Bay bridge of San Francisco

How to Get 3D Models with Accurate Data Capture using gNext

Whether you’re conducting infrastructure inspections, monitoring a construction project, or surveying the damage assets have endured, one thing holds true: it’s paramount to have the most accurate data possible. Simply put, well-collected data provides for better digital twins and better inspection outcomes.

The gNext Platform is a valuable resource for processing large amounts of data. Being able to visualize data from numerous perspectives is key to grasping the true condition of your asset. However, having more images increases gigapixels and the cost of using the platform, so it’s crucial to find balance. To give more context, let’s break down the methods gNext provides for visualizing critical infrastructure.

The gNext Platform enables clients to visualize their assets in 3D. These 3D models are formed based on photogrammetry. gNext uses image data and our proprietary photogrammetry engine to process and create layers. This enables users to measure their assets, detect areas of interest, identify defects, and more. gNext provides a wide range of model layers, including:

  • Textured mesh: These are versatile model layers that represent complex 3D objects, enhancing their realism with detailed shapes and textures.
  • Point cloud: Like pixels in a photograph, point cloud models consist of data points that define 3D structures. The models are crucial for infrastructure inspection due to their ability to offer hyper-realistic visualizations of asset data with ease and efficiency.
  • Contour and elevation maps: These maps provide clear asset visualization by showing ground variations, slopes, and terrain features like hills, valleys, and rivers without cluttering surrounding elements.
  • Terrain: Terrain is a topographic representation of the ground surface, displaying elevation data relative to a grid and digitally excluding surrounding vegetation and assets to emphasize the underlying terrain.
  • 2D Orthophoto: This precise photographic map eliminates distortions caused by factors like camera tilt and terrain dimensions, providing an accurate representation of infrastructure.

The Importance of Data Capture for Asset Owners

Good data capture enables clients to stay ahead of the curve and quickly address any issues or defects. In the last several years, the infrastructure inspection industry has been using drone technology and image sensors to collect asset data.

Drones can navigate complex and hazardous elements in ways human inspectors cannot. Highly tuned sensors can be used to take images of infrastructure assets and processed using photogrammetry to produce actionable digital twins. These types of technologies allow users to capture assets to the full extent and other essential data that could’ve been missed otherwise.

When asset data isn’t captured properly, not only will you not get an accurate or complete model, but you’re left having to put in extra work to tie up loose ends. This is precious time and money lost on additional labor.

Insufficient Data Capture Example:

Asset types with thin structural elements, like water and cell towers, are notoriously difficult to capture. In the case below, a drone was flown in between the thin elements, causing incomplete data capture.

3d model made from insufficient data capture gnext
The model presents defects in the thin structural elements due to insufficient data capture or overlap.
3d model made from insufficient data capture example with camera positions gnext
Camera positions show the drone flight between the thin structural elements and a low amount of image overlap.

Accurate Data Capture Example:

In the case below, a drone was flown around the tower instead of in between the thin elements. This resulted in accurate data capture.

3d model made from good data capture gnext
The model shows the thin elements processed with high resolution.
3d model made from good data capture example with camera positions gnext
Camera positions show the drone flight as an orbit around the tower, with plenty of images and overlap.

Data Capturing Methods for Accurate Data Models

Because every asset is different, data capture varies based on factors such as the size, location, underlying terrain, and weather conditions. But it’s still possible to capture high-quality asset data despite circumstances.

First and foremost, you’ll want to find the right camera for the job. Camera options fall into two categories: RGB and multispectral. RGB cameras are great for capturing the natural color of elements, while multispectral cameras can map vegetation and land cover.

When photographing infrastructure manually or via vehicle, you’ll want to use a camera with at least a 12 MP resolution (more is always better) and preferably a fixed lens. Zoom lenses are also acceptable if you don’t change the zoom level while taking the photos.

Using drones can eliminate many of the obstacles one may face while manually collecting asset data. Nowadays, most drones can capture the necessary imagery to conduct inspections. But you may want to choose either a multi-rotor or fixed-wing drone depending on constraints like the survey area size, your project goals, and your team’s budget. Some drones are equipped with GPS and may also be fitted with sensors and other tools to capture every element of the infrastructure.

Capturing Asset Data Properly: From Lighting to Sensors

data capture with drone
Industrial Unmanned Drone Survey, Monitoring And Discovery

Lighting: One of the most important steps in creating an accurate data model is to do your best to reduce the shadows surrounding the structure. Overcast days are your best friend when it comes to collecting asset data. But if you’re working indoors, consider using a lightbox.

Resolution: Depending on the type of data you need to collect, you’ll want to adjust your resolution accordingly. Consider using a higher resolution for when you need the highest level of clarity (i.e., detailed inspections) and a larger quantity of photographs. A standard resolution (12 MP or 20 MP) is better suited for capturing larger sites, such as site planning, construction project management, or survey work.

Images: For successful 3D modeling of objects with photogrammetry, overlapping images are key, as the object itself is reconstructed through matching pixel groups across photos. Ultimately, you want images of each face of an object from multiple angles to make your processing as simple as possible.

Data Capture: Because photogrammetry is essential in the inspection and surveying process, you’ll want to shoot your assets strategically. Try moving your camera around the structural elements to capture each perspective of the asset. This will allow you to obtain a broad view of the mapped area. If you use a handheld camera, consider positioning it on a tripod to eliminate any issues with movement or lighting.

For example, extra precautions should be taken for thin elements like water and telecom towers. To ensure an accurate model, the drone should orbit around the tower. If the drone flies in between the thin elements, the data for the model will not be accurate. gNext partner Skydio offers Skydio 3D Scan, which is the preferred method for capturing thin elements.

Camera Sensors: Popular sensors include RGB images, thermal, and LIDAR, depending on the inspection and project requirements. Understanding the desired outcome will help determine the best sensor for the job.

Some of the top drones equipped with image sensor payloads include Skydio, Wingtra, and WISPR Systems.

Skydio

Skydio is the leading U.S. drone manufacturer and world leader in autonomous flight. Their drone technology leverages breakthrough AI to make capturing quality data a breeze. They’re a longtime partner of gNext and countless other organizations.

Wingtra

Wingtra created an easy-to-use fixed-wing drone that delivers consistent survey data faster and at a lower cost than many other surveying tools. The WingtraOne can fly higher than some other popular drones, which allows you to capture more ground and more detail.

WISPR Systems

WISPR Systems is a U.S.-based company that builds drones for rugged jobs like surveying, mapping, construction, and more. Their drones offer advanced navigation tools and real-time data analytics, all while being user-friendly.

Take Inspections to the Next Level

gNext is here to help. The gNext Platform is a cloud-based tool that combines artificial intelligence and 3D modeling to enable inspectors to analyze the integrity of assets safely, accurately, and collaboratively from a remote location. Our platform can automatically process data with 40+ pre-defined asset types, saving you time and money. gNext is user-friendly and designed to empower its users, so no manual tie points, photogrammetry skills, or other experience is required to use the platform.

Request a demo of our platform to see what we can do.

Follow us on LinkedIn for more updates!

Similar Posts