 Hello everyone. Welcome to today's lecture in the course remote sensing principles and applications. We started discussing about different remote sensing systems and the ways of image acquisition and in this lecture 2 we are going to continue with that particular topic. In the last class we discussed briefly about the different types of remote sensing sensors which collects spatial information, spectral information or radio and metric information. I also told you some of the most commonly used satellites data like from which we are acquiring data such as Landsat or Modus etc. They collect all sort of information, they collect spatially, they collect spectral information and also they collect radiometric information. We also started to discuss about how like photography was used in the earlier days like the people will send the normal film cameras like specifically designed for such this sort of like earth observations. They will put such camera in aero planes or balloons whatever it started with balloons slowly progress to aircraft with the invent of aero planes. Then people will fly this, take photographs of the space and will use it for interpretation purposes that is identifying what is there on earth surface or also for mapping it that is a different field called photogrammetry that is what I introduce to you in the last class. This class we will see how images are being acquired or how those systems work when the satellites are being launched, when satellites launched into space they have to collect images 2D images how it is being done. I also told you like satellites are commonly launched into two kinds of orbits one is called near polar orbit, one is called geostationary orbit. First we will discuss about the image collection from satellites that are located in the near polar orbits we will first see that. So, what a near polar orbit is say this is earth in space. So, earth rotates from west to east we know and the satellites will be rotating or revolving around the earth from north to south around the poles. So, most likely they will not exactly pass over the poles, but they will be having like a slight inclination away from the polar axis that is why we call this as near polar orbits. So, when satellites are placed in such near polar orbits earth is moving in a direction that is almost perpendicular to the direction of satellite motion, satellite is moving like this earth will be moving like this. So, north, south, east, west this is how the orientation is. So, the images will be collected like in one dimension like in the direction in which the satellite is moving the image will be collected due to satellite motion and in the direction perpendicular to it the image will be collected by the motion of the sensor or the scanner attached with the satellite. First we will see in detail how, but before we move on to the two terms I would like to introduce the direction in which the satellite is moving we commonly refer it as along track direction and the direction perpendicular to satellite motion we call it as across track direction. So, along track is along the direction of satellite motion across track is in a direction perpendicular to it. So, keeping this in mind we move on to the figure labeled as B in this particular diagram. So, what it shows it shows that the sensor or the satellite is moving in this particular direction let us assume this is like along track. So, in this particular direction image will be collected by satellite motion. So, first of all in earlier days when satellites was launched like digital camera technology was not as advanced as we are having it today like in the 70s and all early 70s period. So, what they did is they developed small detectors send the detector to space and that detector will be attached with what is known as a scanning element. So, the detector which is essentially like you can think it of like a simplest version of the camera we are using. So, it will be attached with the scanning element. So, what it will do the satellite will be orbiting in space let us assume the satellite is now moving towards the like in this particular direction like coming in towards your screen. So, when this is moving there will be like a detector attached with the net and there will be like a element known as a scanner. So, what the scanner will do the scanner will be oscillating like this in a direction perpendicular to satellite motion satellite is moving like this and the scanner will be oscillating like this. So, what essentially it will do is a detector is like only a one chip or one detector element will be there in this will be there in the sensor or in the satellite satellite will be moving like this. The scanner through its motion in a direction perpendicular to it will collect all the energy in the across track direction that is this is land surface this is how satellite is moving. So, the scanner it will be like it will be like standing from this side to the exact nadir point and move to this side. So, what it will happen whatever the energy coming in from the ground it will be collected by the scanner and that will be passed on to the detector for recording purposes. So, the detector will convert that energy to Dn and store it in form of like a image. So, that is how it was building. So, the scanner essentially provides the data collection in the across track direction and as the satellite is moving in along track two dimensional image will be formed that is just imagine starting