 Hello everyone, welcome to the next lecture in the topic active microwave remote sensing where we are discussing about the imaging radar. The last lecture we discussed about the resolution concepts of imaging radar system especially the real aperture radar, what are the limitations with it and briefly got introduced to the concept of synthetic aperture radar. So, in this lecture we will discuss about the geometrical characteristics of the images acquired by synthetic aperture radar systems. So, geometrical characteristics of image basically means what sort of distortions may occur, how features may appear. So, we are going to discuss briefly about that. So, consider this in analogy with the distortions we discussed in the optical remote sensing when we discussed about the scanner elements like we discussed about visc broom scanners, push broom sensors and all, there I told for visc broom sensors like the JFOV will be keep on increasing and it will become coarser and coarser as the scanning angle increases which will cause image distortions all these things we discussed right. Similar concepts apply to radar remote sensing also, but here it will be occurring in a difference since we will see how it happens. So, I told you that radar measures distance in the slant range. So, the slant range distance can be converted to ground range distance easily if the surface is flat, but most of the surfaces are not flat they will have some sort of topography. So, presence of such topographic features say there may be like a small hill. So, presence of such features will change how land surface feature appears in the radar images. So, that is what we are going to discuss now. Here in this slide I have marked 3 concepts or 3 ways in which images or features will appear distorted in radar images. First thing is foreshortening, then it is layover, then it is shadow. What exactly foreshortening is let us say we have a mountain or hill like this. So, the radar system is somewhere here. So, this slope let us say this point is I will say this as like B double dash okay. So, the ground point here is A, the ground point here is B, the ground point here is C, the top of the hill is labeled as B double dash. Let us say this line the length AB double dash is the slope is same as B double dash C. Similarly, the horizontal distance is also the same that is horizontal distance AB is equal to BC. Similarly, even the sloping distance AB double dash is equal to B double dash C. So, it is like symmetrical mountain, symmetrical hill let us assume okay. So, this is also same, this is also same. Now, if the radar system is here, this slope is phasing the radar antenna, this like AB double dash is phasing the radar antenna and B double dash C is phasing away from the antenna. That length of these two things, these two slopes AB double dash, B double dash C are same. So, naturally we will think in the image they appear symmetrical that is if we look from the top let us say this is like one pixel in optical image. How this will appear? This will be one point say point A then comes point B say I will just draw it in a magnified manner. So, let us say this is point A, this is that point B double dash, this is that point C and the distance between these two even in the image should be equal then only the image is correctly representing the topographic feature even in image point AB double dash should be equal to B double dash C, image distance. In radar images, this will not be the case. Due to the slant range image acquisition what will happen is the radar signal that is coming in will make sure or will compress the four slope that is this is the image of the four slope. Four slope means the slope facing the antenna. So, the length of the slope, four slope will become shorter than the slope that is facing away from the antenna. So, the slope distances that normally we will see in radar images they will it will not be the same even though the hill is like symmetrically in our example the distance image distance A dash B dash will be much shorter than B dash C dash. So, this concept of shortening the length of the slope facing the antenna is what is called four shortening. So, this four shortening effect decrease with increasing look angle I will like expand further in the next slide. So, four shortening is in image in radar image the length of the topographic features having slope facing the radar antenna will appear shorter than the length of the slope facing away from the antenna. This is one thing. The next concept is layover. Here there is like a gentle slope like AB double dash. Let us say there is another hill which is relatively steeper. So, this maybe I will label this particular point as C double dash. So, B this angle is like a much steeper angle. If that is the case if this angle is steeper what will happen is when the radar signal is coming in for the radar pulse the slant range distance between the top point C will be shorter than this ground point B or C double dash. So, that is the slant range distance for point C double dash will be less than the slant range distance measured for point B. Then what the radar image will think or what the radar system will image the radar system will image as if point C is near to the system than point B. So, when the image is formed say if this is like the azimuth direction when the image is formed first this point C double dash will be map then point B will be map. So, that is rather than mapping image like this. Essentially, if you look at in the horizontal direction if you walk in this particular direction from ladder to the hill we will first encounter the ground point B then only we will be able to enter encounter this point C right. So, essentially the image also will should look something like this. If this is the direction in which flight is moving image point B should be nearer to the system C double dash should be nearer to the system and this is the reality in terms of horizontal direction. But for radar since the slant range distance of C double dash is smaller than slant range distance of B C double dash or the top of the hill will be image first then the bottom of the hill will be image later. So, in