Bentley’s ContextCapture: A powerful Enterprise tool

We’ve been interested in Bentley’s software solutions for Enterprise drone operators for some time. So we thought it would be great to sit down virtually and learn more about what the company’s products do. We did that. And now we’re here to share some of what we’ve learned.

If you’ve been seeing some of those wild digital twins produced by Skydio’s 3D scan software, it could well be the incredible job of stitching together all of that data into those stellar digital twins was done by Bentley software. It gets the job done and does so exceptionally well. There are other photogrammetry engines out there that Skydio has partnered with, including DroneDeploy and RealityCapture, but Bentley is also one of Skydio’s preferred partners for carrying out this task.

Let’s take a closer look.

A huge variety of products

When we had our briefing with Bentley, we focussed largely on how its software integrates with drones – specifically when it comes to Enterprise tasks. But we realized, after taking a deeper dive on the company’s website, that it has created a huge number of industry-specific software products. For example, MAXSURF allows you to create and test hull designs for ships or assess the stability of existing vessels. AutoPIPE helps to design industry-compliant pipe layouts for projects up to and including nuclear facilities. The MineCycle offerings help with the design and planning of mines, including maximizing the efficiency of material handling.

And this, literally, is just scratching the surface. The company has a huge number of highly specialized industrial software products.

But we’re here for the drones.

Data, refined

There’s been tremendous growth in the use of drones for commercial purposes in recent years. This is especially true on the Enterprise side of things. Drones are being used for critical infrastructure inspection, mapping, surveying, and much, much more. As the use-case scenarios for drones have grown, so too have the software offerings that allow you to stitch, organize, and interpret your data.

And that’s where ContextCapture comes in, explains Mark Barkasi, Senior Application Engineer with Bentley Systems. Barkasi has been with the company since 2005, so he’s seen the growth of drones since the very early days.

The software, ContextCapture is kind of agnostic. It can work with million-dollar camera systems or the camera that’s in your phone. And it will use the same photogrammetric process and computer vision regardless of which camera you’re working with.

Does that mean you could, for example, capture some images from the air and other shots with different devices from the ground and integrate them? Yes. And the projects can be huge. Says Barkasi: “We can produce datasets with hundreds of thousands of images with ContextCapture.”

The big picture(s)

Imagine that you’re running a major construction project. Or doing mapping, surveys, volumetric calculations, or simply want to keep track of changes of anything you’re working on over time. In the old days (and still current days, for companies that aren’t using drones), this is a hugely laborious task. It involves putting people out in the field, taking measurements and doing calculations manually, and then – likely – multiple documents or spreadsheets that have to somehow be compared over time.

Using a drone, and with software like Bentley’s ContextCapture, is vastly more efficient and saves money. It also, depending on the task, removes people from tasks that might involve personal risk. So, as many of you already know, there’s a huge advantage to deploying drones and software in many situations.

Plus, this software stitches amazing digital twins. While the image below is just a still, if you hit this link, you’ll be able to manipulate the image in 3D and zoom in on areas of interest.

If you think that looks cool, play around with this link

The big picture, the small picture

As with other photogrammetric software, this is a huge advantage for those who want to zero in and inspect particular parts of an asset. You have multiple options for viewing the data, but one of the key advantages of software like this is that you can quickly go from the big picture and zero in on an area of interest. Take a look at this example: A digital twin of a bridge reveals a spot an inspector/engineer wants to examine more closely.

Does that mean poring through hundreds (or, more likely, thousands) of photos to find that single high-res image? Nope. Simply identify the spot you’re interested in, and the software will pull up the associated photos for a closer look.

Images courtesy Bentley ContextCapture

As Mark Barkasi explains: “Now we’re looking at the individual image. The model in and of itself is an excellent form, but it may lose a bit of detail. But if we look at the actual image, we may pick up things that we didn’t previously see or aren’t as apparent on the model.”

And the key here is that it’s a very simple task to find the individual photo associated with the area of interest.

Machine Vision, AI

If it’s a big project, and you’re looking for changes over time or wanting to identify anomalies (such as cracks in concrete), that can still be a big job for a human being with two eyeballs and not enough caffeine. Here, too, Bentley can assist. The software can be trained to look for the specific defects you want to quickly identify. Think of the time that can be saved. Again, here’s Barkasi:

Any type of AI that we run is going to be currently be run in ContextCapture on the images themselves so we can pick up defects, if we train detectors to look for those defects…In the case of the dam that you saw (an image shown during the briefing), we can perhaps use the AI to look for cracks.

Again, as we’ve seen elsewhere in the industry, the trend is toward using Machine Vision, Machine Learning, and AI to efficiently identify features that might take a huge amount of time — reducing the risk a person might miss the anomaly entirely. After all, we’re only human.

Huge tasks, simplified

So picture this: Instead of having people trying to capture all of this manually, a drone (even in conjunction with a ground-based camera) can recreate a digital twin of a real-world structure. If there are changes over time, those changes can be detected. This applies to everything from the diminishing volume of a pile of gravel or ore through to new anomalies that might need a closer look.

This is powerful software…and it keeps getting smarter

And the software is not limited to simply photos. You can add survey data, LiDAR (Light Detecting and Ranging, which uses lasers), and more to create a model that can be picked apart on a highly granular level. Here’s Bentley’s Mark Barkasi again:

Essentially the models are either going to be a three-dimensional mesh or a point cloud, or a Digital Surface Model or DSM. We have many different outputs; outputs that work in ESRI software, CAD, in GIS or web viewers.

And yes, there are tools available within ContextCapture so that you can easily extract accurate measurements (and more) from these models. In addition, they can easily be shared with multiple team members. As Barkasi points out: “In the cloud, team members can collaborate.”

In its own words

This is really just scratching the surface (and if a surface was scratched, ContextCapture could probably pick it up). But in closing, we’ll take a paragraph from Bentley’s website:

With ContextCapture, you can quickly produce even the most challenging 3D models of existing conditions for infrastructure projects of all types, derived from simple photographs and/or point clouds. Without the need for expensive, specialized equipment, you can quickly create and use these highly detailed, 3D reality meshes to provide precise real-world context for design, construction, and operations decisions for use throughout the lifecycle of a project…Develop precise reality meshes affordably with less investment of time and resources in specialized acquisition devices and associated training. You can easily produce 3D models using up to 300 gigapixels of photos taken with an ordinary camera and/or 3 billion points from a laser scanner, resulting in fine details, sharp edges, and geometric accuracy. Dramatically reduce processing time with the ability to run two ContextCapture instances in parallel on a single project.

DroneDJ’s Take

There are a number of photogrammetric engines out there, and as the use of drones continues to grow exponentially in the Enterprise world, there’s increasing demand for such tools. Bentley’s ContextCapture is one of the leaders from a company with a huge background in industrial software. There’s also much more that we haven’t been able to touch on here, even with 1,400+ words. If you’re an Enterprise company, ContextCapture and some of its other offerings are definitely worth a look.

FTC: DroneDJ is reader supported, we may earn income on affiliate links

Subscribe to DroneDJ on YouTube for exclusive videos

Load more...
Show More Comments