What Miners Want

I attended the Commercial UAV Expo in Las Vegas at the end of October.  I gave a talk entitled “Mine Site Mapping – One Year In.”  This talk was on our experiences with performing mine site mapping services with our AirGon Services group.   Our services group is primarily about Research and Development (R&D).  We use our engagements with mining companies to discover the products that they need, accuracy levels and, most of all, how to reliably create these products.  These experiences inform both the development of our technology (the MMK, Topolyst, Reckon, the BYOD Mapping Kit) but also help us develop best practices for both collection and processing.

As I prepared for this presentation, I reviewed the mine site mapping projects we have performed over the past several years to tabulate the products our customers have requested.  These turned out to be, in decreasing order of popularity:

  • Site Volumetrics with a priori base line data
  • Site Volumetrics with no prior data
  • Site contours (“topo”) – 2 foot interval
  • Site Contours – 1 foot interval
  • Time series volumetrics (“borrow pit”)

In every case, the customer desired a site orthophoto.  In fact, they usually want an ortho of the entire site with analytic products of a subsection of the mine site.

I thought in this month’s section, I would review these products from the acquisition and processing point of view.

 Volumetrics with baseline data

I have written a few articles about injecting a priori data into a mapping project.  This is the situation where, at some time in the past, the customer has done a site survey and wants to use these data as the bottom surface of stockpiles.  Their primary desire here is for consistency from inventory to inventory.

An example of this, a large limestone quarry that we fly, is shown in Figure 1.  Here baseline data as well as a reclaim tunnel model have been provided to us as a DWG data set.  The illustration of Figure 1 shows these data being used by Topolyst to create a 3D base surface.

 

Figure 1:  Bottom Data with reclaim tunnel model

Figure 1: Bottom Data with reclaim tunnel model

The primary challenge that we have when receiving a priori data is the accuracy of the data.  We often find that these data were obtained by traditional stereo photogrammetric collection techniques so we do not have a point cloud from which to assess accuracy.  Now, done properly, stereo photogrammetry produces survey grade data.  Unfortunately, much of this a priori data was collected with the surface obstructed by existing stockpiles; in other words, it was not a stockpile free base data mapping.  This means that the stereo compiler had to estimate locations under the existing data.  We find that in most cases, these estimations are simply linear interpolations from one side of the obscured area to the other.  We often find these bottom models extending above the current surface.  It is difficult to tell if the data were incorrectly modeled or if the ground has actually changed from the time the baseline data were collected.

A second big challenge we have with these data are a lack of knowledge by the provider as to the exact datum to which the data are referenced.  We are often concerned with elevation differences of just a few centimeters.  The Geoid model really matters when you are approach survey leveling accuracy goals.  We have found, on more than one occasion, a priori data with an incorrect vertical model.  This usually occurs (at least in the USA) as a result of using the incorrect NAD83 to WGS84 transformation.

Over the past year, we have added a lot of refinements to how Topolyst handles this a priori data.  Those of you who do LIDAR or photogrammetric processing will immediately recognize this as the problem of introducing “breaklines” and “mass points” into a model.  LP360 (Topolyst is just a variant of LP360) has always been a very strong product in terms of breakline modeling.  We have added a few features in this area to improve the modeling as it typically applies in UAS mapping.  We are now at the point where we really do not have any software issues with this sort of modeling but the interpretation problems will always remain.

This type of modeling requires:

  • Direct geopositioning (RTK/PPK) on the drone
  • Multiple surveyed check points on the site for data validation
  • Strong modeling tools such as Topolyst
  • A conference or two with the customer to understand the models
  • A lot of patience when defining stockpiles

Volumes with no a priori data

Here the customer is interested only in the volumes of the piles, without regard to location.  The deliverable is generally a spreadsheet with volume, material type, density and tonnage.  Of course, our customer deliveries are via our cloud data platform, Reckon, so we want the toes to be correctly georeferenced.

