Geometric Correction

Nathan Sylte
Lab 6

Geometric Correction

Goals and Background:  

    Lab six involved the introduction to geometric correction, which is extremely important in image processing. Specifically, lab six is designed to better ones abilities on image to map rectification, as well as image to image registration. Both of these processes are important in geometric correction and are often required to maximize data extraction from satellite images.

    There were two parts to lab six. Part one involved image to map rectification. In this case, a digital raster graphic (DRG) of the Chicago Metropolitan Area was used to correct a Landsat TM image of the same area. Part two involved image to image registration of an image of the Sierra Leone region.

Methods:

    To achieve the image correction of the Landsat TM image of Chicago, two images were brought into two separate viewers. The two images included the Landsat Chicago image as well as a reference image of the same area. The Landsat image served as the input image. The control points tool was selected and the polynomial geometric model was used. Specifically, the first order polynomial model was used. Four ground control points were then added to both the reference image and the input image with a control point error of less than two percent. The image below shows the ground control points used in part one of the lab.

 Figure 1. Figure one shows the Chicago Metropolitan Area, and the ground control point locations for part one of the lab. The image on the left of the frame is the reference image.

    Image correction of the Sierra Leone satellite image involved a similar process. The control points tool was used again; however, this time the third order polynomial model was used. Twelve control points were selected. For resample methods, the method was changed from nearest neighbor to bilinear interpolation. After the processes were complete, the new corrected images were compared to their reference images to look at spatial accuracy. The figure below shows the ground control points used in part two of the lab.

Figure 2. Figure two shows the Sierra Leone region, and the ground control point locations for part two of the lab The image on the left is the reference image.  

Results:

    Below is the result from part one.

Figure 3. Figure three shows the geometrically corrected image of the Chicago Metropolitan Area.

    Below is the result from part two.

Figure 4. Figure four shows the geometrically corrected image of the Sierra Leone region.

Sources: All images are from the Earth Resources Observation and Science Center, United States Geological Survey.

Geological Survey

.

Geological Survey

.

United States Geological Survey. (n.d.). Earth Resources Observation and Science Center. Retrieved from http://eros.usgs.gov/

   

   

    





Lab 5 LiDAR Remote Sensing

Nathan Sylte
Lab 5

LiDAR Remote Sensing

Goals and Background:

    Lab five involved taking on the roll of a GIS manager where the goal was to work on a project for the City of Eau Claire, Wisconsin. Some important aspects of the project involved obtaining LiDAR point cloud information in LAS formation for the City of Eau Claire. Also, there was a quality check that occurred to ensure data quality.

    The overall objective of lab five was to become more knowledgeable about the structure and processing of LiDAR data. Specifically, lab five involved the processing and retrieval of surface and terrain models. Another important aspect of the lab was the processing and creation of intensity images and other results from point cloud.

    The lab was divided into three parts. Part one involved point cloud visualization in Erdas Imagine. Part two resulted in the generation of an LAS dataset, and part three involved the generation of LiDAR derivative products. 

Methods:

    Part one involved point cloud visualization in Erdas Imagine. Initially, the LiDAR point cloud files were saved in the (.las) format. This allowed them to be displayed in Erdas Imagine. ArcMap was then opened and a shape file of the AOI (area of interest) was added. The Erdas image was then compared with the ArcMap image.

    In part two, a new folder was created in arc catalog in which a new LAS dataset was to be stored. The LAS files where then converted into a LAS dataset. Statistics were also calculated for analysis purposes. Next, x, y, and z coordinate systems where assigned to the dataset. For the x and y coordinate system, the NAD 1983 HARN Wisconsin CRS Eau Claire (US Feet) coordinate system was used. For the z (vertical) coordinate system the NAVD 1988 system was used.

    Part three (generation of LiDAR derivative products) involved the LAS dataset to raster tool. This tool was used to generate a DSM (digital surface model) of Eau Claire that showed elevation. The parameters used are shown in figure one below. The hill shade 3D analyst tool was then used to enhance the image. The LAS dataset to raster tool was used again to generate a DTM (digital terrain model) of Eau Claire. However, different parameters were used. The cell assignment type was switched to Minimum. Last in section two of part three, a LiDAR intensity image was generated. The LAS dataset to raster tool was used again. The changed parameters included the value field reading INTENSITY, and the binning cell assignment type was changed to Average.

   

   

Figure 1. Shows the parameters used for the digital surface model.

Results:

Figure 2. Figure two shows the digital surface model from part three with the hill shade tool applied.

Figure 3. Figure three shows the digital terrain model from part three with the hill shade tool applied. Many surface features have been removed for analysis of terrain features.  

Figure 4. Figure four shows the intensity image displayed in Erdas Imagine for better viewing purposes.

Sources:

Eau Claire County. (2013). Lidar point cloud and Tile Index.

 

 Lab 5



Nathan Sylte
Lab 4 Miscellaneous Image Functions


Goals and Background:

    There were a multitude of goals for the miscellaneous image functions lab. First, we delineated a study area from a large satellite image. Second, we showed how the spatial resolution of images can be maximized for specific visual interpretation. Third, we used radiometric enhancement methods in certain optical images. Our fourth goal was to link a satellite image to Google Earth, which can be used as supporting visual information. For our fifth goal, we were exposed to different methods in satellite image resampling. This led into the sixth goal of the lab which was the introduction to image mosaicking. Our last goal, which was goal seven, was to use binary change detection by using graphical models to analyze land change.

