AT 319 Final Project

    The final project of AT 319 was a culmination of the labs of the course. First we needed to understand mission planning and how to operate a drone safely. Then we need to accurately gather the data and in the right projection. The project also called for the gathering data from ground control points for post processing. After the completing of the data collection the lab required good file management and meta data to maintain organization. Finally, we had to process the data and use the GCP data to correct the data in post processing. Then we were called to perform different types of analysis and represent that in generated maps. The first map generated did not require any analysis and was simply about outputting the processed data. 

The second map generated was to represent the ground control points and the coordinates recorded on or phone in comparison to the coordinates recorded by the ground control points themselves. 

The Third map generated was a comparison between original orthomosic and a hillshaded Digital Surface Model. 

The Fourth map generated was an orthomosic with an additional feature class used to digitize the roads between the fields. 

The Fifth map was generated by classifying the orthomosaic into four classes. 

The Sixth map generated was a DSM that showed all areas above 10m. This was done by using raster functions and the greater than tool to find all the area 10m above the lowest elevation.

For the last part of the lab we were required to generate a map that showed all grass areas that are within 20m of the roads and more than 20m away from the trees above 10m. To accomplish this we created two buffer maps that would be used with the extract by mask tool to find the remaining area of grass. The first buffer map showed all areas that were within 20m of the roads that we ditigitzed.

The second buffer map generated showed 20m around the areas above 10m. 

With these buffer maps we took the classified map and remove all of the classes besides grass. Then we took the first buffer map and used it to extract all the grass areas inside the 20m of the road. That new class was then used with the second buffer map of 20m near trees above 10m to extract the area outside of that range. Then we took that final class of the grass area with both buffers implemented and placed it on top of an orthomosic to provide an overview. 

 

Ground Control Points in the Field

 Previous labs have shown how to use ground control points while performing post-processing but this lab handles actually gathering GCP data in the field. We utilized ESRI Field Maps to create a feature layer over the already provided one. This new feature layer would hold the coordinates recorded on our phone for the ground control points. The feature layer provided had a sketch that provided recommendations for the GCP locations. Then we entered the field and recorded GCP points as precisely as possible in comparison to the sketch layer.   

As you look at the above generated map you can see the difference between sketch and the recorded ground control points. This highlights the importance of accurate ground control points and the ineffectiveness of mobile devices as equipment. 

Object Based Classification Cont.

 This lab is a continuation of the last lab. We are working with classifying raster images and then performing simple analysis based of those classifications. The first step of this lab was to take raster of County Park and classify it. Then we had to use the value count provided by the classification and use the pixel size to generate area for each class.

Then we took that already classified raster and broke the classes into two parent classes of permeable and impermeable to find their areas.  

Then we were provided a road and asked to provide an analysis of its condition. To accomplish this we classified the road into pavement, cracks, and vegetation. Once this was done we again used the count value to find the area of each class. Then we generate a map showing classes and their areas to shows the amount of cracks compared to the total pavement. 

 

Object Based Classification

     This week labs deals with classify rasters by using object classification inside of ArcGIS Pro. The instructor provided a demonstration on operating classification wizard and then asked us to create steps to perform object classification in the future.  My generate steps are listed below:

1. Use the Extract Bands function to better identify your object by selecting which bands to extract
1. EX vegetation
2. Open the Classification Wizard
1. Confirm Object based classification is selected
2. Select Default classifciation schema
3. Segment the Images
4. Confirm the desired
1. Spectral Detail
2. Spatial Detail
5. Create the Preview Segment
3. Training Samples
1. Remove the default classes
2. Create new parent classes
1. set values
2. set colors
3. Create specific classes
1. Select the class and use the polygon tools to train the ai
2. Collapse the polygons to create a single class
4. Select the Classifier (Support Vector Machine)
5. Run the Preview Classification
6. Run the Classification and set the output
7. Reclassify any errors
8. Finish the Classification Wizard

After completing our instructions we were ordered to classify a neighborhood into permeable and impermeable surfaces. 

Volumetrics

 This weeks lab focused on using the Surface Volume Tool to generate volume outputs alongside with understanding resampling effects on data. So first we took the mission area and clipped away the area that does not belong to the stockpile. Then we complete a multitude of resampling to 10cm and 100cm. Then we use the surface volume tool to find the volume of the pile at the non resampled pile, 10cm resampled, and the 100cm resampled pile. This was completed for three separate days of data. The table below shows that volumes calculated at each sampling size and on each data. The pictures below show generated layouts of the dates and samples sizes.  The meta data on each map indicates the data and sampling size.

Feature	Resampled 10cm Volume	Resampled 100cm Volume
0722	42,269	42.054
0827	83,888	83,667
0930	54,815	54,564

Raster Calculations

     This weeks lab focused on map algebra and utilizing it provide specific outputs. Additionally, the lab introduced the topic of resampling. The idea of resampling to create less precise but more useful information for analysis. For example, without resampling you will have much more variation inf surface level and it will be more difficult to establish a surface level for analysis. The first requirement of this lab was to generate a map depicting all of the flat areas inside of the mission area. To accomplish this goal we used the aspect analysis tool and the map calculator to find the area without a slope. 

   Then we were instructed to find all of the areas between the elevations of 233 and 234 to have an idea of water collection. 

Then we were instructed to find all the areas with a slope exceeding 30 Degrees. To accomplish this we utilized the aspect tool.

  

Then we were instructed to again show areas that exceed 30 Degrees but only above an elevation of 245 meters. To accomplish this we kept the aspect results from the last map and used map algebra to combine the limitations. 

 

 

 

Pix4D Processing and GCP Correction

     This weeks lab we worked with the Pix4D program in order to complete some post processing and to implement the idea of ground control point corrections. The first portion of the lab dealt with understanding projection and correctly importing the images into the program. Then we had to import the ground control points into the program with emphasis on projection and coordinates. After both the images and the GCP points were imported we had to input corrections for the GCP points to be accurate. This process required selecting each ground control point and then locating it inside of multiple overlapping images. The images below show the corrected ground control points on the mini box on the right side. 

GCP 1

GCP 2

GCP 3

GCP 4

GCP 5

GCP 6

GCP 7

GCP 8

   After the post processing was complete and the data was corrected using the ground control points we were instructed to generate a layout of the ground control points. To accomplish this task we utilized ArcGIS Pro to show the overall data along with detailed insets that show the location of each ground control point.