from here energy is coming like this. So, it will be keep on collecting energy coming from the earth surface. So, it will cover certain distance in the across track. So, one line of image has been formed. So, if you take a photograph a photograph is a 2D space each row it is built up of several rows. Okay. So, this now it has collected one full row of data one line of image then the satellite will move to the next line. Again the scanner will collect the data for that particular line that is how it was collecting the data in the earlier stages. So, we call that type of scanner as line scanner. So, a line scanner essentially has one detector element per band. So, let us say the satellite is collecting data in four bands blue, green, red and NIR for an example I am telling. So, it will have four detector elements for collecting data in each band. So, just four bands, four detector chips. So, now the scanner is actually collecting the data passing it to detectors inside and each detector will store the energy coming in that particular bandwidth. So, essentially this type of scanner will build the image line by line one line in the across track direction then the satellite will move to the next line then the second line will be imaged then the satellite will move to third line then the third line will be imaged. So, it is like line by line as the image is formed such scanners they are known as line scanners they collected image line by line. We will also see that this type of operation having only one detector per band will be disadvantages, disadvantages in a sense the time available for the data collection by this scanner is extremely low in the order of micro seconds. We will see how it is in the later part of slides, but since the time is very low what people thought is instead of putting one detector per band can we add more detectors in each band and send it to space. That example is shown in this particular slide in this figure what is known as a visc broom scanner. So, what exactly is a visc broom scanner? Let us assume the data is being collected in say again like four different bands blue, green, red and NIR, but instead of having just one detector element per band now we are having four detector elements per band. So, this four is for blue, the next four is for green again it is for red like that. So, four or whatever it will be like equal number of detector elements will be spaced four or three whatever they will space more than one detector per band for collecting the data. So, what will happen when data is being collected? So, now instead of the scanner having to collect each line completely like the satellite is moving the scanner is moving in the across track direction. So, the scanner has to complete that one full line before the satellite moves to the next line only that amount of time it has the scanner has to rotate extremely fast. But if it has three, four detectors behind it rather than just having one detector element now it has say four detectors per band aligned in the along track direction. So, now the scanner has somewhat more time like four times now because of presence of four detectors it has four x time. Like if it is was like say 5 microseconds before now it is 4 into 5 microseconds 20 microseconds time it has. So, schematically I will explain it in this slide. Let us assume this is the along track direction. So, there will now be instead of having one detector element let us say we have four detector element per band and the scanner will be rotating in this axis this is across track. So, what will happen first this scanner will be looking at one particular place like collecting data over it like collecting data and whatever data is coming from this particular point let us say will be collected by detector number 1. Then after it passes now detector number 2 will be overhead of this. So, detector number 2 will be overhead of this and this will collect the data from the same point then detector number 3 will be there it will collect data from the same point then detector number 4. So, essentially what happens is each ground point is now being imaged by four detectors one after the other. So, that is why such sensors as the number of detectors in the along track direction grows that time available for imaging certain pixel or certain area on the ground is increasing by the number of detectors times. See if you have four detectors so, then now the time is increased by four times. If you have six detectors the time is now increased by six times in comparison to when you have only one detector element per band. So, here you note earlier case in the line scanner we had only one detector element per band. In case of this type of scanner which is known as visc broom scanner we have say 4, 6, 10 whatever depends on sensor configuration people will have different number of bands sorry different number of detector elements. Different number of detector elements will be aligned along the along track direction in the along track say if the satellite is moving like this the detectors will be oriented along the direction of satellite motion and they will scan together. So, they will be attached to scanner now the scanner will be collecting data together for all the detectors. So, since as the detectors or the scanners is moving so, each ground point each area on the ground will now be imaged by say n number of detectors 4, 6, 10 etc that increases the time available for data collection. So, such scanners