Lab 8: Spectral Signature Analysis

Goal:
 The main goal of this lab was for us to gain experience with measuring and interpreting spectral reflectance or signatures of different features that are found on the Earth's surface and near different surface materials that have been captured by satellite images. The outcome of this lab is for us to be able to collect multiple spectral signatures from remotely sensed images, look at the bands of each feature analyze them and graph them.To practice these skills we used Landsat ETM+ images of Wisconsin and we were to digitize and record the different spectral signatures of the various features below.

To find all of these spectral signatures, all we had to do was upload the image then go to the draw tab on the application and click on polygon. We needed to draw a polygon around any land or feature that was on that list or that one was interested in using the signature editing tool on. The signature editing tool is located under the raster tab, after collecting all of the data that was on the list this is what my graph ended up looking like when comparing all of the features that I had collected.

After all of the data was collected and then projected on the graph above, one can tell which features are absorbing the most energy and that plants generally reflect the most green light since they are absorbing the longer wavelengths. This lab was one of the more helpful labs, I will be able to use this skill in my future with remote sensing and in our final project, because this skill helps us track the function of different systems that are found in the environment.

Photogrammetry

Lab 7: Photogrammetry

Goal:

                The main goal of this lab was to help us develop our skills in performing key photogrammetric tasks on aerial photographs and satellite images. But to get more in depth on the goal we are more specifically training our overall understanding in the mathematics behind certain calculations such as photographic scales, measurement of areas and perimeters of features, and calculating the relief displacement of an image. To sum up what the overall goal that we are trying to learn are different ways to look at images on Erdas and this lab is going to help us perform Orthorectification on different satellite images as well as introduce us to stereoscopy.

Methods:

                While going through the lab, I had to use many different tools that we have never used before on Erdas Imagine before and we had to perform multiple tests on different images in order to complete different tasks in the lab. The first step of the lab had to do with scales, measurements and relief displacement. The first question was asking us to create scales for different images based off of different numbers that were given to us. We had to calculate what the scale was for an aerial photograph while showing all of our work. We had to go through this procedure a couple of different times for different images and different areas on an image. After we found out all of the measurements we had to calculate the relief displacement of a certain object on a map or in this case the smoke stack of a building. In order to find the relief displacement we need to use this equation; d = (h*r)/H, this allows us to see which objects have a higher elevation than the local datum.

                The next step started going into less measurements and scales, and more stereoscopy. This procedure involved a 3D perspective view of the City of Eau Claire. We used this data that we collected about the elevation of the city of Eau Claire to determine which features have the highest elevation and which had the lowest. In order to figure out that we need to use a new tool called the Anaglyph Generation of two different images. This would give us an image that would show the different areas in the City of Eau Claire that have a higher elevation of different areas. We needed to use Polaroid glasses to observe the image and the surrounding areas of the image this would help us describe the elevation of the different features in Eau Claire.

                The final step was Orthorectification, this part of the lab will introduce us to Erdas Imagine Lecia Photogrammetric Suite or LPS which is mainly used in digital photogrammetry for many different reasons, some might include triangulation, Orthorectification of images collected, extraction of digital surface and elevation models etc. In order to do this we needed to use the Imagine Photogrammetry tool in the toolbox, from here we created a new block file so than we can begin to set up the model. What we are working on currently is going to be a horizontal reference coordinate system instead of a vertical we will be working on the vertical reference coordinate system later in the lab. Then we had to start by adding the Spot_pan.img to the program, after that we needed to make sure all of the properties for the program were correct and set to the right setting. The next step was to start activate point measurements and collecting GCPs by comparing to images that were showing the same area. Then we would just keep adding on points with different reference points that would show the same image on both maps. We had to do this procedure nine times in a row to get all of the reference points that were needed, than we had to reset the horizontal reference source and edit an entirely different images than before while keeping the one of the original images. Then we would go on to add some more reference points to both images points eleven and twelve.

                The next step was to change it to a Vertical Reference source, while using a DEM and update all of the Z values on the selected points that we have made so far. We then changed the “Type” and “Usage” columns from “none” and “tie”, to “full” and “control”. After that we closed the point measurement tool and were returned to the LPS project manager. This is where we started adding more reference points to these two images comparing the two through the use of these points. We had to add twelve more reference points to the two images, this will lead us to starting to triangulation and ortho resampling. For the triangulation we needed to use same weighted values for all of the points. The next step was to create a Triangulation Summary Report to see if we would be able to actually triangulate the data that we are working on. Once we finished with the final steps for the triangulation we started the ortho resampling process, when starting the final process we were using a Bilinear Interpolation as the resampling method, our end goal was to try and get all of the reference points from both instances onto the same block figure. The final image will be of two images that are displaying the same image but at different locations, this meaning once you put the two maps over one another it will have a special overlap at the boundaries of the two orthorectified images.