     Overall, lab four exposed us image preprocessing, image enhancement, delineation of study areas, image mosaicking, and the use of graphical modeling to analyze land change over time.

Methods:

    Part one resulted in the sub setting of an image in which we created an AOI (area of interest). This was done by first bringing the desired image into ERDAS IMAGINE. The raster tab was then clicked on followed by the opening of an inquire box. An inquire box was then opened around the Eau Claire region. Next, we used the subset and chip function followed by the create image subset option to create our sub image.

    Part two involved the pan sharpening of an image with poor resolution to visually maximize the image. First, the desired images were imported into ERDAS IMAGINE. Then the pan sharpen tool was used along with the resolution merge option to merge the images. The multiplicative pan sharpening method was used along with the nearest neighbor resample technique to generate the pan sharpened image.

    Part three ended in the reduction of haze from an image. The raster processing tool called radiometric with the haze reduction option was used to eliminate haze from our desired image. This proved to be a quick process resulting in a more clear image.

    In part four we linked our image viewer to Google Earth. First, we brought in an image of Eau Claire that was to be linked to the Google Earth View for visual aid. We simply clicked on the Google Earth tab and then used the connect to Google Earth option. The link Google Earth to view option was then activated to match the views. Next, we used the sync Google Earth to view option to sync the views.

    Part five resulted in the changing of image pixel sizes (resampling). The spatial raster tool along with the resample pixel size option were used first to resample the image. Nearest neighbor was the method of choice, and the output cell size was changed from 30 by 30 meters to 15 by 15 meters. The resample method was then changed to bilinear interpolation and the process was repeated.

    Image mosaicking was the focus of part six. First, the image was selected that was going to be added. However, the image was not immediately added to the viewer. The multiple tab was selected first and the radio button multiple images in virtual mosaic button was then chosen. The raster options tab was then selected followed by the checking of the background transparent and fit to frame boxes. The image was then added to the viewer, and the process was repeated for the second image that was to be mosaicked. Next, the mosaic express tool was used to mosaic the images. Section two of part six focused on the use of mosaic pro. The process started by bringing in the images in the same previous manner. The mosaic pro tool was selected and the selection of the add images option followed. The image area options tab was then selected along with the compute active area button. This process was repeated for both the images that were to be mosaicked. After these steps were completed, the send selected images button was chosen to order the images. The color corrections option was also used to achieve the desired image. The mosaic was then run.

    Part seven involved the use of binary change detection to analyze changes between two images. This involved the use of the raster functions tool. The two image functions option was utilized. The new image was then brought into a viewer for analysis. The histogram of the image was used to look for changes. The change and non-change areas on the histogram were calculated in the following manner. The upper limit was calculated by taking the mean + (1.5)standard deviation and adding it to the middle value of the histogram. The lower limit was calculated by subtracting the mean + (1.5)standard deviation from the middle value. Section two involved the mapping of change pixels with the use of spatial modeler. The equation deltaBV= BV(1)-BV(2)+c was used help calculate an input value for the model. Model maker was used to generate models for part seven.

Below are the models used in the analysis in part seven.

  

After running the models, a new image showing land change was generated. ArcMap was used to generate a map using that image to show land change around the Eau Claire region from 1991 to 2011.


Results:


Part 1 Section 1.

(Figure 1.) Part one section one result showing the Eau Claire area. The Eau Claire area is our area of interest.


Part 1 Section 2.

(Figure 2.) Part one section two result showing the sub setting with the use of an area of interest shape file.


Part 2.

(Figure 3.) Result from part two showing the pan sharpened image of the Eau Claire area.


Part 3.

(Figure 4.) Result from part three showing the haze reduced image of the Eau Claire area.


Part 4.


There is no result from part four. We simply used Google Earth as an auxiliary view of the Eau Claire area. This result can be achieved by following the instructions from the methods section (part 4).


Part 5.

(Figure 5.) Result from part five showing the resampled images. On the left is the image resampled using bilinear interpolation. On the right is the image resampled using the nearest neighbor method.


Part 6 Section 1.

(Figure 6.) Result from part six showing the image generated using mosaic express (right), compared to the original image (left).


Part 6 Section 2.

(Figure 7.) Result from part six showing the image generated by using mosaic pro (right) compared to mosaic express (left).

Part 7 Section 1.

(Figure 8.) Result from part seven section one showing the histogram generated. YY represents the upper limit and -XX represents the lower limit. Upper and lower limit values are written in blue. Yellow highlighted areas show change areas.

Part 7 Section 2.

(Figure 9.) Result from part seven section two showing the area of change from 1991 to 2011 in the Eau Claire area. Areas that changed are shown in red while areas that didn't change area shown in grey.


Sources:


Satellite images are from Earth Resources Observation and Science Center, United States Geological Survey.


This is the link to lab four Link to Lab 4