Eyes on the sky for a groundbreaking safety and efficiency solution

Looking skyward is solving safety and efficiency challenges for geotechnical engineers, with drone photogrammetry helping to photograph, generate 3D models, measure, and analyse site terrain faster than ever, informing critical decision-making while keeping personnel out of harm’s way. Senior Engineering Geologist Adam Kerr explains why this arguably underused tool is so powerful for data acquisition, and why the process should be rolled out as best practice across the industry.

The sounds of a site radio, haul truck horn or overburden blast alert are all part of aural safety measures in motion on mine sites. And for geotechnical engineers working to optimise a mine’s safety and efficiency, the whirr of an Unmanned Aircraft Vehicle (UAV) – or drone – should join this chorus, in greater numbers, too.

Drone photogrammetry, the process of launching a UAV to capture images and metadata in the field to eventually emerge with a 3D model of the terrain, may not be a new concept in our field. But through rapid advances in technology and more budget-friendly price points, it’s a tool more accessible than ever, bringing value in efficiency and speed to plan delivery, as well as safety of those on the ground.

Data collection and mapping – we’ve come a long way

Structural mapping has historically, and in some places still is, done terrestrially, and often at great risk to the field operator.

In its early form, it has involved physically scaling walls, going right up to the surface and taking measurements with compasses and inclinometers, at a considerable safety risk.

As technology progressed, land-based surveyors would set up scanners, take a single high-resolution photo from location. Then they’d move all their gear somewhere else, set up, and do it all again. Rinse and repeat for many hours in the field, still with risk to personal safety while trying to access hard-to-reach locations.

The advance of drone technology means this time-consuming, risky data capture is a thing of the past. Now, we spend an average of 15 minutes flight/photography time, to capture images and vital corresponding metadata including approximate location co-ordinates, focal length/aperture, direction, angle, azimuth, and more. This metadata in each photo is used by photogrammetry software to stitch images together into a superior 3D model in considerably less time, with substantially less personal risk. As technology has progressed, the algorithms of photogrammetry software have improved considerably, and have become far more user friendly.

How it works – drone photogrammetry in action

The whole process can be understood through the lens of a highwall photogrammetry case study, which we recently undertook at an open-cut coal mine.

There are three main steps; data capture, data processing, and data interpretation:

1.     Data capture (via drone):

• Hardware used: DJI Mavic 3 quadcopter

• 50 photos taken along a 400m long by 50m high section of highwall

• Photos all taken manually, at various heights, angles, and distances from the wall

• Data capture (including travel from office to pit) less than 1 hour

• Metadata captured in each photo used by photogrammetry software to stitch images together into a model

2.     Data processing (via software and assessing scan detail):

• Data processing: Photogrammetry software: 3DF Zephyr

o   The first ‘layer’ involves a point cloud image, the foundation for the rest of the modelling.

o   This is followed by textured mesh.

• Data processing: Assess detail. Detail of the photogrammetry stitched image is checked for clarity and correctness.

• Data processing: Register to Survey

o   The scanned model is then translated to correct position and orientation

o   Register the textured mesh to the latest pit survey

o   Position accuracy – within 1m spatial accuracy between photogrammetry model and pit survey

o   Orientation accuracy –  +/- 1°

3.     Data interpretation:

The final step is to interpret defects in the wall, including joints, faults, and bedding planes.

We then export defect attributes, including dip, dip direction, length, and positions to understand persistence and spacing.

Defects can be used to assess kinematic risk of failure, such as planar, wedge, or toppling style failures.

We then determine operational controls and to inform future design.

The resulting 3D model:

Why drone photogrammetry should be rolled out far and wide

The cost of commercial drones – upwards of $10,000 – has often been a prohibiting factor. In today’s market, however, and for geotechnical applications, retail UAVs are more than up to the task, at a fraction of the price.

While commercial UAV features – including ultra high-resolution cameras, advanced obstacle avoidance systems and precision sensors, with real-time kinematic processing (RTK) – all help deliver the millimetre accuracy needed in civil engineering, for example, on a mine site work is at a much larger scale.

Retail UAVs offer powerful capabilities and features to capture all required images and metadata at a friendlier entry price, without compromising quality where it is needed.

Most retail models still have excellent safety features including crash avoidance, which means you won’t fly into a wall. If you lose signal with a controller, the UAV will remember exactly where it took off from, fly back and land at exactly that spot if control is lost.

Environmental considerations – such as wind and weather – can also be made through the usual checklists of mine site ‘take fives’, following standard operating procedures, and other site safety rules.

An often-overlooked check – for birds of prey – should also be performed. Eagles, commonly found across mine sites, can be territorial and we’ve witnessed them taking down little quadcopters before.

For decision makers on site, the use of UAVs can optimise mine design through photogrammetry and also offers benefits in just being able to see what’s going on from a different perspective.

Simply flying a drone up to take photos to see a crack not visible from the ground, for example, provides information we wouldn’t have otherwise, giving confidence that we know what is really going on and can act quickly to control any risk.

Drone photogrammetry also helps to create permanent records of sites, allowing for better monitoring and future comparisons.

Avoiding barriers to implementation

While it is now easier to access capable UAV technology, the proliferation of drones in society is becoming more and more regulated.

With so many UAVs in the air, there is a justifiable need to ensure rules are followed – across mine sites and within a broader context. This means it can be more difficult to obtain the required remote pilot’s licence, and become certified through the Civil Aviation Safety Authority (CASA), to be able to fly a UAV for work. In addition, each site usually has its own rules to abide by when operating a UAV.

Navigating these governance and legality factors can be the most difficult part, but engaging an accredited, experienced operator ensures a smoother process and compliance with a mine site’s own flying rules and CASA regulations, too.

In conclusion

With such ready availability and powerful capabilities for data acquisition and decision-making, implementing UAV photogrammetry into geotechnical work is best practice the industry would be wise to embrace.

It’s a powerful tool that minimises risks by allowing geotechnical engineers to assess potentially hazardous areas remotely, while saving time and money compared to traditional, land-based methods. Helping sites reach production goals faster – and safer – on current projects and beyond is a flying start, indeed.