the image when you look it will appear as if the top is laying over the bottom. So, the image will look something like this the mountain will be sloping like this rather than looking like this. So, this is what we call it as layover mountain top overlaying on ground ahead of mountain like the base. So, it will be image like this. So, this is called layover effect the top will be appearing in front of the bottom then shadow. Shadow effect will occur if this back slope is actually much steeper. Let us say the back slope is here like the slope facing away from the antenna. If this is like steeper than this particular depression angle then what will happen is the slope facing away from the radar may not get illuminated by the radar pulse and hence whatever the portion that does not get eliminated by radar pulse will appear much darker in the radar image like it will appear black in the radar image. So, this is called radar shadow. So, the three types of relief displacement foreshortening layover and shadow foreshortening means if there is kind of like a topographic feature on the surface the slope that is facing the antenna the length of the slope will appear shorter than the length of the slope that is facing away from the antenna. This is called foreshortening shortening of the length of the foreslope. Foreslope here means the slope facing the antenna. Layover means layover is kind of like an extreme case of foreshortening that is the top will appear in front of the bottom of a topographic feature. Then shadow is for slopes facing away from the radar antenna in the direction opposite to it some parts of the slope may not be illuminated properly by the radar signal and they may produce like a black portion in the image without any radar input. So, some portions will appear completely dark this is called shadow. We will just look a little bit deeper into this. See again like the concept is explained in detail in these three figures. So, for foreshortening naturally will happen when the slopes are gentle like the foreslope is like kind of like a gentle slope not like a very steep slope when we compare this with this instant angle or I will call it as like a look angle. When we compare with this with look angle the slope is like gentle in such cases foreshortening will happen. This point AB will be imaged as will appear like a short steep slope and it will produce like a bright signal because let us say whatever features are present within the slope everything will be compressed within like a very short distance in the image. So, this will produce like a bright signal but for like a shorter length. So, this is called foreshortening. In case of layover the foreslope facing the antenna is steeper than when we compare with the look angle of the system. Under such circumstances the shot the top will be imaged first and then the bottom will be imaged giving an impression that top is laying over the bottom. This is layover. So, here comes shadow that is here the slope facing away from the antenna. The backward slope is much steeper when we compare this with the depression angle of the system. If that is the case whatever the portion that is lying between this depression angle point and the ground point will not receive any radar signals and hence will produce like a completely dark signal in the image. So, image this particular portion will appear completely dark. So, some characteristics of the foreshortening are like the foreshortening depends on object height first thing. Taller the object or taller the top of the object when compared with the bottom foreshortening effect will be severe and if it crosses a certain threshold then it will translate into layover that is layover is kind of like an extreme case of foreshortening. So, if the top height is much if the top is much taller than the bottom of the topographic features say there is like a huge tower standing if that if this angle is much steeper. So, the angle or the slope of the line covering the top and the bottom will be much steeper. Under such circumstances foreshortening will become very extreme and gets converted to layover. Then look angle with respect to look angle the foreshortening and the layover effects will be severe in the near range than in the far range. Actually this was like look described in the previous slide. So, the foreshortening effects decrease with increasing look angle that is in near range when the look angle is shorter the foreshortening effects will be much stronger that is slopes that are slopes will be foreshortened to a greater extent in the near range than slopes that are in the far range. So, the look angle and the location of the objects will affect foreshortening and layover. So, both foreshortening and layover depends on these factors. Then some characteristics of radar shadows. For radar shadows they will be completely dark. So, radar shadows will not have any kind of signal it will be pitch black because this is because like just think it in analogy with our optical remote sensing. When we discussed optical remote sensing we saw that even when direct sunlight is not falling on an object the diffuse skylight can irradiate like there is a building there is like some other object here say the sun is shining from this direction. So, this building will obscure the direct sunlight falling on objects here. But still the diffuse skylight present in the atmosphere will illuminate whatever features present there right. So, we will get some image of objects that are present within the shadow in optical remote sensing that is not the case in radar remote sensing. Radar remote sensing we are not receiving or measuring any outside signals whatever we send only we will receive it back. So, if the objects are not eliminated by the microwave pulses that were transmitted by the radar antenna they will not reflect anything back and hence whatever portions that do not receive any radar elimination they will appear completely dark no other external source of energy is present it will be the radar shadows will be pitch dark. And also the radar shadows appear only in the range direction that is say if the flight is flying like this shadows will be