If you leave out the correct georeferencing (meaning you compute the volume of the pile but do not necessarily try to align it with an existing map), you have the sort of processing offered by a myriad of web-based solutions such as Kespry.  Under this business model, you typically upload the raw drone images which have been georeferenced by the navigation grade GNSS for x, y and the drone barometric altimeter for elevation.  This typically provides horizontal accuracy on the order of several meters and vertical accuracies at about 5 meters.  So long as the camera is properly calibrated, this methodology leads to volumetric accuracies that are accurate to within about 5%.

We never do these projects without some check points.  These are surveyed image identifiable points that we use to check horizontal and vertical accuracy.

The biggest issues we have encountered with this type of project is the definition of the stockpile toe – it is somewhere between comingled piles, it traces along an embankment such as the pit, the stockpile is in a containment bin and so forth.   There requires a lot of careful toe editing in a three dimensional visualization environment such as Topolyst.

We never have issues with accuracy because we always fly with a direct geopositioning system.  For our MMK, it is a Post-Process Kinematic, PPK, GNSS system.  For the senseFly eBee, it is an onboard RTK system.  We always lay out some checkpoints for project verification.

A very clean mine site with stockpiles sitting on a surface is nearly non-existent (except in our dreams).  While you sometimes encounter sites where you can just manually draw a toe, these sites are nearly always at inventory transfer locations, not working mines.  In fact, of all the mine sites we have surveyed, we have encountered only one “groomed” site (see Figure 2).  Even at this site, the upper left and lower right piles required some disambiguation (wow, that’s a big word!) work to separate the pile edge from encroaching vegetation.

Figure 2: A "groomed" inventory site

Figure 2: A “groomed” inventory site

 Site Contours (“topo”)

A surprising number of customers want contours.  As you know, these are elevation isolines at a particular interval.  Most customer want either 2 foot or 1 foot contour intervals.  These data, in DXF or DWG format, are used as input to mine planning software.  I find this a bit odd since I would think by now that this downstream software would directly ingest a LAS point cloud or at least an elevation model.

Contours are always absolutely referenced to a datum (a “Network”).  This can be a local plant datum or, much more commonly, a mapping horizontal and vertical datum such as a state plane coordinate system for horizontal and NAVD88 with a specific geoid model for vertical (at least in the United States).

You can tie to the datums using either direct geopositioning with onboard RTK/PPK or you can use dense ground control points.  I personally would never collect data that must be tied to a datum without having a few image identifiable checkpoints.  Unfortunately, this means that you will need at least an RTK rover in you equipment kit.

A good rule of thumb for contours is that the accuracy of the elevation data should be at least three times the accuracy of the desired contour interval.  This says if you are going to produce 1 foot (30 cm) contours, you need 4” (10 cm) of vertical accuracy relative to the vertical datum.  When you measure your checkpoints, don’t forget to propagate the error of the base station location (which you might be deriving from an OPUS solution).

Preparing a surface for contour generation is perhaps the most tedious of mine site mapping work.  It is generally the only site mapping you will do that requires full classification of ground points (the source for the contour construction).  An example of 2 foot contours within a mine site is shown in Figure 3.

Figure 3:  An example of 2' contours

Figure 3: An example of 2′ contours

Sites with a high degree of vegetation in areas where the customer wants contour lines will have to be collected with either manual RTK profiling (very tedious!) or with a LIDAR system.  You simply cannot get ground points with image-based Structure from Motion (SfM).  No surprise here – this is why LIDAR was adopted for mapping!

If the customer does not want to pay for LIDAR or manual RTK collection, the vegetated areas should be circumscribed with “low confidence” polygons.  You can either exclude the contouring completely from these areas or classify the interior to vegetation and let the exterior contours just pass though the region.  In any event, the customer must be aware that the data are quite inaccurate in these regions.

The SfM algorithm gets quite “confused” in areas with overhead “noise” such as conveyors and vegetation.  This confusion (actually correlation errors) typically manifests as very low points.  You will need to find and clean these points prior to contour generation.

Conclusions

Product generation for UAS mapping requires a lot of front-end planning.  This planning needs to be product-driven.   If you customer (you, yourself, perhaps) needs only volumes with no tie of the toes to a datum, you can get away with no control so long as some other information such as camera calibration and flying height are correct.  By the way, we recommend never collecting this way since you are precluded from doing any meaningful time series analysis.

