return. So, if at all I told you like even we can count the number of trees over certain areas using like lidar we may like have like a wrong count of the trees in such circumstances. So, the footprint of the lidar and whatever the returns coming in from that particular footprint is effectively going to influence the way we are going to observe the terrain or see the terrain. So, this is really important we need to have kind of like a feel for this or understand the concept. So, not all the points which are like in the first return I told you like if at all you are talking about like a discrete lidar system we can separate the first it all the first return separately second return separately last return separately and so on. So, essentially for like a discrete return system especially with like a small footprint all the first returns may not come from the tree top or building top and so on. It can be captured on either sides or similarly in kind of like a large footprint system it is not necessarily bad a large footprint system can effectively average at least in the kind of like an average way it may bring out like the topography to us. So, based on our needs and applications we need to decide which laser system we have to use a discrete form or like a full waveform lidar what should be the size of footprint that we may need and so on. So, essentially we have to choose based on the applications for which we are going to use lidar system. In addition to giving information about elevation lidar will also store the intensity or the power that is returning back from the ground surface. So, even if there is like a canopy whatever is like reflected from the canopy will be the power will be recorded back by the sensor and that is again is really useful to get more information about the terrain. In the initial lecture I told you that generally we use green wavelength or NAR wavelength for the lasers in lidar system 532 nanometers or 1064 nanometers. If that is the case then let us talk let us talk in terms of like NAR based lidar system. We all know that in NAR wavelengths vegetation will have like a very high reflectance. Similarly, like even soil surfaces has typically higher reflectance or naturally land surface has higher reflectance water bodies has lower reflectance and so on right. But vegetation or this kind of relationship between optical remote sensing and lidar sensing is not really direct that is we cannot always imagine if there is some vegetation present on the ground in lidar intensity image also it will appear very bright. Normally in NAR band image our normal image it will appear very bright because of high reflectance. But in lidar intensity it need not be the case. An example is given in this particular slide. Say in the figure A what we have is like the lidar last written elevation. So, it is giving information about like elevation of different points. So, it is interpolated. I told you that whenever we have like a large number of points the points will be kind of interpolated to give you like a surface because we naturally earth is kind of like a surface not a collection of points. So, essentially we interpolate the points to produce a surface. We can identify trees here. So, these are all some high elevation points. So, bright appearance and elevation map is like high points with higher elevation all these things. Looking at the intensity this is like the intensity that is you can think it off in terms of like the power returned. So, from the return power and also like some other information we can sense here there is like vegetation. This is like again this is like vegetation portion and so on. So, again these are like grass here. But if you look at it the trees which will have like a very high reflectance in NIR portion is actually appearing dark than in between barren surface or open surface. Naturally we expect vegetation to have very bright reflectance than soil surface. But here trees has lesser reflectance actually the grass identified as this particular boxy has a higher reflectance and so on. Why is this happening? Why trees appear actually dark? Trees should naturally appear bright in NIR images. It has a very high reflectance. True it has very high reflectance. But we also need to look at the nature of reflection or nature of scattering happening in the terrain element like in the earlier classes we discussed that reflection can be happening from the surface scattering or it can happen from within like the volume within the terrain feature which we call like volume scattering basically. So, if you have like a large tree there will be like plenty of gaps in between. So, the lidar pulse may penetrate into it and multiple scattering will happen from different different portions of canopy naturally right. It is not only like the one either top portion reflects everything. Some light will always penetrate especially like NIR we have seen that it penetrates through the leaves to certain extent and we also know this additive reflectance property because of the penetration. So, what essentially will happen is lidar is essentially like a backscattering observing system. If the sensor is here essentially the signal should come towards it in kind of like a straight direction for the sensor to record it. When volume scattering happens within the canopy the reflected signal need not only go in the direction of incoming beam it can also go in all other directions because it may became more diffuse in nature rather than being like okay wave is coming like this it need not only go back specularly in this direction from towards nadir and towards zenith it need not be like this. When volume scattering happens it will become diffused and a large fraction of energy will be scattered in all other directions and only like a small portion will go back in the same direction from which like energy came. So, it is not like not even like specular reflection it is like backscattering energy came in and energy went back in the same direction. So, the backscattering is essentially like a small component whereas energy is now actually being scattered into several directions surrounding it. So, that is the reason vegetation appears dark but rather than appearing backscattering if you have like you have like a normal NIR sensor which may see all the information going in certain direction like within the IFOV then it will be able to capture like a some amount of larger signal coming in from the vegetation that is the thing. And also like grass appearing brighter means here in this figure labeled as C so, grass surface naturally there is not producing lot of volume