cast only in this direction in the range direction. This is because in the azimuth direction there is no concept of shadow if there is like a topographic feature standing here when the flight is flying like this then the shadows will fall only in range direction. So, by looking in the direction in which the shadows were aligned we will be able to get some information about the flight direction that is possible. And also the shadows the shadow effect will become very severe in the far range. That is if there is like a slope in the near range there are chances that the back slope will get illuminated shadows may not be there but in far range shadow effect will be very severe. This is like opposite to this foreshortening effect foreshortening effect will be very severe in the near range will decrease in the far range shadow is exactly opposite shadows will be occurring to a less extent in the near range it will be occurring to a larger extent in the far range. So, these are some characteristics of radar images. So, this particular slide tells us like an interesting picture of like an example like all these hills A, B, C, D are like almost identical and they are aligned positioned in different positions in the range direction A is very near to the antenna and B, C, D are like farther away from the antenna. So, just look at this what is happening here at point A there is like a layover effect that is extreme case of foreshortening. So, here the effect is much larger A will be imaged first that is the top will be imaged first then the bottom will be imaged. So, this will be like the top point this will be like the bottom point. So, the layover is much and the image will look something like this like it will appear as if the top is much farther away from the bottom. When you slightly move away from the move away in the range direction when we come to this hill B this effect will become little this effect will become slightly reduced say the top may appear something like this rather than appearing very steep it may appear like slightly as if both of them are in both of them are in almost like straight line. So, this is for hill A this is like hill B. So, slightly reduced effect of layover at point C the layover effect is further reduced. So, the top and bottom will appear as if they are at one single exact point that is the slanted in distance of top and bottom are one and the same that means both the points will appear exactly same thing it will appear as if there is no top and no bottom it is like a single dot it will appear as something like this then in the far range some foreshortening because foreshortening is like layover is translating into foreshortening. So, your extreme effect is translating into something if layover has not occurred here if the slope is much shallow here itself then there will be layover here and the layover effect may vanish in the far range. So, that is the concept. So, the layover effect which is like in very extreme case of foreshortening it is like slowly decreasing as we move away in the range direction whereas look at the shadow this backward slope here there is some sort of return for the radar signal there is some weak return whereas from this mountain B itself the shadows will be like dominating and the length of the shadow is keep on increasing as we move away in the range direction. So, the layover or foreshortening effects they are like they will move together. So, they will the effect will decrease in the range direction for layover and foreshortening whereas shadow effect will increase the length of the shadow will increase as we move away from the the nadir point as we move away in the range direction. So, these are essentially the characteristics of geometrical characteristics of radar image. So, we have to be really careful and this effects the geometric characteristics will be clearly visible in image acquired by aircraft systems because normally in aircraft systems due to like the shallower flying height what will happen is the look angle will vary a lot the aircraft has to have like a what in order to cover like a large area there will be like a large variation in the look angle leading to a drastic change in the surface images in the surface with there will be like a drastic change in the image acquired. So, these effects layover or foreshortening will be clearly seen. When we move to satellite systems the look angle will not be varying to a large extent because of the very large flying heights involved due to which the effects may be slightly reduced like visual appearance the image may look okay to our eyes. So, whenever we work with like aircraft based radar images we have to really take this into account. So, that is why like in radar image processing there are like certain tools to process SAR images even open source tools are available for example, SNAP by this ESA and Copernicus team. So, those tools will have like an inbuilt terrain correction feature like there will be we will be using like a digital elevation model in order to correct like the effects of this terrain due to this layover and foreshortening effects in order to get like a more true representation because like a top should not come before the bottom these are like really like a large scale distortion right. So, those distortions has to be removed. So, terrain correction has to be performed in radar images whenever the images are acquired over land surfaces having very large variation in surface topography. The next important feature of SAR images are radar spickle. So, what exactly radar spickle? So, when I briefly discussed about the SAR systems I said like it works based on measuring the phase relationship right between the waves that got transmitted and between the waves that got received. So, the phase information is also getting stored and the phase information is being used to process the data to produce one image and in order to improve the azimuth resolution. When phase information are used to add several waves let us say like there is like one single element this is like one pixel in the radar image. There can be many different features present right because even though if you assume this is