On the other hand, most meaningful data (that is, you can quantify the accuracy relative to a datum) will require a very careful control strategy as well as a rigorous processing workflow with the right tools (meaning Topolyst, of course!).  No matter what geopositioning strategy you employ, you should always have some independent methods for verifying accuracy.

If all of this seems a bit daunting, you can get assistance from us.  Remember, our services group is really our R&D lab.  Our real goal is to sell technology to owner/operators and production companies.  No matter what drone you are using, you can always avail of our consulting services.  We have gained a lot of experience over the past few years, mostly by first doing the wrong thing!  Save yourself this time and money by engaging with us!

 

 

 

FacebookTwitterGoogle+LinkedInShare

AirGon Happenings

I am pleased to announce that AirGon’s request for amendment to its Section 333 waiver for flying commercial small Unmanned Aerial Systems (sUAS) was approved in April.  Our amendment adds all current and future 333 approved aircraft to our 333.  AirGon can now fly any sUAS that has ever been approved by the FAA as well as all future approved systems.  This list currently contains 1,150 different sUAS (AirGon’s own AV-900 is number 207 on the list).  This provides us a lot of flexibility in working with clients; for example, in situations where a glider sUAS is more efficient than a rotor craft.

The FAA has also recently streamlined the process of obtaining an N number for a sUAS.  Prior to the change, a paper process that required several months was the only option.  Now an online system is available, greatly simplifying this procedure.  Note that this is not the new online registration system for hobby drones but rather the system used for obtaining an N number for a manned aircraft (if you are confused, join the club!).  Combined with our new 333 amendment, we can now get a new aircraft legally operating within days.

We continue to do a lot of work to optimize the accuracy of point clouds derived from dense image matching (DIM).  DIM are the data of choice for sUAS mapping since they can be generated from low cost prosumer cameras using standard application software such as Pix4D Mapper or PhotoScan.  The question always remains as to how good these data really are.

It has taken us a lot of experimentation and analysis but we think we have fleshed out a procedure for assuring good absolute vertical accuracy.  It involves the use of Real Time Kinematic (RTK) Global Navigation Satellite System (GNSS) positioning on the sUAS, a local base station that we tie into the national Continuously Operating Reference Station (CORS) network and the National Geodetic Survey’s Online Positioning User Service (OPUS) to “anchor” the project to the network.  We have also discovered that high vertical accuracy cannot be obtained without camera calibration.  We typically use an in situ process for calibration.  We have flown many dozens of sites (primarily mining), giving us a rich set of test data.

I cannot over emphasize how critical network vertical accuracy is.  Most customers want elevation maps of their sites.  These are usually delivered as contour vector files.  As we all know, a 1 foot contour requires vertical accuracy of 1/3 of a foot.  This is a very tight requirement!  A three inch vertical bias error over an acre is an error of about 400 cubic yards – this is significant.

We see a lot of drone companies processing site data with no control and no RTK/PPK.  While, with the introduction of scale into the model (many companies do not even do this), one might obtain reasonable difference computations (such as volumes), the network accuracy is very poor (obtained from the airborne navigation grade GNSS only) and hence the data are of limited use.  We have discovered that these techniques (where no control and/or RTK/PPK is used) can result in the vertical scale being incorrectly computed.  This means that even differential measurements are not accurate.  Why spend all of the money to collect these data if they are of unknown accuracy?

A more difficult area that we have studied over the past several years is what I refer to as “conformance.”  That is, how well does the DIM actually fit the object being imaged?  DIM processing software (again, such as Pix4D and PhotoScan) do a miraculous job correlating a 3D surface model from highly redundant imagery using the general class of algorithm called Structure from Motion (SfM).  In addition to the obvious areas where SfM fails (deep shadow, thin linear objects such as poles and wires), a lot of subtle errors occur due to the filtering that is performed by the SfM post-extraction algorithms.  These filtering algorithms are designed to remove noise from the surface model.  Unfortunately, any filtering will also remove true signal, distorting the surface model.