which has more than one detector element per band in the along track direction and which scans the ground surface using a scanner or known as visc broom scanners that is given here in this particular slide visc broom. Again we will note later part of the slides that even for visc broom scanners the amount of time collected is in the order of like few microseconds only, but still it is higher than the line scanner, but still it is in the order of microseconds. In order to overcome this people thought instead of having detectors in the along track direction and putting it in in form of like a scanner cannot we change this orientation because what will happen is first of all the time of data collection is very small in the order of microseconds and also as the scanner is moving there will be lot of geometric distortions. Geometric distortions means the features on the earth surface will not be imaged properly there will be some distortions we will see again later in this particular lecture topic. But there will be lot of geometric distortions because of this scanning and this scanning has lot of limitations attached to it. So, what people thought is can we remove this scanner part because scanner is like a mechanical mover there is some element which is moving mechanically. If the scanner system fails let us assume the worst case scenario if the scanner fails then the detectors are almost like useless they cannot scroll without the help of the scanner they will be just moving along with the satellite. So, they will be just collecting a small amount of data along the direction in which the satellite is moving that is all they would not be able to collect image in the across track direction. So, they will be just limited to like one line over which the satellite is moving. So, without the scanner the satellite the detector will not be able to collect data some mechanical movement is there which affects the satellite system basically. So, people thought can we remove the scanning part we do not want the scanning part because scanner always having mechanical movement in say the satellite is not recommended it may cause some problem in the latest stage. So, people thought instead of putting say 4, 6 or 10 detectors in the along track direction people decided to put many number of detectors in the order of like say 1000 also in the across track direction that is now you have a large number of detectors per band in order of say hundreds or even thousands in the across track direction and the satellite will be moving like this. So, each detector will be collecting energy from certain area on the ground continuously. So, there will be like thousands of detectors like this in the across track direction they will be moving like this collecting data from the ground simultaneously at the same time you just compare this with scanner. Scanner is if the scanner is here it will collect data only coming in from this point if the scanner is here it will collect data only from coming from this point if the scanner is here same thing. But if you have say like 1000 detectors in the across track no scanner involved all the 1000 detectors are seeing the ground at the same time means all the detectors will be collecting energy simultaneously from different different points on the ground. So, as the satellite is moving like this all detectors will be keep on collecting data and building a two-dimensional image. So, such systems where there is no scanning element but you have hundreds or thousands of detectors in the across track direction and while it is moving it is collecting data in two dimensions such scanners are called push broom scanners. So, push broom scanners yeah it is given here push broom. So, it will have like say n number of detectors in the across track direction they will be collecting image from all the pixels in this line together simultaneously collecting data in this form in push broom form has lot of advantages. First thing the time available for observing each and every ground area is increased many folds like for many push broom scanners the time for data collection will be in the order of milliseconds. Earlier for the line scanner or visc broom scanner the time available was in the order of microseconds whereas, now the time will be in the order of milliseconds we will see later and also the geometric distortions will be reduced significantly for a push broom scanner. So, push broom scanner has several advantages. So, three different types of scanners we have seen one is line scanner putting only one detector element attach a scanner to it which will oscillate like this collecting data. Then we saw visc broom where you have more than one detector per band in the along track direction which together will scan like this which will provide say if you have 4 detectors will provide 4 times more observation of every point on the ground. On the other hand push broom scanners push broom sensors they are not scanners push broom sensors will have large number of detectors aligned in across track direction and they will be collecting image from entire line simultaneously and they will be moving like this ok. So, such sensors are called push broom sensors. Then the digital era came like many advancements happened in digital image acquisition. So, people thought of like putting even if you put like a push broom sensor still it collects one line per per image it collects line by line right even if in case of like push broom but still