reflection most of it happening within the surface. So, essentially what is coming in from the zenith is going back towards the zenith like this. So, that is like the difference between a normal optical remote sensing system and LiDAR system. So, there is no one to one correspondence between them. We cannot interpret the LiDAR intensity images similarly what we do with the normal images we need to say some caution the patterns may differ but essentially LiDAR intensity provides us additional information rather than just looking at x ways at points if we have the additional information of what is the power return it is really helpful to us it will help us to know what is there on the surface to like a very good extent. But still we need to exercise some caution while interpreting such images. Till now we are discussing about like the general properties of LiDAR systems and in the course like remote sensing course what we are dealing with essentially deals with satellite based remote sensing. So, essentially since we are talking about LiDAR it is necessary for us to know if there are any space based or satellite based LiDAR systems. Yes, there are few LiDAR systems that are there in the space operating from space based platforms. Like in earlier like from the year 2003 there was a satellite called ISAT with the LiDAR instrument called GLAS which was in orbit up to 2009 for measuring the topography of poles, ice sheets, glacier and all. So, that is why it is called like ISAT it is major application is to collect topographic information especially about the glaciers, ice sheets and also about like world's forest. This is one of like the very first space based or satellite based LiDAR system. Now, we have two LiDAR systems operational at the time of recording we have like two operational LiDAR systems in space one is ISAT 2 with an advanced laser altimeter system like called ATLAS advanced topographic laser altimeter system which is there which is developed version of like the glass which is like much more developed version of the glass system glass sensor that was there in the earlier ISAT 1 satellite. So, ISAT 2 again has major aim of measuring the topography of ice sheets glaciers and also like world's forest. See if you look here so the some of the basic characteristics is given here. So, the ISAT 2 satellite has an orbit of around like 500 kilometers like orbital altitude with a laser footprint size of like 17 meters and orbit return period of 91 days. So, Javel just discuss about this. So, the orbital inclination is 92 degrees it is in like a near polar orbit again it uses a green wavelength 532 nanometers because that is like the characteristic wavelength used for studying ice and all. So, it has a swath width of 6 kilometer along the ground track and the type of pulse turn it is called photon counting it is like slightly different from the discrete and full waveform they use what is known as like a photon counting technique. So, this is like one of the operational satellite based LiDAR system that is available. So, this will produce like this will have like a repeat interval of like 91 days. So, once on an average once every 3 months it will collect the topographic information over the same spot on the ground. But why do we have like such a very large number of repeat cycle? You just look at like the swath width actually it is like a very narrow swath what the ISAT 2 covers because it is like beams it is not like a scanner. ISAT 2 is not scanner system ISAT 2 produces like multiple beams and each beam will have its own footprint. So, within this each beam the elevation will be stored and actually the swath width is very small. Because of the very small swath width it has to go like large number of orbits in order to cover like a global coverage. Let us say if you have like one orbit like this say in terms of like lamps that you have a swath width of say 185 kilometers. So, within one go you are collecting information about 185 kilometers on either side of your orbit. But let us say now your swath width is just for example, entailing it is just 50 kilometers. So, what essentially happens rather than collecting one full 185 kilometers you are now collecting only 50 kilometers. So, now you have to make at least 3 orbits in order to have like a good coverage even for the same portion right. So, the number of orbits the satellite should make before it covers like the entire globe increases. So, the orbital in the discussions about like the platforms I told you some systems like almost all polar near polar satellites in sun synchronous orbits it will have like a repeativity cycle once every 16 days once every 24 days and all after making certain number of orbits say after making 233 orbits lamp that satellite will return to like the same position again towards the first orbit when it started. So, it is like orbital cycle. ISAT system before it makes completes one full cycle it has to it is it will undergo more than 1300 orbits because of its very narrow swath. The number of orbits it should make in order to cover like the entire globe is extreme large it should undergo 1300 plus orbits in order to cover like the entire globe and hence the revisit time is very large because of its very narrow swath width. So, the swath width and orbital characteristics will define the repeat cycle I told you and also like the repeat cycle swath width will further affect the temporal resolution. So, all these things are interconnected that is why ISAT 2 has like a very long gap between its repeat cycle almost 3 months once. There is another sensor called GEDI GDI which is Global Ecosystem Dynamics Investigation essentially a lidar to map or monitor vegetation in forest. This is not a satellite this sensor is there in the International Space Station which is now housing more number of like remote sensing sensors like GDI is one of the sensors there is another sensor called EcoStress. There are like multiple sensors remote sensing sensors are now being housed in the International Space Station. It is not a satellite we all know it is kind of like space based research observatory which is orbiting in like a lower earth orbit. Say it is not orbiting in terms of like around like 500 to 600 kilometers in elliptical orbit and it is not it will not produce like a global coverage. ISS it has an inclination of about like 51.6 degrees that means it will not cover like the entire large part of globe it will be mostly confined within this 50 degrees to 50 degrees kind of thing. So, that is one of the thing we need to remember. So, G-ray sensors essentially covers forest vegetation in tropics and