Author Archives: admin

Workflow of a roof inspection using a drone (DJI M3E/M3T)

The increasing number of commercial buildings in Switzerland is driving up demand for safe and effective roof inspections. Each roof has unique requirements, and various areas may need inspection, such as leaks, holes, and components that require regular checks.

Drones have transformed the way roof inspections are conducted over the past decade. Their use allows for efficient data collection without the need for ladders or entering the attic. The information gathered can then be easily shared and processed.

In this article, we will look at the process of a roof inspection using a drone (DJI M3E/M3T)

Table of Contents

collect data

Data collection: At the start of the inspection, relevant information about the building is collected. This includes data such as the size, shape, and pitch of the roof.

1. Property inspection: The roof is visually inspected for potential damage and defects. Drone footage can be used to conduct a detailed assessment.
2. Defining the inspection area: The area to be inspected is defined. Specific areas such as solar panels or ventilation systems can be the focus.
3. Checking sensor settings: The drone's sensor settings are checked and adjusted to ensure the best possible data acquisition.
4. Flight planning: A flight plan is created to ensure optimal coverage of the inspection area.
5. Data collection: The drone is launched and collects data through visual and thermal imaging.
6. Manual inspection: A manual inspection can be carried out to examine specific areas more closely.

Data Processing:

Data processing: The collected data is processed to enable an accurate analysis of the inspection data.

1.  Thermal and visual datasets: The thermal and visual datasets are combined to create a detailed inspection analysis.
2.  Ground control points/checkpoints: Ground control points are used to ensure more accurate referencing of inspection data.
3. DJI Terra settings: DJI Terra settings can be used to perform effective data evaluation and analysis.

Collect data and define the inspection area for drone inspection of roofs

To successfully conduct a drone inspection of a roof, certain aspects must be considered. First, the size of the roof must be taken into account to understand the scope of the project and plan accordingly. Smaller roofs can be inspected in a few minutes or even seconds, while larger commercial roofs may require longer flight times.

The building's height is another important factor to consider when planning the mission. A short flight time to the top of the building allows for better operational planning.

Define inspection area

To define the objective of the inspection, it is important to know what objects are located on the roof. The type of inspection and the associated requirements for accuracy and resolution can vary depending on the objective.

The most important use cases for roof inspections are

• Detection of cracks and leaks
• Air conditioning inspections
• Inspections of photovoltaic systems
• Chimney/exhaust gas inspections
• Requirements for the measurement or desired results
• Roof inspection

When thermal imaging sensors are used to inspect, for example, solar panels, leaks, or air conditioning systems, the flight should be conducted immediately after sunset to avoid thermal stress from direct sunlight. When searching for leaks, at least 24 hours should be allowed to pass after a rain shower to understand the drainage/leakage. The size of the building should be taken into account when planning the flight route to avoid unnecessary flight time.

Data accuracy requirements are another important aspect to consider when planning drone inspections. With the Mavic 3 Enterprise With the RTK module, centimeter-accurate data can be generated without the need for ground control points. This data can also be compared with other construction site data to achieve greater accuracy. RTK, PPK, and cloud PPK technologies can help achieve a high level of accuracy.

Checking camera settings for a roof inspection

To ensure safe flight conditions, the sensor settings must also be checked. Several factors must be considered when choosing camera/sensor settings, such as exposure time, ISO sensitivity, and color correction. By selecting these settings correctly, clear and accurate images can be captured, contributing to an effective roof inspection.

Recommended settings

• shutter speed A shutter speed of 1/1000 or faster is recommended for daytime flights. Motion blur plays a significant role during night flights. Therefore, try to set the shutter speed as fast as possible without losing sight of the roof.
• Take advantage of the ISO valueto compensate for the shutter speed. During the day it is best to set the ISO value to Auto, but for night flights you can use it to "brighten" the image if you need to use shorter shutter speeds.
• Image format: JPG
• Image format: 4:3
• Mechanical trigger: EIN
• Sensors to be detected (in thermal imaging): ALL

For thermal inspections, we generally recommend setting the color palette to iron red, as there are large color differences at different temperatures within the camera's field of view.