We are working with several of our mining customers to quantify these errors and, once these errors are characterized, to develop best practices to minimize or at least recognize when and where they occur.  An example of an analysis is shown in Figure 1.  Here we are analyzing a small pile (roughly outlined in orange) of very coarse aggregates with a volume of about 340 cubic yards.  This site was flown with a very high end manned aircraft LIDAR system and with AirGon’s AV-900 equipped with our RTK system.  The DIM was created using Agisoft PhotoScan.  We obtained excellent accuracy as determined by a number of signalized (meaning ground targets visible in the imagery) control and supplemental topo only shots.  We used in situ calibration to calibrate the camera (a Sony NEX-5 with a 16 mm pancake lens).

As can be seen in Figure 1, we created a series of cross sections over the test pile.  These cross sections were generated using the Cross Section Point Cloud Task (PCT) in LP360/Topolyst.  This tool drapes cross sections at a user specified interval, conflating the elevation value from the user specified point cloud.  We ran the task twice, conflating Z first from the LIDAR point cloud and then from the DIM.   In Figure 1 we have drawn a profile over one of the cross sections with the result visible in the profile view.  The red cross section is derived from the LIDAR and the green from the DIM.

Comparing LIDAR (red) to DIM (green)

Comparing LIDAR (red) to DIM (green)

Note that the DIM cross section (green) is considerably smoother than the LIDAR cross section (red).  This is caused by several factors:

  • The aggregate of this particular pile is very coarse with some rocks over 2 feet in diameter. This leaves a very undulating surface.  The LIDAR is fairly faithfully following this surface whereas the DIM is averaging over the surface.
  • The AV-900 flight was rather high and the data was collected with a 16 mm lens. This gave a ground sample distance (GSD) a little higher than is typical for this type project.
  • Due to the coarseness of the aggregate, significant pits appear between the rocks, creating deep shadows. SfM algorithms tend to blur in these regions, rendering the elevation less accurate than in areas of low shadow and good texture.

The impact of lower conformance is a function of both the material and the size of the stockpile (if stockpiles are what you are measuring).  For small piles with very coarse material (as is the case in this example) a volumetric difference between LIDAR and SfM can be as great as 20%.  On larger piles with finer aggregates, the conformance is significantly better.   For example, in this same test project, we observed less than 0.25% difference between LIDAR and the DIM on a pile of #5 gravel containing about 30,000 cubic yards.

There still remains the question of which is more accurate – the volume as computed from the LIDAR or the volume as computed from the DIM?  I think that if the LIDAR are collected with a post spacing ½ the diameter of the average rock, the LIDAR will be the most accurate (assuming that it is well calibrated and flown at very low altitude).   However, the DIM is certainly sufficiently accurate for the vast majority of aggregate volumetric work, so long as a very strict adherence to collection and processing best practices is followed.  For most high accuracy volumetric projects, manned LIDAR flights are prohibitively expensive.

We continue to do many experiments with local and network accuracy as well as methods to improve and quantify conformance.  I’ll report our results here and in other articles as we continue to build our knowledge base.

Your Business Model, not Ours!!

We have invested a tremendous amount of resources (monetary, development, knowledge) into developing technology and services for mapping sites using dense image matching collected with small Unnamed Aerial Systems (sUAS). Our focus is applications suitable for an sUAS (non-populated areas, smaller sites) that require near survey grade accuracy. The most common example is small open pit mine sites such as quarries. We have not considered agricultural applications since these tend to be very large areas where radiometric analysis is the focus rather than geometric correctness.

Like most other companies involved in this emerging market, we are trying to predict the most palatable business model. However, I would say that unlike many other technology providers, we are seeking the business model that makes the most sense for the customers, not for us.

AirGon LLC has a very big advantage over companies funded by venture capitalists. We are funded both by GeoCue and by investments from our small group of inside shareholders. This allows us to focus on a long-term vision of the market. We plan to become the “go to” company for sUAS mapping, much as GeoCue has become the “go to” company for airborne and mobile laser scanning.

One of the big questions that Venture Capitalists have in funding a startup is that of scale. If the venture will not scale up to a sufficient size to provide a comfortable multiple on the initial investment, the venture is not considered financially viable. In the sUAS business, it is hard to devise a model that will scale that does not require significant involvement on the part of the customer. The most popular model is a leased plan where the customer flies the drone and uploads the image data to a cloud-hosted system provided by the vendor. In some of these models, the customer may even do the data extraction such as defining the base of a volumetric stockpile.