temperate regions even not even in the like the very high latitude boreal regions. But this is one of like the demonstration system to prove that lidar based remote sensing is really helpful for monitoring the global biomass like the biomass of forest, carbon cycle and so on. Using this height information we will be able to model the biomass and also get information about like the carbon cycle happening within the forest ecosystem. All these things are really vital information in order to understand the earth's climate even like the earth system as a whole. So, the G-ray sensor uses an NAR wavelength swathed with of like 340 kilometers like it is not again it is not like a scanning system not like that it will have like again multiple beams in the across track direction and cover it. So, space based system they are not scanning based systems they have like in individual beams 4 beams 6 beams and all simultaneously it will be covering the ground. And this is like a full waveform lidar as I told like for vegetation monitoring a full wave waveform lidar can be useful and the laser footprint has a size of roughly 25 meters. So, this is the these are like the 2 sensors lidar based sensors that are there in orbit. So, after this G-ray one more sensor called MOLI multi footprint observational lidar and imager is planned to be launched again to be placed in international space station. So, this is kind of like a second demonstration tool like G-ray will be taken off and MOLI will be replaced that is like the information available in the literature and it is planned to be launched in the year like 2021 again it is like a full waveform lidar observing in NAR wave in 1064 nanometers with SWAT with a for like 1000 meters and so on again in ISS. So, these are some of like the some examples of lidar based or satellite or space based lidar systems which provides data ISAT 2 data and G-ray data like publicly available we can download them and use them. So, before we conclude we will just see like few applications of lidar briefly I am not going to explain any case studies and all just going to highlight few applications for which lidar systems can be used. The first major application for which lidar was used is like topographic surveying in order to obtain XYZ of the various features on the earth surface and that is like this major application still being used whenever we want like a high precise or highly accurate digital elevation model like three-dimensional representation of the terrain then lidar is kind of like one of the go-to technologies it will provide really quick accurate and precise digital elevation model and actually like scientists from different organizations led by like IIT Bombay created like the flood map or like the flood forecasting system for Chennai city in which like the DEM was collected from a lidar system because for flood modeling and all or flood forecasting applications we need like a highly accurate and dense DEM. Dense DEM means the spatial resolution has to be like really small it has to be in order of like less than a meter with very high vertical and horizontal accuracy in order to properly model the way in which the flood water may run. So, the lidar system was used to collect the DEM information for that particular project and also for various large scale civil engineering projects lidar system can be used like the high speed railway project between like Mumbai and Hamadabad a lidar system is used to carry out the survey along the path because the survey has to be completed in order of few months within the entire stretch and if we resort to like ground based survey it is going to take us years to cover that 500 600 kilometer stretch. So, people used lidar system and this one of like large scale application of lidar system for engineering purposes and also like as I told lidar based elevation information is also useful for monitoring and modeling global ice sheets glaciers and all like as demonstrated by ISAT satellite normally people will not launch a satellite if some data is not extremely helpful for a certain application launching a satellite is extremely costly. So, from that itself we can infer how important elevation information is for cryospheric applications and also for ecosystem based or especially forest ecosystem based applications we also use lidar system like for understanding forest canopy height estimating biomass and all lidar system is often used. And one more application is bathymetry like in order to map the topography of lakes or water bodies below the water how the topography is like the water surface maybe let us say that we have like a lake let the water surface be this. So, how the topography of this bottom surface lies lidar has been used to map certain in certain bathymetric studies. Most of the bathymetric lidar systems will use both the wavelengths 532 nanometers and 1064 nanometers. So, we all know that NAR wavelength has like extremely low penetration water bodies. So, whatever like the returns we get in NAR wavelength will most likely would have returned from the top of the water surface and green wavelength has certain penetration capacity into the water bodies. So, the last return may might have originated from some features present in the water or sometimes even like the bottom or the bed topography. So, the difference between them between the returns of green band and NAR band will give us information about the bathymetry of that particular water body. So, this is like another application for which lidar system is used. So, essentially lidar is one of like really advanced technology or I will say it is like combination of different technologies combined together will help us to get information about the earth and features on the earth with a much rapid turnaround time. So, with this we end this particular lecture and also this particular topic. So, essentially we have covered almost most of the principles that normally like a new coming students want to know about remote sensing like optical remote sensing like visible NAR wavelengths, how it will interact with earth surface features, thermal infrared remote sensing, their physical bases, passive microwave, active microwave all the basic principles we have discussed about at least to in some detail. So, from the next lecture onwards we will slowly move on to like application part or the analysis part I will say where we will talk we will just briefly get introduced to various data products then we will see like a few analysis and also like few applications in the last few lectures of this particular course. With this we end this lecture. Thank you very much.