Field Activity #3: Unmanned Aerial System Mission Planning

Introduction

The goal of this exercise is to improve critical thinking when planning for different scenarios encountered by geographers. Five different scenarios were given with a goal of devising a plan on how to solve the scenarios.  While planning for the scenarios the use of a UAS (Unmanned Aerial System) was highly recommended to be a big factor in the solving process because the scenarios involved an image of the area to be taken.  For each scenario a plan was thought through to include: costs, type of UAS, type of sensor, GIS software, time of year and any other factors that were needed to complete the process.  However, because of the inexperience of the class, only the leg work of the scenarios were thought through to give an overview on how to solve the mission. 

Scenarios



 Scenario 1

v  A military testing range is having problems engaging in conducting its training exercises due to the presence of desert tortoises. They currently spend millions of dollars doing ground based surveys to find their burrows. They want to know if you, as the geographer can find a better solution with UAS.

Using UAS to survey for desert tortoise burrows is a much quicker and more cost effective way to discover where the burrows are compared to ground based surveys. There are two main options that can provide high quality data for this kind of survey; LiDAR and supervised classification using aerial imagery.

LiDAR can be used for this project because it collects elevation data in the form of a point-cloud. The LiDAR sensor shoots a laser at the ground and as the beam is reflected back it records the elevation it was reflected at. The LiDAR sensor requires a large UAS because of its weight so most rotary propeller UASs are out of the question but some fixed wing options will work such as in figure 1 below



figure 1: A fixed wing UAV, capable of being equipped with a LiDAR sensor

Once the LiDAR data has been processed a DEM (digital elevation model) will be created. After knowing how deep the tortoise burrows are a base height should be set that is that many feet/inches above the base height of the data. This will create a DEM with the negative elevation representing the tortoise burrows.

This option is costly but if millions of dollars are being spent on ground based surveys it would be well worth it to use a UAS in this fashion. A second option which will most like be much less expensive would be to fly a UAS and to have it take images of the ground and from these images use a supervised classification to automatically pick out where any tortoise burrows may be.

A supervised classification works by having the user select representative areas using reference sources such as high resolution imagery. The software then characterizes the statistical patterns of the representative areas and classifies the image. The use of a multi-band camera makes the classification scheme much more accurate. This is because the camera records data from a scene as individual color values. From these values a spectral signature can be derived. Using this signature, software such as ERDAS Imagine, will select pixels on the image which are within a specified range of the signature creating an image with one color representing a specific feature such as blue for all water.

This will reduce time in discovering tortoise burrows because the burrows have a unique spectral signature. Since the upturned soil will stand out from the ground it will be easy to select the burrow on an image and specify that all pixels with similar spectral signatures should be classified the same.

This process does involve some ground truthing to verify that the classified burrows are actually burrows and not randomly selected pixels on an image that happen to be similar. Having the person classifying the images will be best because they will know the exact area of where the burrows are.  

A camera that captures imagery in multiple bands that would be excellent for this kind of task is the UltraCam shown in figure 2 below. This camera will produce high quality images with the capability to be used in a supervised classification.

Figure 2: Ultra Cam camera capable of taking images in panchromatic, red,
green, blue, and infrared channels



Scenario 2

v  A power line company spends lots of money on a helicopter company monitoring and fixing problems on their line. One of the biggest costs is the helicopter having to fly up to these things just to see if there is a problem with the tower. Another issue is the cost of just figuring how to get to the things from the closest airport.

Instead of using a helicopter and having someone investigate power line issues it would be much safer and more cost effective to use a rotary UAS (unmanned aerial system). The rotary UAS will be able to fly extremely close to the power line without risk of major damage to the pilot or anyone else if it comes in contact with the line. This is because of how the propellers on the UAS are positioned; they allow for a stable flight with the ability to make sharp turns. Figure 3 shows an image of a rotary UAS. Notice how the propellers are evenly distributed around the center of the vehicle. Pictures of any damage can be taken with ease because the rotary UAS is able to hover in place and can provide not only pictures of the damage but real time video of any issues.