These “self-service” business models proliferate in the rollout of new technologies that are generally called “Web 2.0” (or are we Web 3.0?). You now see it with everything from reservation booking systems to the Uber taxi concept (in the Uber case, the job of “dispatch” has been handed over to the customer). Even grocery stores have gotten on this bandwagon with self-service checkout kiosks.

We certainly believe that self-service will play a major role in the emerging sUAS mapping business. However, at the current time one size does not fit all. This is particularly true in light of the draconian FAA regulations that currently exist for commercial sUAS operations. A mine site leasing a “fly-it-yourself” drone would require an FAA 333 exemption as well as an FAA licensed pilot. This is a fairly significant barrier to adoption of the technology. In addition to the legal hurdles, many customers want to nibble into this new approach to mapping rather than wolf it down in one gulp.

We launched our CONTINUUM program as a way to address these customer needs. CONTINUUM allows a customer to pick from a menu of hardware, software and services that best suit her needs. A few customers want to buy a mapping kit and do it all themselves. For this customer we offer the AV-900 Metric Mapping Kit (in both base and RTK versions). Other customers want to fly their own equipment but have the data processed as a service. Still others want to have us provide full services where our Field Service Analyst shows up at their site and performs the complete job. Under CONTINUUM, we can provide what the customer wants, not what we think might be the best business model for us.

One of the real values behind CONTINUUM (and the reason for the name) is that most customers do not know what they will want to do as a final business model. They would like to be in an environment where they can experiment a bit. This is exactly what we provide through the CONTINUUM program. A customer can modify the business model from one of AirGon doing everything to they, themselves,  internalizing the entire process or any mix in-between, all without the need to change vendors.

I am not sure what will be a profitable business model for AirGon. We are still very heavily in the Research and Development mode. However, one thing I do know for sure – the successful business model will be the one that is deemed successful by the end-use customer. We intend to be the provider of that ultimate solution!

 

 

Top Ten Considerations for Selecting a Drone Mapping Services Vendor

You realize that significant benefits would be realized by transitioning mine site mapping/volumetrics to drones (more properly, small Unmanned Aerial Systems, sUAS). You have decided, at least for the immediate future, to use an outside service provider rather than internalize the process.

Since you have, at least for the present, decided to outsource drone-collected mapping and volumetrics, the task now is to select a qualified company to perform these services. A checklist for evaluating a potential service provide should include these questions:

  • Is the vendor authorized to fly by the appropriate regulatory body (e.g. in the USA, the Federal Aviation Administration requires a Section 333 exemption permitting commercial drone flights)?Drone Picture
  • Does the vendor have sUAS aircraft liability insurance?
  • Are the rights to the collected data clearly spelled out?
  • Do you feel confident that the vendor’s methodology for rigorous network/local accuracy (surveying accuracy) will meet your requirements? For example, a 4 inch vertical error in a borrow pit computation amounts to about 538 cubic yards of volumetric error per acre!
  • For projects that require Network Accuracy (anytime you intend to extract information such as elevation models, contours or are performing time series analysis, you will need Network Accuracy), can your service provider tie their results to a reference network that can be independently verified?
  • Does the vendor have a plan for incorporating surveyed quality assurance check points that will be captured in the aerial flight?
  • Does the vendor understand how to incorporate design information such as “bottom” lines, reclaim tunnel models, complex a priori stockpile toes and so forth into the modeling process?
  • Does the vendor have a reasonable approach to allowing you to collaborate on resolving project boundaries, stockpile identification, stockpile toe definitions, occluded areas and so forth?
  • Have proposed ground personnel worked on mine sites and have safety awareness? For example, for USA mine site operations, do they have basic MSHA Part 46 training?
  • Can the vendor provide references?

You should engage in a pilot project with your candidate vendor. This will limit your initial investment and give you an opportunity to fully vet the proposed provider before committing to a long term relationship. You will want to have independent test data to validate the vendor’s solution.

