Now that you understand the reasons behind using photogrammetry to create 3D models of indoor spaces and how to gather suitable data with Elios 2 for this purpose, you’re prepared to move forward with processing those data to construct your first 3D model. In this post, we’ll walk you through how to use Inspector to prepare your dataset for processing with Pix4Dmapper or any other photogrammetry software. Following that, we’ll guide you step-by-step on how to create your first 3D model using Pix4Dmapper.
To follow along with this guide, you’ll need an Elios 2 dataset, Inspector, and Pix4Dmapper. If you haven't got the necessary equipment or software yet, here's where you can find everything:
- **Download an Elios 2 dataset**
- **Download Inspector**
- **Obtain a trial version of Pix4Dmapper**
---
### Preparing Data with Inspector
The initial step involves extracting video frames and saving them as individual images. Open your video file in Inspector and utilize the "Export frames as images" tool located under the "Export" menu.
This tool lets you select the start and end points for frame extraction and the frequency at which frames will be pulled out. Given that the video records at 30 frames per second (fps), selecting "one image every 30 frames" will yield one image per second, which is usually a good starting point. However, if you were flying at high speeds or if the images lack sufficient visual features, you might opt for "one image every 15 frames," increasing the number of images. Just remember, more images mean longer processing times!
For Pix4D users, we offer two processing templates that automatically set all recommended parameters for you. If you check the relevant box, the template will be saved alongside your images.
- **Fast processing**: Ideal for datasets with plenty of visual features and high overlap. Images are downsized for faster processing, resulting in fewer points in the final point cloud.
- **Robust processing**: Best suited for datasets with fewer visual features or suboptimal trajectories (low overlap). Matching strategies are more resilient but lead to longer processing times. Full-resolution images are retained for denser point clouds.
---
### Processing with Pix4D
#### Camera Model
When creating a new Pix4D project and adding the video frames extracted by Inspector, the software will automatically detect the correct camera model via tags in the image exif data. For 4K videos, this is the Elios2_2.7_3840x2160 model.
If you've extracted images using another tool, they won’t be recognized, and you'll need to manually select the camera model from the dropdown menu. Note that the Elios 2 camera model is available for both 4K and FHD video formats, as well as the 12MP still format.
#### Processing Options Template
You can import a processing options template (.tmpl) following these instructions. To generate the template file in Inspector, just tick the corresponding checkbox in the Export frames window. Once imported, the template is stored in Pix4D and can be reused for future projects.
#### Combining Multiple Flights
You can process images from several flights simultaneously to build larger models and align all data on the same model. Projects containing around 2,000 images (approximately four flights with one image per second) run smoothly in Pix4D. Processing time depends on your hardware and chosen processing options (templates).
Remember, photogrammetry in a GPS-denied environment relies solely on visual data. When merging multiple flights, it’s essential that each flight contains images similar to others, enabling seamless connections between flights. Review the guidelines for acquiring appropriate data with Elios 2 to build 3D models, particularly focusing on loop closure.
---
### Using Other Photogrammetry Software
While Pix4D is a popular choice, other photogrammetry software options are also viable, provided they accept non-geotagged images (images without location data).
**[Related Read: What Is Simultaneous Localization and Mapping (SLAM)?]**
If you're using another software, you may need to select a different camera model with distinct parameters. Many photogrammetry programs let you specify approximate camera parameters, allowing the software to optimize the model during processing. If you proceed this way, ensure you begin with an easy dataset. Recording a video specifically for this purpose can be beneficial:
- Outdoors, with optimal lighting conditions
- Environments rich in visual features
- Multiple flight paths with high overlap
Below are the key characteristics of the Elios 2 camera:
| General | Details |
|--------------------------|-------------------------------------------------------------------------------------------|
| Sensor size | 7.564 mm (H) x 5.476 mm (V) |
| Focal length | 2.71 mm |
| Lens distortion | Fisheye |
| Format | Resolution | Active sensor size | Pixel size |
|----------|------------------|---------------------------|----------------------|
| 4K | 3840 x 2160 | 5.952 mm (H) x 3.348 mm (V) | 1.55 μm (H) x 1.55 μm (V) |
| FHD | 1920 x 1080 | 5.952 mm (H) x 3.348 mm (V) | 3.10 μm (H) x 3.10 μm (V) |
| Photo | 4000 x 3000 | 6.323 mm (H) x 4.743 mm (V) | 1.55 μm (H) x 1.55 μm (V) |
The software may let you fine-tune processing parameters like image resolution or matching strategy. These adjustments can significantly affect processing time and output quality. Refer to your software's documentation for guidance.
---
### Referencing the Model and Taking Measurements
By default, the 3D model lacks proper scale and orientation since the images aren’t geotagged. To scale, orient, and reference the project accurately, you’ll need ground control points (GCPs) and/or scale and orientation constraints. Scaling is necessary for taking measurements, while orientation helps with visualization. Referencing the model in a specific coordinate system enables integration with other models and geodata.
This section assumes Pix4D usage, but similar steps apply to other photogrammetry software. After the first processing phase (camera position and orientation), you receive a quality report showing the number of calibrated images. Other indicators reflect the overall quality.
By examining the 3D view, you should recognize the asset's shape. If too few images are calibrated or the model appears distorted, adjust processing options for more robust settings and restart the first processing step. Alternatively, extract more images (e.g., two per second) and begin anew.
At this point, you can add:
- **Scale constraints**: Use known object lengths visible in images to scale the project. Learn more about scale constraints in Pix4D.
- **Orientation constraints**: If the model isn’t upright, defining a vertical axis can enhance visualization. Identify two points forming a vertical line, such as corners of a room or a vertical pipe. Learn how to set orientation constraints in Pix4D.
- **Ground Control Points (GCPs)**: Georeference the model in a specific coordinate system by adding GCPs. These points must have precise 3D coordinates in your chosen system. Typically measured with GPS or total stations.
Once these elements are added, reoptimize the project—a quick process that recalculates image positions. Proceed to Step 2 (point cloud densification and 3D mesh).
Step 3 (DSM, Orthomosaic, and Index) is only necessary if you require those specific outputs.
---
### Share Your Work
Photogrammetry is a powerful tool for inspectors and surveyors to create digital twins of nearly anything. We’re eager to see what you achieve using these tools. Feel free to share your results by posting screenshots and fly-throughs of your models on social media, tagging them with #elios2photogrammetry, @flyability on LinkedIn and Facebook, or @fly_ability on Twitter. We’ll share your fantastic work with our followers and tag you in return. If interested, send us your favorite works, and we’ll showcase them on our blog and social media channels.
This article is part of a series about photogrammetry:
- Why use photogrammetry to build 3D models of indoor spaces?
- Mining drones: Elios 2 creates photogrammetric models
- Building 3D models with Elios 2: How to acquire appropriate data for photogrammetry
- Building 3D models with Elios 2: Processing data with photogrammetry software
- Elios 2 Tested for Indoor Stockpile Volumetry, Produces 3D Maps Accurate to within 1 Centimeter
Pure Ptfe With Adhesive,Ptfe Film With Adhesive,Pure Ptfe Film With Adhesive,High Temperature Resistance Ptfe Tape
Taixing Yingxing Composite Material Co.,Ltd , https://www.ptfecnyx.com