Something

Best practice tips for successful drone mapping

Best practice tips for successful drone mapping

Using drones for mapping had become common in all aspects.

Creating an orthophoto image, a digital elevation model or a 3D model takes minutes.
But, producing a good quality mapping product is not straightforward as you might think…

Utilizing easy-to-use free flight planning applications (such as DroneDeploy , PIX4Dcapture, etc)

Afterwards, building orthophotos and models with proprietary photogrammetry software (such as AGI methashape, Pix4d, or a lot of others).
Whether installed or as a paid software as a service.
I am mostly using the excellent open source software: ODM (Open drone map).

All of the above had democratized the process of creating fast and self-made mapping products. Utilizing different use cases for a lot of users profiles.

Producing a good quality mapping product is not straightforward as mentioned.

We had worked with a few of our clients aiding them on the process of creating a good quality mapping products.
We mostly use the open source “ODM” (Open drone map or WebODM) software.
After a lot of learning, trial and error I would like to share some tips and best practices.

This blog is focusing on best practices for how to plan the optimal flight for creating quality mapping products.
We will focus on what to do and not to do, and common mistakes.

Of course, there is also optimization for the photogrammetry software, especially ODM, but this will be a subject for a special future blog.

drone images sample

A lot of drone mapping users, whether they are flying the drone or hiring one for the job, are not fully aware, or inexperienced with the special requirements that the photogrammetry process needs to actually do the magic and get a good mapping product.

 

We saw a lot of users that were disappointed with photogrammetry software results. Those users were trying to change the software parameters endlessly just to find out the results are the same .
(and that the process takes time and money to run, each time).
While the key is as in any other image processing software: “Garbage in, garbage out”.

It means, that if you supply the software images with not enough light or other artifacts that you even didn’t know about them, the product will not look good, or even in some cases, the software will fail to build it.

This will cause costly and timely repeating field days. Making you repeating flights over and over, running the software again.
Furthermore, if you are supplying the product to a customer, he will be very unsatisfied.

Those are the tips we have gathered for planning drone flights that will produce the best quality mapping product:

This list is intended to use whether you fly the drone or want to hire a pilot and instruct him*:

  1. Sunlight: It is a crucial factor for quality. Try to keep good light conditions, full sunny or full cloudy, not a mix. Partial clouds which make the images parts shadow parts sun is not good for the quality.

    HalfClouds Mix Cloudy and sunny

  2. Flight hours: The sunlight should be with minimum shadows so try to fly not in the morning or the afternoon. In those hours, the sun makes long shadows for objects and confuse the photogrammetry matching algorithm.

    Hours recommended are 10:00 AM – 2:00 PM (varies and depends on season and location of course) the less shadow you have, the better results will be.AfterNoon After-Noon image-shadowy and soft light

  3. Flight height: Need to be balanced with the local low of flight, obstacles, the resolution required, the area and its nature, topography.
     (i.e, if it’s very different such as a city or very much looks alike like a regularly looked agriculture field) and the area

    If the light is not good and the area is very look alike like a field, the software cant find matching points between images. To the software, it all looks the same, you need to fly lower, so more details will appear and the software will find them and match images.

    40-50 meters height is a good choice for a start.
    Keep in mind that if you have high objects in the area, such as buildings, You might have to plan higher height, in order to get good overlap for them, too.
    Plan 3-4 times the height of the tallest object if you want it to look well in the output.

  4. Flight speed: The speed of the flight might also impact the quality. It depends on the drone type and model, the topography, and the resolution wanted. Flying too fast at a high altitude might blur the images. While flying too slow might also make images blurry. If the quality of the images is not good you might want to try on high/less speed, it varies to the specific case.

  5. Flight plan: Overlap is the most impacting factor. consider  80/80  or even more overlap if the results or conditions are not good, especially for agriculture/trees area.
    Try to minimize moving objects. (cars, persons, water) change to quiet days/hours when the traffic of objects will be minimal.