We also recommend taking the time to do a short flyover of the roof at the beginning. This can help you find the best camera settings before the flight. A roof can often be much brighter than you think, and if you manually adjust the camera settings at the first waypoint, the images are often "washed out".

Planning a flight for roof inspection with a drone

The most common method for inspecting a roof is to collect enough overlapping photos to create a high-resolution map and a 3D model of the roof. This can be done with the DJI Pilot 2 App will be performed if you Mavic 3 Enterprise Use Series drone.

When planning inspections, it is best to select the Mapping Mission option. Here you will find a guide to help you get started with Mapping Missions.

And here are some settings we specifically recommend for roof inspections:

• Use the default overlap settings of 70% and 80% front overlap. This should be sufficient for high-quality 3D model reconstruction for the visual sensor.
• If thermal imaging is required, we recommend 80% sidelap and frontlap.
• When selecting the altitude, you should use the two sliders: flight route altitude and target area to takeoff point. The optimal flight altitude over a residential roof is 5-15 meters above the roofline. This resolution may not be achievable for larger commercial buildings, so planning for 15-30 meters above the roof should be sufficient. A short flight to verify the building height will help you determine the appropriate mission altitude. For example, if you are checking the height of a residential roof and it is 8 meters high, set the target area to 8 meters and the mission altitude to 15-25 meters. If you are checking a commercial roof that is 15 meters high, set the target area to 15 meters and the flight route altitude to 30-45 meters.
• The Target Surface to Takeoff Point slider allows you to achieve the correct overlap settings even when the drone is launched from the ground. With the 4/3″ sensor of the Mavic 3 Enterprise You can capture incredible detail with a wide dynamic range.
• Here are some GSD estimates using the M3E ( DJI Mavic 3 Enterprise ):
  • 7,50 meters 0,2 cm/pixel
  • 15,00 meters 0,4 cm/pixel
  • 22,50 meters 0,6 cm/pixel
  • 30,00 meters 0,8 cm/pixel
• Here are some GSD estimates using M3T ( DJI Mavic 3 Thermal ):
  • 7,50 meters, 0,26 cm/pixel visual, 1 cm/pixel thermal
  • 15,00 meters, 0,53 cm/pixel visual, 1,98 cm/pixel thermal
  • 22,50 meters, 0,78 cm/pixel visual, 2,97 cm/pixel thermal
  • 30,00 meters, 1,05 cm/pixel visual, 3,96 cm/pixel thermal

• The software offers the following for creating a 3D reconstruction: Mavic 3 EnterpriseThe series offers the Smart Oblique function, which allows for taking oblique images during flight, instead of just taking NADIR shots. This allows the Gimbal It can be controlled during flight. However, Smart Oblique should not be used for solar inspections on roofs with thermal updrafts, as this can affect the accuracy of temperature measurements.
• Flight direction and speed are other important factors that should be considered. Mavic 3 Enterprise It is equipped with a mechanical 4/3″ shutter that enables fast shooting with minimal image distortion. The flight speed is suitable for the Mavic 3 Enterprise Not so important, but if the goal is thermal inspection with the M3T, the maximum speed should be below 4,4 m/s (10 mph) to avoid image blurring and incorrect image measurements from the thermal sensor.
  • If you are planning the flight direction and only want to capture visual images, you should fly in the most efficient direction. When inspecting thermal solar panels on roofs, it is recommended to fly parallel to the panels to achieve the best results in data processing.

Data collection

When you are ready to capture the location, you should ensure that you can maintain a line of sight with your drone, which can be difficult when capturing building roofs.

Make sure to keep a close eye on the drone's flight plan and the FPV camera to ensure you do not fly over people.