Figure 3: Rotary UAS equipped with a camera, propellers allow
the the camera to stay stable

A major advantage to using a UAS like this is that you can launch and land the vehicle from virtually anywhere. Not only will this rid the need of an airport but it will also eliminate having to waste time waiting for a helicopter to arrive near the power line. Having a helicopter fly close to power lines creates an issue of pilot safety and also the safety of anyone who may be on the ground. Cameras can take amazingly high quality images from a distance but even then you could receive higher quality by using a similar camera mounted onto a rotary UAS and have it fly in and hover much closer to the power line.

A disadvantage to using the UAS is that typically these types of vehicles have less flight time. This is where a helicopter outdoes the UAS. Even though the flight time may be less the cost of a potential injury to anyone involved in surveying is nonexistent with the UAS since the pilot can be stationed almost anywhere.

Scenario 3

v  A pineapple plantation has about 8000 acres, and they want you to give them an idea of where they have vegetation that is not healthy, as well as help them out with when might be a good time to harvest.

When examining the task of finding healthy vegetation over an 8000 acre area the cheapest option I can think of would be to download a LANDSAT image for the area then examine the infrared color band. LANDSAT is an abbreviation for Land Remote-Sensing Satellite which is in orbit around the world with an interval rate of 16 days for the newest satellite (LANDSAT 8). What that means is that every 16 days there will be a new image for the same area. LANDSAT has sensors which are able to record light reflectance from the ground similar to what a normal camera would do but it can also record the infrared energy being emitted which can be used for vegetation analysis because the healthier a plant is the more infrared energy it will emit which will be recorded by the sensor. The files downloaded from LANDSAT represent each band the satellite records light in (red, blue, green, infrared, shortwave infrared, etc.). These bands come in black and white TIFF files which are able to be used/opened in virtually any kind of image manipulation software. The TIFF files are black and white because of how the sensor records the color for each band. For anything blue, such as water, the pixels that make up the water will have a higher pixel value than pixels for land. The same principal applies to green objects such as plants and grass and so on for other colors. The infrared band will give higher pixel values to pixels representing objects that emit more infrared radiation than other objects. The infrared band would be opened using any kind of standard image viewing software. The more white an area is the more infrared energy being emitted thus the healthier the vegetation. In figure 4 below you can see that agricultural fields are much healthier and ready to be harvested than other natural areas in the image. 

Figure 4: Landsat image of healthy vegetation appearing in white, the red

circles are showing the healthy vegetation

This option is completely free as long as you have an internet connection and a way to unzip the downloaded file then be able to view the files. Although this option saves a lot of money it does have a few downfalls. First, since the satellite is on a 16 day interval you won’t be able to have images be taken on demand and even if you find an image for a date you want there is a chance it could be filled with clouds which would distort or even block the ground altogether. Assuming you go with this method of using the LANDSAT images you may run into an even bigger problem which would make you start over completely; satellite failure. This has already happened to the previous LANDSAT 7 satellite. The images taken from LANDSAT 7 would be of similar quality to LANDSAT 8 but they include a large amount of missing pixel data so all of the images produced are virtually useless for any kind of analysis like checking on the health of a pineapple plantation.

A second option would be to attach an infrared camera onto a fixed wing UAS (unmanned aerial system) and have it fly over the plantation recording infrared radiation producing an image which would be very similar to the one produced by LANDSAT. Figure 5 below shows an infrared camera capable of being attached to a UAS. This option of using a UAS will include a cost of a couple thousand dollars, most of which going to infrared sensor and UAS, but the money saved in not having workers check on the entire plantation’s health might be worth it. By using the UAS you would be able to have on demand infrared images taken of the plantation instead of waiting and hoping that the image from LANDSAT is of high quality.

Figure 5: Infrared scanner capable, used to take images in infrared

To discover the best time to harvest you could examine the infrared images to see when the plantation is mostly white meaning healthy. By using LANDSAT images you have access to images from previous years so you could start to see a trend in when the plantation is at its peak health and ready to be harvested. The LANDSAT images would give a good approximation of time to see this trend but the use of a UAV with an infrared would give a better look at exactly when the plantation is at peak health. Since LANDSAT is free to use it may not be a bad idea to investigate those images and to use the UAV in conjunction.

Scenario 4

v  An oil pipeline running through the Niger River delta is showing some signs of leaking. This is impacting both agriculture and loss of revenue to the company.