like 10 meter by 10 meter, but still within this 10 meter by 10 meter there can be different different features present within this pixel. So, when radar interacts with these objects based on the object characteristics the reflected signals or the we call it as backscattered signals the backscattered signals will have differing phases like phase relationship can be the same or it can be different based on the features present. All these things will be processed together by the radar system to produce this one particular pixel. So, when all these are combined together along with the phase information what can happen is the phase of each of this returned signal may add up constructively or they may add up destructively. Like if they still like microwave pulses that got transmitted are coherent. So, if even after reflection if this coherency is maintained then what will happen is the radar system will or the reflected backscattered radar signal will interfere with each other and this interference will either can add up constructively or destructively. If the different reflected microwave signals if they add up then essentially they will produce like a very bright patch in the image for that particular pixel like schematically I have represented if there are like 2 waves and if there are like sorry if there are like 2 waves and if they are like in phase with each other then these 2 will add up and the net resultant wave will have like kind of a increased amplitude. So, this will appear like a brighter image. So, this is like constructive interference or adding up of waves whereas if 2 waves are out of phase with each other they will cancel out and they will produce like a very dark patch. So, that is like these patches basically which are like really dark. So, the waves within like each resolution element of radar pixel can add up during this data processing radar data processing and they can add up in either constructive manner or in destructive manner leading to bright and dark spots in radar image. We call this as radar speckle. Especially like certain images especially acquired over like urban areas and all may have like speckle effect to a large extent. So, on a colloquial note we call this as a salt and pepper pattern bright dark bright dark randomly appearing bright and dark patches. Image with this kind of characteristics will not be helpful for us like they may not be pleasant to look we will be having like random patterns or random patches of bright and dark patches which may not look good. So, the speckle has to be kind of reduced. So, this is like a characteristic of SAR image not like real aperture radar. It is a characteristic of synthetic aperture radar because in synthetic aperture radar the phase information is also recorded and used in data processing thereby enabling the addition of 2 different waves or super imposition of 2 different waves. So, this is due to radar data processing. So, this speckle effect has to be reduced. So, there are like plenty of ways in which this radar speckle can be reduced we are not going to discuss everything in detail. But one of the simple ways that people use in radar processing is multi look that is even when the radar system is acquiring images like what I told you in for synthetic aperture radar processing is the same object will be appearing in the radar system for a long distance much longer than the physical length of the antenna. So, all the distance for which the object is visible will be combined together in order to synthesize a very long antenna length that is what I said conceptually. So, let us say one particular object is visible for like several kilometers of distance for several kilometers distance they will be broken up all the things will not be combined together. So, they will be broken up into small small patches say 3 or 4 different let us say it is broken into 4 patches each patch will be processed separately and then they will be added together we call this multi look. So, each broken up let us say this is like the entire sorry this is like the entire length for which an object here is being image by the radar system instead of taking this entire length L let us say this is broken to small small chains. This is 1, 2, 3, 4 for this 4 it is broken to small small chains and these are these will be processed separately. So, what exactly is happening is conceptually speaking the phase information is just added for a small amount of distance covered rather than adding up all the phase for this entire length. So, the each of this thing we call it as one look ok. So, then all these looks will be added together to produce what is known as a multi look image. So, essentially a multi look means rather than synthesizing a very long antenna the antenna length will be broken during data processing itself in order to reduce this spical effect. So, here essentially we are sacrificing the azimuth resolution because as the antenna length decreases azimuth resolution will become coarser and coarser. So, we are sacrificing the azimuth resolution in order to reduce the spical effect. So, here this is like a one look image, but here it is like a four look image ok. So, you can see like the spical effect is reducing. Apart from this multi look there are different ways different kind of filters available in image processing chain in order to reduce this spical effect. Anyway in this course we will not see everything we will just get introduced to what is radar spacal. So, radar spacal is like salt and pepper pattern that is appearing in synthetic aperture radar images which will produce like a random pattern of bright and dull spots. And in order to reduce this radar image processing system itself will do what is known as multi look processing for reducing this effect. Apart from this we can use some sort of image processing filters to reduce this effect further. So, in this lecture as a summary we have discussed about the image characteristics or relief displacement features of radar images. And also we have discussed about the radar spacal which is a characteristics of image acquired by synthetic aperture radar system. With this we end this lecture. Thank you very much.