the amount of time available to collect data is increased manifold. But with the advent of digital systems and sophisticated digital sensors and people moved on to developing a two-dimensional array of detector elements like what we now have in our normal cameras. So, if you open our normal cameras we can see like a two-dimensional array of detector elements. So, essentially what are we doing we are having like something in front of our eyes we are focusing our camera towards it and it will have a two-dimensional matrix let us say 100 by 100 detector elements are there 100 rows by 100 columns. So, it is essentially 10,000 detectors arranged in form of like a square matrix. So, those 10,000 detectors will simultaneously observe 10,000 ground points that is on the object space and it will form a 2D image. So, such systems when sent to space and used to collect information we call such detectors as digital frame arrays. So, that is explained in this particular slide digital frame camera array areas labeled here. So, that will have a two-dimensional detector element. So, here in this example we have a 5 by 5 matrix of detector elements per band. So, blue, green, red and air for 4 band you have 4 2D arrays. So, essentially what will happen? So, it is like almost like similar to taking photography the simple analogy. So, here is the lens. So, each of these 5 by 5 detector element will observe 25 ground points simultaneously together it will collect data and store it in each band. This further increases the time of observation. So, now I am going to introduce you to one particular technical term known as dwell time. I am repeatedly telling time of observation 4 over like each pixel or something like what is the time available for data collection over each ground area and technically that time is known as dwell time or time of integration. The time available for observing one particular ground area we call dwell time or time of integration. So, if you look at these four type of sensors line scanner, visc broom scanner, push broom sensor and a sensor with two-dimensional array detector element. So, the dwell time will be increasing in the order, dwell time will be leased for line scanner, dwell time will be little bit higher for visc broom scanner, much higher for push broom sensor and still higher for a two-dimensional array type of detector sensor. So, but a sensor with a two-dimensional array will have some sort of memory limitations like now we will experiment with the availability of large format digital cameras. Now we will be experiencing each image will be like several MBs in size. If you take even like a short video with high resolution cameras it will be like going to gigabytes. So, it is very easy to fill the memory of our memory card that is we are using now. So, this will be the constraint with respect to using a two-dimensional array camera. It will collect lot of information together, it will fill the memory of the system and also there will be time to transmit the data right because satellite is not in the ground, it is in space and the data has to be transmitted from the satellite to the ground remotely without any wired communication. So, that will take lot of time and that will be like a major disturbance. So, mostly the two-dimensional array of detectors are essentially limited to aerial remote sensing not to satellites still most of the satellites are using visc broom or push broom technology, most of the satellites are now using push broom sensor type. But having like a two-dimensional array matrix for each band is essentially limited to remote sensing from aircrafts. Because of this memory limitation quickly the memory will fill up there will be like a lot of time taken for transmitting the data from space to ground which will be problem essentially for satellites that are farther in space. And now till now we were talking about just four bands blue, green, red and yellow as an example. There is a class of sensor known as hyperspectral sensors, hyperspectral sensors means they will collect data over large number of bands say 100, 200 bands at the same time. Like instead of collecting data from 0.4 to 0.5 in the entire blue band it will collect data in 0.4 to 0.41, 0.41 to 0.42 and so on large number of continuous bands with much smaller bandwidths. Such sensors are called hyperspectral sensors. So, which actually improved the way in which we collected images and identified objects on the ground's base. So, such sensors will have large number of bands data to be data will be collected in large number of bands every like simultaneously. For such sensors how the data is being collected one such example is again given in this slide. So, for hyperspectral sensors I told you there will be like say 100 to 200 bands. So, they will use a push-pwn type of sensor. So, that is let us say we have 10 detector elements in the across track direction. And let us assume we have 100 bands in the hyperspectral sensors. So, what it will have? It will have like a detector array it is also a 2D array, but each row will correspond to this 10 detector elements in the across track direction like each element in one row whereas different different rows correspond to different different bands say this is band 1, this is band 2 and so on that is one row will collect data in one particular band. So, if there are like 10 detector elements in the across track direction 10 different ground points will be image simultaneously and the detector