First many factors need to be accounted for, the agriculture could be also affected by other factors including a drought, bad soil, and over production.  Also the Niger River is known as being one of the most polluted rivers in the World, thus fixing the oil leaking might not lead to wasted agriculture area or crops.  Many questions will need to be asked before starting the project including: what time of the year is it?  This will affect the river water level and the spread of the oil.  If the Niger River water level is high the disperse of the oil leakage will be effecting the crops more.  Also, the description of the crops should be known, are they being harvested at this time or is the season in a transition?  First an image of the area should be taken to find out where the leakage is occurring.  When looking for an oil leakage, areas of black should be identified, the color of oil.  Also the area of black will be most heavy near the leak and then start to spread out as it travels down the river.  If the river is relatively clear, which should also be known before taking the image, the oil leak should be relatively easy to find.  This image can be taken either by an UAV (unmanned aerial vehicle) controlled by a computer or by a balloon, depending on the expense of the equipment and weather.  The disadvantage of using a UAV to take the image is it will be expensive ranging in the thousands, but it will be the easiest and most efficient way to take the image with the range the UAV can have.  A ‘normal’ high quality camera should be fine for finding the oil leak, no special effects on the camera or image should not have to be used.  The advantage of using a balloon to take the image is it will be very cheap and relatively easy to use compared to flying a UAV.  The disadvantage is the balloon may be hard to control with the wind and the range the balloon has compared to the UAV will be less.  However, a third option can be used, to get more accuracy, to determine the oil leakage by looking at vegetation health using a near infrared sensor. The health of the agriculture should be in most danger surrounding the oil leakage then getting healthier when moving away from the leak.  The near infrared image will show the healthy vegetation appearing in white and the unhealthy vegetation converting from gray to black.  Knowing where the agriculture is most unhealthy will help determine the area of oil spill.  This device will be more expensive and will have to be used by an unmanned aerial system because of the risk of losing the sensor. 

Using the UAV to take an image of the Niger River Delta to find the oil leak is the best option in this scenario.  It will on the higher end of the cost but with a serious problem, like an oil leak, the best option should be used.  Also using a near infrared scanner to look at vegetation could also be used along with the UAV.  After these steps are taken and clean images are produced the oil leak should be able to be found and fixed, helping the revenue and stopping the contamination of crops.  Two links that sell UAS; the first is less expensive of less quality and the second being more expensive and having more options of UAVs.

Figure 6. CAPTION: Image of a UAV, being placed in the air ready to be flown
around and used to capture images.  The military uses UAVs to
capture aerial images of images. 

  

Scenario 5

v  A mining company wants to get a better idea of the volume they remove each week. They don’t have the money for LiDAR, but want to engage in 3D analysis.

In order for you to figure out how much ore you are removing from the open pit mine, you will need to obtain 3 dimensional images of the mine to ultimately create a DEM (digital elevation model) of the mine. Obtaining these 3 dimensional images can be done through Photogrammetry camera systems mounted on a fixed wing UAS. Photogrammetry camera systems have automated film advance and exposure controls, as well as long continuous rolls of film. Aerial photographs should be taken in continuous sequence with an approximate 60% overlap. This overlap area of adjacent images enables 3 dimensional analysis for extraction of point elevations and contours. Once the images have been shot by the fixed wing UAS, a technique called least squares stereo matching can be used to produce a dense array of x, y, z data. This is commonly called a point cloud. A DEM image like the one below (figure 7) can then be modeled in ArcGIS to accurately reflect contours of the mine as well as the elevation levels of the mine.

Figure 7: Digital Elevation Model, showing elevations.  Red higher
elevation and blue lower elevation

Since you will know the elevation levels of the mine, every new DEM created with subsequent point clouds will reflect the elevation changes that occurred over a given period of time. This change in elevation will allow you to see the volume of ore being taken out of the mine. Obtaining an elevation point cloud with a fixed wing UAS equipped with a photogrammetry camera system, is much faster than manually surveying the mine. It can be done as often as needed with relative ease, saving your company massive amounts of time and ultimately money. This method is not as accurate as using LIDAR data, but it is much cheaper. If you were to take weekly readings of the mine using LIDAR you would spend a fortune on data collection. I see photogrammetry as your most viable option if you are set on taking weekly volume tests.