Lab 6: Geometric Correction

Goal

The name of this lab is "Geometric Correction" and our main goal was to fix the distortion between two maps that are representing the same image and creating a new map with less to no distortion. During this lab we learned a few new skills while doing it, for instance we learned two major types of geometric correction, one was the pre-processing of activities prior to the extraction of the biophysical and sociocultural information from satellite images. In order to complete these tasks and to learn these skill sets, we needed to use Erdas Imagine to help us complete each of the tasks that were represented to us in the lab. 

Methodology/Results 

The first step of the lab was to upload two maps that represented the same location and try to fix the distortion for one of them so than both maps would match exactly. In order to do this we had to upload both maps into Erdas, and than find the control points that were made on the map originally, the points that were causing the distortion in the first place. Once we deleted all of the control points it was our job to make our own and try to get the points to match correctly in each map so than the points would be in the exact same location. Once we created three control points and attempt to match them up every point after that would automatically match up because the polynomial order is at 1 meaning that there only needs to be three points for the points to start matching up by themselves. After we got our four points on their our goal was to get the Root Mean Square (RMS) error down to 2.0 or lower, the lower the RMS was the less distortion we would have in the image. Below is an image of the maps known as figure 1 with the four control points on it and with the RMS lower than 2.0 or is this case at 1.5757.

Figure 1

The next step for the lab was the same objective as the first, compare two images that are representing the same area and try to find and get rid of the distortion by using Erdas Imagine. This time its a little more difficult though because we were asked to change the polynomial order from 1 to 3 which means that we need to use more control points before they automatically get added to the corresponding map. In this case we had to add ten different control points all over the map and try our best to match the points from one map to the other before they would automatically match. After we added the ten control points we were asked to add two more and then try to get the RMS error below 1.0. In order to do this we needed to change the location of almost every single control point ever so slightly so than the RMS would read below 1.0. Below is an image known as figure 2 and this figure represents the two new maps being compared, with 12 control points and with an RMS of 0.7137. 

Figure 2

This lab was very helpful is finding out different ways on Erdas to remove distortion from images or at the very least get rid of some of the distortion. I thought of this lab as a great learning experience, and plan to use all of the skills picked up this lab in the future.

Lab 5 Remote Sensing

Goals

The goal of this lab was to teach the students how to properly build a data structure by using Lidar and also properly process data in Lidar. There were a few specific objectives that we were supposed to keep in mind while going through this lab and these included; processing and retrieval of various surface and terrain models, and processing and creation of intensity image and other derivative products from point cloud. 

Methodology/Results

In order to complete this lab we had to apply ourselves in different ways that we haven't tried yet for instance we had to use different aspects of both Erdas Imagine and ArcMap in this lab in ways we haven't. First we had to create a LAS data set using ArcCatalog, this was the first step and we had to create this LAS data set as a Point Cloud in order to access different lidar point clout tiles. After we created the LAS data set we had to add all of the different .las files that we were given to it so than the data set would actually contain some data. We then had to explore all of the properties of the LAS data set so than we would know everything that was in it and have a better understanding of what we were trying to do for this lab. The data that we were using was representing the city of Eau Claire, we had to use this data to find out different max and min values for both X, Y and Z. In order to do this we would use the statistics tab in ArcCatalog which would give us all of the information that we needed to complete this step. After we had finished up fine tuning all of the data on ArcCatalog it was time to move the LAS data set over to ArcMap for the remainder of the lab. We had to use multiple tools during this lab, these tools included; the Identify tool, the LAS Dataset Tool bar, and many more. The LAS Dataset Tool bar helped us with a lot of different steps in this lab, it helped us look at the map in many different forms, for example we looked at the map where it specifically defined all of the areas with elevation, slope and contour which was very helpful in understanding all of the different things that ArcMap can do. The next step was to look at the layer properties of the map and change around the different classification codes and returns which would present the layout of the map in different ways. For example some would focus more the ground level of the area in question and other settings would focus on non ground areas of the map such as rivers and mountains. 

Once we got through all of those steps the next step was for us to change the LAS data set into a raster so than we could get different aspects of the map which would lead to us finding out different information about the surrounding area of Eau Claire. The first raster that we created we used 3D Analyst to help us perform the task correctly and once we had the raster created we wanted to use another tool on ArcMap called Hillshade which would help define the hills from the rest of the map. If you look at the image below you will see the first Hillshade map that we created and how it defines the hills or areas with major elevation over the rest of the map. 