  6. Flight method: flight planner apps provide several options for relevant mapping. You should select the relevant plan for the output you need, so you will not do non-relevant flights and too many images if you just want orthophoto, while don’t expect to get a 3D model if you fly only one grid.

    The rule of thumb is if you want just orthophoto to choose a one-way grid.

    if you want an orthophoto and 3D model and the area is without specific objects such as a building choose a double grid.

    Add two circles plans in different heights with overlap and camera angle (45/60)  around a specific object you need to model. You need much more detail (circles added to a grid will usually make a model better, but it is unnecessary just for orthophoto)

    Creating a circle only around an object such as a building without a one-way grid will not be enough for orthophoto, just for the model.

    In the case of a 3D model with vertical walls or obstacles that block the drone camera, you can use a good camera or a phone camera if you don’t have one. Make overlaps (manually or with an app) with GPS on.
    Adding the camera images to the drone images will upgrade the results of the model.

  7. Flight plan: When you sketch the flight area of interest, enlarge it, so the software will have more points outside the area to help it create good quality even in the borders of the area. 
    Keep the same: Drone (camera type), height, speed and resolution all over the plan and flight.
  8. Pattern: If the area has a pattern of straight lines such as roads, or fields with rows, plan the flight path parallel to the line pattern, not diagonal.
    Flying above shiny/reflective objects is a challenge (water, solar panels) try to avoid it, or fly in non-sunny day. 

  9. Post-flight QC: After getting the images from the drone, we recommend those steps and workflow:

    • Open the folder in a file system explorer which can show you medium previews of the images. Open a few images to see if they are sharp or blurry, shadowy, etc. If all is blurry don’t expect magic in the output.

    • Scroll fast and stop if you see blurry, non-focused, damaged images or one that stands out as different. Delete or cut those to another folder.

      It takes 2 minutes which can save you time and frustration for bad results afterwards.

scroll images
Scrolling drone images

10. Flight and images location QC: I will also recommend viewing the location and path of the raw image to check for inconsistency/islands and out-of-area images.

If the software doesn’t have this feature, You can use a GIS desktop software such as Qgis or Arcgis or Qgis or Javascript EXIF library.
If you see out-of-area images, delete them or move them before running the photogrammetry process.

Checking the geotaged images location, for issues Checking the geotag images location, for issues (see points on the left)

11. Rare effects: Sometimes even when you fly with good lights and hours, the images still look blurry and the output is bad. If you check the images and there is a “glow” or light in the middle, this phenomenon is called Cellini’s halo or Heiligenschein (halo in German) and it is named after the famous Italian artist: Benvenuto Cellini.

Commonly called hotspot.
Source: Wikipedia

You can enhance reading in the link above, but to keep it short concerning drones:

It is happening when the sun is behind/above the drone camera (typically afternoon or early morning) and the image includes the drone shadow on it, creating this “glow” around the shadow (which is unseen on the image due to its small size, opposed to the Baloon above)

Orthphoto result from hotspot effect in the original drone images Orthphoto result from hotspot effect in the original drone images

Making the ortho result, bad:

(Example for orthophoto  with bad quality from the internet. the effect can be seen especially on the upper areas )

To prevent it, keep the recommendations as before: best hours to fly are when the sun is high in the sky usually 10:00-14:00. Try to plan the flight path directions so the sun won’t be behind the camera and will create drone shadow on the images

Happy mapping!

* The list not covering all and might not fit any case. partial use it with common sense and mostly keep security and rules at first priority.

If you need help, we will be happy to assist…🙂

——————————————-
I’m Ran Tzkhori, a Geo-Spatial data and Spatial systems expert.

We help organizations find fast answers with the aid of spatial data and spatial SQL in their existing or new systems in house or in the cloud.

If you need advice, don’t hesitate to reach me.⬆️

Share Post

Facebook
Twitter
Pinterest
LinkedIn
WhatsApp
Telegram

Leave a Reply

Your email address will not be published. Required fields are marked *

מעבר להודעה
1
Hi There,
I am Ran, Mikoom’s Founder.
I will be happy to answer any question.
Just click on the button below to reach me.
Skip to content