top of page

How TraceAir Archives and Verifies 1/10 of a foot Accuracy

There are 4 key steps in our process which result in high accuracy.
Each step plays a critical role and is described in detail below.

1. Create Flight Markers

 

Ground Control Points (GCPs) are used to georeference the scan accurately. Our experts recommend locations from extensive experience. GCPs located along the perimeter determine the boundary of where the scan will be accurate. These GCPs coupled with a limited number of GCPs populated throughout the site reduce distortion
by anchoring your scan to known points.

 

Quality Control Points (QCPs) are used to verify the accuracy of the scan. By placing QCPs between GCPs in select locations, omitting them from photogrammetry, then comparing the actual elevation of the points to the elevation of the points in the scan we can quantify the quality of the scan.

 

Flight markers (GCPs and QCPs) are created by placing two foot
by two foot “X” marks on the ground, so they are visible in the drone photographs taken from above. Precision GPS equipment is used to capture the coordinates and elevation of the center of these marks.

2. Perform Drone Flyover

 

For those with a drone program, TraceAir will help streamline your operations and improve the accuracy of your scans. Common guidelines to follow are:

 

  • Fly at the proper altitude and keep a constant absolute altitude
    for the entire scan

  • Maintain the proper frontlap and sidelap settings

  • Perform the flight as close to midday to minimize shadows
    and fly as fast as possible to minimize variable lighting conditions

  • Upload the photos of your scan to our platform to automatically begin the photogrammetry process

 

For those looking to start a drone program, TraceAir will help with the above and provide recommendations for which drone and camera combination to purchase. Not all hardware repeatedly produces good results in photogrammetry so we help select and test the drone and camera that fits your needs.

For those without a drone program, TraceAir will utilize its network of third-party pilots to coordinate all of the drone operations to save you the financial and time investment.

Drone in construction.jpg
TraceAir laptop.png

3. Photogrammetry: Data Processing and Georeferencing

 

Upon completion of the creation of the sparse point cloud, the TraceAir team will georeference the scan. They use the GCPs visible in the pictures to assign specific coordinates and elevations to these markers which stitch the 3D sparse point cloud into the proper system of coordinates. Both global (NAD83, WGS84) and site-specific systems of coordinates may be used.

 

After georeferencing, the dense point cloud, orthophoto, and digital elevation models are automatically generated.

 

When a site-specific system of coordinates is used, the matrix of transformation is created in order to relate the site-specific coordinates to global ones. This process will ensure accurate linear and volumetric measurements within the resulting orthophoto and digital elevation models. The map itself will be located on the global map in the approximate location of the job site. Running measurements between the points on the drone map and the underlying online map should be avoided.

4. Quality Procedures and Scan Publishing

After photogrammetry is completed, the following quality procedures take place:

  • The model is compared to previous scans to see if there were any errors in photogrammetry

  • Volumetric calculations are run as a sanity check to verify site balance hasn’t changed significantly

  • Elevation error of QCPs are calculated

  • Quality report of QCPs errors are distributed via email

  • Flight marker maintenance recommendations are distributed via email

The entire process is completed overnight. The scan is published and ready for use early the morning following the flyover.

Cut_Fill map.png

TraceAir Bulk/Shrink Calculation

Accurate bulk/shrink estimates are crucial for a reliable balance prediction. Usually, estimates are provided by geotechnical engineers and derived from either soils tests or observations of dirt behavior on sites nearby. Lab tests imply projecting data from limited sample size on the whole site, and resulting coefficients often differ from real soil behavior during grading. Historical data from other sites may not be applicable to a particular site.

 

TraceAir suggests improving the procedure by measuring the real bulk/shrink coefficient of the dirt moved on site from the very beginning of the project to any given flyover date. For that, four components (models) should be taken into account (also see the screenshot below):

  1. Original Ground surface (OG)

  2. Recent flight surface

  3. Up-to-date Design Grade surface (DG)

  4. As-built Remedial Surface (from sub-ex bottom shots)

0 Bulk-Shrink Drawing 2.png

To come up with an average Bulk/Shrink coefficient of the soil moved, one needs to calculate the Total Cut and Total Fill completed to date. The flight surface and the remedial bottom shots surface should be made on the same date.

  • Total Cut (Completed) = Raw Cut (Cut Regions) + Remedial Cut (Cut Regions) + Remedial Cut (Fill Regions)

  • Total Fill (Completed) = Raw Fill (Fill Regions) + Remedial Backfill (Fill Regions) + Remedial Backfill (Cut Regions)


 

After the Total Cut and Fill completed by the specific date are calculated, real Bulk/Shrink coefficients may be determined as:

  • Shrink (Bulk) = 1 – Total Fill/Total Cut

 

We use the software script to process four surfaces and run pixel by pixel calculations throughout the whole grading footprint to determine the Total Cut and Fill completed. Then the Bulk/Shrink is calculated. Given that the DG and OG surfaces are available and the remedial bottom shots and the fly-over surfaces can be collected on a regular (weekly) basis, TraceAir solution allows calculating the real Bulk/Shrink values over time and get insights from the trend (see an example below):

Screen Shot 2018-08-23 at 18.31.04.png

TraceAir Balance Prediction

There are a few crucial components that one needs to take into account while calculating the initial site balance prediction (see the drawing below):

  • Original ground topo (OG)

  • Design Grade surface (DG)

  • Suggested (estimated) remedial (corrective) surface

  • Bulk/shrink coefficients (estimates) for different types of grading:​

    • Raw Cut (OG to DG in cut regions),

    • Corrective cut in cut regions (over-ex), and

    • Corrective cut in fill regions (removals)

1 Initial Balance Prediction Drawing.png

Calculations should be pretty straightforward and may look like the following:

 

  • Raw Cut (Remaining) = Raw cut * (1+Bulk/Shrink for Raw Cut)

  • Bulked/Shrank Remedial Volume = Over-ex * (Bulk/Shrink for Over-ex) + Removals* (Bulk/Shrink for Removals)

  • Balance Prediction = Raw Cut + Bulked/Shrank Remedial Volume – Raw Fill

 

The problem with this method is that only one parameter (raw cut/fill numbers) is reliable, provided the OG is accurate. Remedial quantities and all the bulk/shrink coefficients are usually the results of the guesswork, and this downside has been almost impossible to overcome so far since what’s underneath, and dirt behavior are real unknowns.

 

The TraceAir balance prediction method helps to take a big step forward as far as accuracy of these estimates in that it gradually replaces the assumptions with the most recent up-to-date data, both for surfaces and dirt behavior (refer to the drawing depicting the shrink scenario below):

 

  • Substitutes the OG with the flight surface allowing to track the remaining quantities vs. original numbers

  • Complements the proposed remedial model with the as-built model wherever the corrective grading is completed

  • Substitutes the initial bulk/shrink estimates with the adjusted coefficients calculated based on up-to-date dirt behavior.

2 Updated Balance Prediction Drawing (1)

Updated quantities are calculated after every update (drone flight, bottom shots, and bulk/shrink) overnight, usually on a weekly basis. As a result, the balance prediction is being improved as you go, giving you more and more accurate information to fix any issues early on.

Skyline-CS.jpg

Case Study

How Pardee Homes balanced 
a 27MM yd³ jobsite to just 20K yd³
using TraceAir

CS.png

Ленинская Слобода, 19 корпус 1, Москва, Россия

© 2022 TraceAir    ООО «ТрэйсЭйр»

bottom of page