Manual verification and inspection of the data

After the automated mission is complete, you can manually collect additional data to get the most out of your inspection. Mavic 3 Enterprise and the Mavic Both 3 Thermal devices feature a 56x hybrid telephoto zoom sensor, allowing you to better understand the target during manual inspection. To keep the zoom and thermal imaging sensors at the same zoom level, we recommend using the Link Zoom function (side-by-side view).

Data Processing:

The processing of thermal and visual datasets begins after site acquisition and includes conversion into a high-quality 2D orthomosaic and 3D model. The use of DJI Terra enables a simple and effective process for generating high-quality datasets. The following step-by-step process shows how to use data with DJI EARTH can be processed:

1. Importing photos/folders into DJI Earth.
2. It is recommended to process visual and thermal data sets separately.
3. Selection of desired output types (2D map, 3D model) and file extensions (Tiff, Obj, etc.) as well as definition of the coordinate system [Switzerland: LV95]
   (if an NTRIP service is used).
4. Start of the aerotriangulation.
5. Optionally, the reconstruction limits can be changed to optimize processing time and output data size.
6. Optional step: Import of ground control point data and selection of the appropriate EPSG code for the region (Switzerland: EPSG Code: 21781).
7. Execution of the steps for reconstructing the 2D map and the 3D model.

Please notice that DJI Terra does not guarantee a radiometrically composited output, only the raw images. Once processing is complete, the accuracy report can be used to verify map accuracy. The data is now ready to view and export. We recommend that you DJI Try Terra with a 1-month trial version, which is based on the DJI The Terra website is available.

Show data

DJI Terra offers various functions for data visualization and analysis. Our annotation tools allow you to measure cracks and leaks, while navigation within the 3D model is controlled via the mouse. A tool is available for long-term viewing, allowing you to orbit the 3D model indefinitely.

For example, in roof inspections, 2D orthomosaics (orthophotos) are often used to analyze leaks, cracks, and thermal irregularities, rather than 3D models. While 3D models facilitate viewing the site from a specific perspective, raw images are often analyzed for thermal inspections instead of the 3D model. For clients requesting a dataset, [the company] offers DJI Terra supported outputs that are georeferenced and can be imported into a third-party analytics tool such as DroneDeploy or PixPro can be imported.

DJI It also has a thermal analysis tool that can be used to examine raw images and processed datasets to gain a complete understanding of temperature measurements. There is also a publicly available tool from Eric Olsen, which allows thermal data to be converted into RJPG for import into Flir's thermal analysis tools.

Third-party analytics providers

For automating inspection analyses, there are many specialized third-party solutions. Providers like DroneDeploy offer cloud-based tools for various industries such as construction, agriculture, oil and gas, and solar. For roof inspections, DroneDeploy offers specialized tools and reports, such as the Roof Report, which determines roof dimensions based on a processed 3D model, but does not offer automatic damage detection.

DroneDeploy offers a radiometric thermal analysis tool that assists in identifying problems within a heat map. The histogram on the left can be used to change the temperature range. Additionally, there is a tool for comparing multiple flight data points to identify differences.

If the focus is on damage detection, then Loveland Innovations and Eagleview Good options for automated damage detection. Their tools can detect not only hairline cracks, but also small holes and indentations caused by hail or tree damage. To analyze the data, the web tool IMGING from LoveLand Innovation or the sample report page from Eagleview are recommended.

For thermal imaging analysis of solar modules Raptor Maps Raptor Maps' tools are well-known. With over 50 GW of solar modules analyzed to date, they are the first choice for solar module analysis. A screenshot from the tool illustrates the solar module inspection process.

Final

In conclusion, it is recommended to use drones for roof inspections, as this can save time, personnel and inspection quality; above all, you have significantly higher occupational safety than before.

If you have any questions about the use of drones in inspection procedures, please feel free to contact us. us .

Links to the mentioned DJI Enterprise Products:

• DJI Mavic 3 Thermal (M3T)
• DJI Mavic 3 Enterprise (M3E)