An immediate red flag is a potential vendor who will not explain their methods in detail, hiding behind a veil of “well, that is our proprietary method that sets us apart from our competitors.” The plain English translation of this is “I have no clue!”

Drones, Metric Mapping and RTK

We have been very busy this first third of 2015 with software development (as we always are).  The thing about software is that it is never static.  It is either undergoing new additions or entering the end of life phase.  We have had a very big focus on ensuring that our products are optimized for LAS 1.4 support as this is the new requirement of the USGS.  Additionally, we like to use LAS 1.4 in our mine site workflows since it supports a few nice capabilities that were not in LAS 1.3.

This is definitely the year of the drone.  Every major geospatial hardware firm has announced a drone system for remote sensing (some for metric mapping).  While the USA is inching along toward some usable drone rules, other countries have clear rules in effect and drone mapping is becoming a standard survey/mapping tool.

We are garnering a very high interest in AirGon’s Metric Mapping Kit (MMK).  This solution provides everything you need to do uncontrolled mapping projects using a small Unmanned Aerial System (sUAS) except a processing laptop computer.  Add in your own surveyed control points to reach survey grade accuracy.

Speaking of the Metric Mapping Kit, we will be hosting a AV-900 MMK workshop in Toronto, Canada on June 11th and 12th.  Thanks to Jim Giordano, we will be presenting live flight demonstrations at VicDom Sand & Gravel as well as an in-depth look at mission planning and post-collection data processing.  Our focus will be on drone-collected volumetrics. Personal protection equipment (steel toed boots, hardhat, safety vest and safety glasses) are required.  Remember that a passport is required for travel between the USA and Canada.  Space is extremely limited so sign up early!

We have been (in a joint project with Applanix, a Trimble Company) researching the use of Post-Processed Kinematic (often erroneously called Real Time Kinematic, RTK) control solutions.  Obviously everyone flying a sUAS for metric mapping purposes would like to dispense with the tedium of deploying ground control.  We will publish the results of our efforts as a white paper when the work is complete.  My goal is a recipe, if you will, of the methods that are appropriate for a given desired accuracy level.

We will be posting an experimental (EXP) release of LP360 (all license levels) within the next few weeks.  Those of you on software maintenance will be able to download this release via the “Check for Updates” option under LP360 Help.  There is a separate article in this newsletter that provides a highlight of the new features.

Till June – Best Regards,

Lewis

GeoCue Group News – May 2015

sUAS – Where will this business go?

The small Unmanned Aerial Systems (sUAS) business is very appealing. For less than US $20,000, you can outfit a complete system for collecting aerial imagery and processing the data into an array of high quality mapping product.

But who will roll out these new low cost mapping systems?  Will it be the major airborne acquisition companies?  Perhaps, but with a business model predicated on large collects, does this fit?  Will it be the owners of the sites that require mapping such as quarry owners, land developers, coal fired power plants?  Or will it be professional land surveyors who offer sUAS mapping as another tool in their toolbox?

In my mind, the professional surveyor is best equipped to roll out this new business tool.  The PS is already tuned to a business model of travelling to small sites, collecting  data, processing results and consulting with the client.  The sUAS will provide a new tool that will allow the PS to offer a broader range of more accurate services to the client base.  For example, rather that delivering estimated elevation models based on a few RTK points, she can now deliver very dense point cloud derived models based on dense image matching.

Perhaps the most exciting new business opportunity is the rapid collection of accurate volumetric data.  Today this is done either by manned aerial mapping or by ground based techniques.  Ground based techniques are very problematic for many situations since accurate data collection of complex or tall stockpiles is very difficult.  Manned airborne methods work extremely well but are prohibitively expensive for high frequency monitoring (even quarterly monitoring is not practical except for the most valuable of stockpiles).  Enter the sUAS.  A flight of 20 minutes can provide the base data necessary for very detailed volumetric computations over a typical 1 square kilometer area.  In fact, the entire process, from mission planning to client deliverable can be performed in less that one day.

The sUAS is upon.  Enterprising folks will figure out very quickly how to produce professional products at a profit.

(Read the GeoConnexion article describing our experience of putting together an sUAS system.)