will have 10 detectors per band. So, it is 10 detector per band and if there are 100 bands it is 10 to 100,000 detector elements. So, this is like an example for data collection for hyperspectral sensors. So, essentially the data can be collected in several ways either having like a scanning mechanism or having like a multiple detector elements. So, we will see few examples of such scanner sensors, scanning elements or hyperspectral array elements and so on. So, if you look at this particular slide, here we are showing the sensor or detector arrangement for a sensor known as Landsat 7 ETM plus Landsat 7 is the name of satellite. This is the name of the sensor enhanced thematic mapper plus it was a vis groom scanner. So, vis groom scanner means it has more than one detector element per band. He collected data in 8 different bands. So, here you have 1, 2, 1, 2, 3, 4, 5, 6, 7, 8, 8 bands it collected data and the spatial resolution is here from bands 1 to 7 the data was collected at 30 meter resolution that is each ground area will roughly be 30 meter by 30 meter and then band 6 this is except band 6. Band 6 data was collected at 60 meter resolution. So, this is a thermal infrared band and band 8 was the panchromatic band which collected data at 15 meter resolution. So, you can see for each this will be like aligned in the along track sorry this will be the along track. So, in the along track you have 1, 2, 3, 4, 5, 6, 7, 8, 16 detector elements arranged in the along track direction. Similarly, you have for bands 1, 2, 3, 4, 5 and 7 you have 16 detector elements. In band 6 you have 8 detector elements in band 8 you will have 32 detector elements because of this variation in spatial resolution. So, this looks like this extremely tiny each detector size. They are like extremely tiny when they when they are arranged in on the space you can see the scale here this is 1 inch roughly 2.54 centimeters all the detector elements are arranged within this 2.54 centimeters. So, this will be like say here if in example in band 1 there are like 16 detectors. So, 16 detectors will be put in the along track direction there will be like a scanner attached to it it will scan like this. So, each ground point will be observed by each one of the 16 detector elements thereby increasing the dwell time. This slide shows another example of a sensor known as MODIS which is again like famously used sensor. The way MODIS works is very similar to viscgroom methodology but there is like a slight difference we call this MODIS sensor as a paddle broom. Paddle broom sensor maybe like it will have like a I earlier told you there will be like a scanner element scanning like this right whereas in MODIS what they have done is they have attached the scanner with mirrors on both the sides like there will be like a very highly polished mirror here on the bottom similar mirror on the top. So, this will be continuously rotating instead of scanning like this it will be rotating like this. So, what essentially happens as the mirror rotates like this it will collect data in the line. Now, the bottom mirror will go to top top mirror will come to bottom it will collect the next line now it will again rotate. So, the data collection happened in one direction only like this next the mirror will turn upside down again the next line like this. So, it is not like oscillating scanner but a rotating scanning element fitted with mirror on both the sides we call it as paddle broom. So, MODIS is attached with such paddle broom essentially it is still a scanner it is doing scanning only but scanning happens in only one direction with director attached in both the sides mirror attached in both the sides. So, MODIS also has many number of detectors per band. So, essentially like we can see band 1 and 2 with 250 meter resolution has lot of bands then band 3, 4 collects data at 500 meter resolution other bands collect data at 1000 meter resolution and so on. So, these things so, this will be like the entire director RHIP it collects data in 36 different bands. So, this is along track this is across track. So, each director element per band is aligned in the along track direction and each band is oriented in the across track direction for MODIS. Now, this is this slide shows an example of an hyperspectral sensor where it will be like in the along track direction there are 210 director elements along track is like this. So, this is 210 director elements in the across track we have 320 director element. So, that is each band has 320 directors, 320 directors per band essentially this is a push broom type of sensor and for each band there will be one probe detector this is band 1 this is band 2 each band will have 320 detectors and like this there are 220 bands. So, essentially 320 into 210 number of detectors this is like a push broom type hyperspectral sensor because it collected data in 210 bands simultaneously. So, just as a summary of what we have done in this lecture in this lecture we have seen how data is being collected from satellites in the near polar orbit that is earth will be moving in this to east satellite will be moving north through south. For satellites in such orbits there will be like satellite motion will be in the along track direction and data will be collected in two dimensions either by using some scanning elements or by using multiple detectors per band aligned in along track or across track direction. With this we conclude this lecture. Thank you very much.