As you can see in the map above that the area that really pops out and is easy to see are the hills because of the Hillshade tool that was used earlier in the lab. The next step of the lab was to do the same exact thing but with different properties being selected while changing it from an LAS data set to a raster. Then since we couldn't really see the map too clearly we had to open the map in Erdas Imagine so than we could get a more clear look at the map. Below is the map that was moved to Erdas and notice that the main visible part of the map are the rivers and man-made features. 

Overall this lab was very helpful in learning different tools and aspects of ArcCatalog, ArcMap and Erdas Imagine and discovering what they can do and being able to use these tools later in life. The lab was challenging but a good learning experience.

Lab 4 - Remote Sensing

Goal

 The main goal of this lab is to give students a new set of skills that will help them with image processing, enhancing images for visual interpretation, be in a position to delineate any study area or AOI from sets of larger satellite image scene, and many other different skills that will help each student learn more about remote sensing. After this lab is complete each student should be able to complete each of these skills and apply them to different projects as well as real life situations in the future. 

Methodology/Results

To be able to complete this lab correctly we had to apply ourselves to different sets of skills regarding Erdas Imagine that we have never used before. The first step that we needed to complete involved "image subsetting of certain study areas" in order to properly subset the image we had to use a feature on Erdas called the Inquire Box. This feature would allow us to change a certain area of the map without doing anything else to the rest of the map. Below is an example of "Image Subsetting" and the use of an Inquire Box. 

Figure 1: Image Subsetting and Inquire Box

As you can see there is only one little box of the entire map that has been changed from the rest of the map, and if you look closely you can see that the part of the map that is located inside the "Inquire Box" actually has a higher resolution than the rest of the map. This is because of the "Image Subsetting" that occurred recently which helped us better define a section of the map that we wanted to have a higher resolution than the rest of the map. 

The next step of the lab was to do basically the exact same thing but this time instead of using an inquire box to change a certain part of the map, we were asked to use a shapefile of the Eau Claire and Chippewa county. Instead of creating a more defined area this shapefile will help separate these counties from the remainder of the map. Below is figure 2 which represents the counties of Eau Claire and Chippewa being separated by a shapefile. 

Figure 2: Shapefile of Eau Claire and Chippewa Counties

The next step of the lab we had to use a few new tools that are included in Erdas Imagine, these tools include pan sharpen and a resolution merge, this of course meaning that we will have both images having the same resolution as one another. Pan sharpen helps us improve the resolution of certain images so than the images become more clear and defined. Another tool that was crucial in the completion of this lab was "Radiometric enhancement techniques" these of course did basically the same thing as pan sharpen but in a different way.

Another way that this lab helped us learn new things on Erdas Imagine is that the next step was for us to first add an image into Erdas, but the next step after that was for us to connect that image to "Google Earth." This of course helps us learn about new ways that we can interpret different images such as using image interpretation keys to determine what kind of features are around certain areas. Resampling was also a great deal in this lab, for resampling the goal was to change the pixel size of the same image and see if it actually changed from the original image and if so how it changed. The conclusion that we came to was that just changing the pixel size of the image was not enough to change the resolution of the image, which means that the actual image didn't change.

After resampling we moved onto "Image Mosaicking" which we would use when a specific study area is larger than the spatial extent of an image scene that was captured by a single satellite. The goal of image mosaicking is to display two images at once while having a smooth color transition between the two especially at the borders of the two images. Below is figure 3 which is the original image of the two images that were used before any work was done to them. 

Figure 3: Image before Mosaic Pro

The next step was to use "Mosaic Pro" to help make a more smooth transition between the two images, so than it looks like one large image because of how smooth the color transition between borders is. In order to do this we needed to put in the two images in question into "Mosaic Pro" and change certain things so than the transition between images was smooth and barley if not noticeable at all. Below is figure 4 which contains the image after "Mosaic Pro" was used. 

Figure 4: Image after Mosaic Pro

As you can see the image has a much smoother transition between one another, you can barely see where one images border ends and the other begins. The final step of this lab has to do with Binary change detection or image differencing. In order to do this we had to use a few different skills that we had recently learned just for this lab, for instance we had to display a histogram of the metadata of two images and through out this final step we had to do multiple calculations so than we could create our final product which in turn helps us in the long run of the lab. Below is the histogram that we were asked to create of the original image that was used in this final part. 

Figure 5: Image Differencing Histogram

The last part of the lab had us use this tool in Erdas called "Model Maker" this is basically a tool that would let us use different images in an equation and a function to help us create a new map all by itself. We used this tool multiple times to create different maps, one of which helped us see the progress of an image and how it developed over 20 years with new buildings and features like that. Overall, this was a very helpful lab and taught me a lot about all the different things that Erdas Imagery can do and will continue to help me learn new things on the way.