In this article, we will once again cover a theoretical topic, namely what skills, background knowledge and software use are required on the part of a person conducting a photogrammetry survey in order to produce suitable files.
The main chapters are the usual ones: data collection, processing, display and data sharing.
This article will not be so extensive, nor so practical, but I would like to convey the basic approach of aerial survey. Why is this necessary? Well, according to my experience, the professional and social acceptance of cartographic and 3D surveys is increasing more and more. Much more people try themselves in the profession – as absolute beginners – than 5 years ago. This is largely because UAV manufacturers offer easy-to-use devices at a low price. Furthermore, photogrammetry processing software is also being released in a user-friendly form, so it is easier to handle them than before. In my opinion, it is good that as many people as possible try this profession, because it has a popularizing effect. At the same time, care must be taken to ensure that the mapping and 3D modeling – if there is a business utilization in the background – is carried out with adequate quality, regardless of whether the entrepreneur is at a beginner level or not. Cheapness and low-quality results can easily have the opposite effect in the eyes of customers.
Photogrammetry DATA COLLECTION
Let’s start from the basics! The fate of the final result depends on the data collection, so it is extremely important to plan it carefully, down to the smallest possible details. Never be satisfied with “That’s fine, they’ll pay me anyway!”, be precise and exact in your work, as it depends on whether they order a survey from you again in the future or not. It’s still better to create a file with higher-resolution and quality that you might not need in the end than it would be if you were pushing the limits of what’s acceptable. Good results can be re-examined at any time and can even be simplified, but bad surveys can only be improved by repeating the whole process.
First, I recommend that you get to know the basics of your chosen profession. For example, if you use photogrammetry, study photography and passive remote sensing (our quick start articles are a good starting point for this).
For data collection, it is essential that you are aware of the weather conditions at the expected time of the survey. Get information from several sources to make sure! If precipitation (rain and snow), fog, strong wind (> 30 km/h), cold weather (< 0 °C) is expected at the time you have chosen, then postpone the flight!
The amount of sunlight falling on the area to be recorded is related to the weather. The condition whether you’re surveying a well-lit, shadowy area or parts affected by side lightning make a huge difference. This also depends on the geographical location of the object to be surveyed, as well as the time of day you want to record. To choose the right time, I recommend this simple but very useful calculator.
Pay particular attention to the UAV if you are flying over water or a populated area. In addition, at take-off points above 6,000 meters, check whether this altitude does not negatively affect the device and the batteries. Also, avoid highly electromagnetic areas and tall metal buildings. Make sure that your devices (e.g., controller and drone) are communicating properly.
Permits related to flights may differ from country to country, so I won’t list everything, but be sure to find out what options you have for legally compliant take-off. What kind of permissions can be considered? This also depends on the category of your drone, at what altitude you would fly, but there are also many other factors, but in general, the following will be required: certain registrations, insurances, airspace reservation, certificate of competence, etc. If you are unsure, ask for help or advice from a competent organization or trusted person. In my opinion, it is best to enroll in a pilot training course (or even several!) where all the legal aspects are taught in detail.
Speaking of the pilot training course… There are quite a few training places available today, which is a big positive. If you want to survey with serious intentions, for the purpose of getting business, then I definitely recommend a course, because it is essential that you get to know your device. As a deeper part of the learning process, regularly read professional literature or articles from reliable sources. I recommend that you also take part in professional conferences, because here you can build a network with people with great expertise who you can turn to for advice later on.
The final thing to bear in mind in connection with data collection is verifiability for yourself and others. Write a check list and keep a note! It may seem a little unnecessary, but it is a huge help when you have a detailed tool-and-task checklist that you can go through before each mission so that nothing is left out.
The note is good for knowing which settings and how you used them in a given situation, so if you do something differently later, you can experiment, since you have a starting point.
When I thought about what to highlight on the subject of data processing, the first thing that came to mind was how many different software exist that can be used for different solutions. Choose the target software according to the requirements of a given project! I know that it can be convenient to use familiar software, but you always face new challenges during analytical work, so be open to learning about new procedures and software.
Let’s start with the basics first, the software for raw data processing. When it comes to photogrammetry, cloud-based processing environments are all the rage. They are easy and simple to use, and what is even more promising is that you do not need to purchase an expensive PC, since the calculations are performed by the service provider’s server. BUT! All this comfort – unfortunately, in the vast majority of cases – comes at the expense of quality. Because in the case of cloud-based software, it is not possible to freely modify our model or map at each phase of the processing sequence. I recommend these solutions more for learning purposes, because apart from one or two really surprising results, I have yet to see that a cloud-based processing would have yielded better results than a desktop-based photogrammetry software. I also recommend desktop software because you can monitor the entire sequence of operations and learn the depths of the processing in addition to the control. After a while, you notice that you collect data differently, because you know that one camera position or setting will give better results, while another will give worse. Consistent thinking and seeing the building blocks of the profession in a unified system is a very important quality. If you need to subject the completed 3D model or map to further analysis (this is the usual situation), then you must also choose a special software. I advise you to choose a target software for which a detailed description of functions is provided. This is important because you can understand how each algorithm works and what kind of error it may have, so you will know whether it will be satisfactory from the point of view of the given work.
At this point we have reached another milestone which is the topic of verifiability and repeatability. GIS work requires experts and not people who can click somewhere with the mouse and get a result that they don’t even understand. You have to tell your client or even your colleagues how you managed to achieve the results you got from a project. You must be able to describe in detail what method or software you used, for what purpose and how. Sadly I encountered some reports that mentioned a certain number on the map which related to a certain calculation (for example a volume calculation). Then everyone should have believed that the data were made with the appropriate precision and accuracy. Needless to say that it was everything but a professional attitude. (ezt a mondatot hozzáraktam, hogy egyértelműbb legyen)
Sidenote: A common mistake is to confuse the concepts of precision and accuracy! Precision is the number of significant digits. If we assume that the accuracy is 1 meter, then the Precision of 1.011 meters makes no sense. However, if the accuracy is 1 millimeter, then the Precision of 1 meter degrades the accuracy!
My next tip is to save all partial results during processing! Many times, I ran into the situation that I thought I would no longer need a created data file and deleted it, then it turned out that it was a mistake. Don’t make the same mistake! After closing a project, you will have the time to clean the contents of the folders and store only the final results. Of course, this does not mean that the folder structure of your project should be confusing. I always create a “Temporary” folder, where the partial results of the calculations end up, marked with the names of the executed functions.
Finally, if you have to write a report as the end result of the project, first of all, make sure you have the right structure. I recommend the following division:
- Introduction – The contractual background, data provided by the customer, commitments and deadlines, purpose of work, contact persons and all important technical information should be displayed here.
- Data collection – By definition, the means and methods by which the data were created must be described. Furthermore, with what precision and accuracy can they be characterized.
- Data preparation – In what software environment did you process the collected spatial data, what results were produced, and why did you use the exact format that you did.
- Data processing/evaluation – The essential part, i.e., the purpose of the commissioned work. Describe in detail the software, functions, etc. used here as well. In addition to all this, try to answer these questions clearly: What were the results of the investigation? What can be derived from all this? How can the result affect your client?
- Summary – Concisely summarize the most important information. This chapter can even go to the beginning of the report as an executive summary. This part should be such that your reader can get answers to the stated problems without having to read the entire text.
The wording of the report is important too, try to make it understandable! If it is essential to use technical terms, then explain them, even with a digression of half a sentence. If you have used a measurement result, function, or recommended method from a literature source, please cite it. For example: “The mathematical background of kriging interpolation provides the most accurate estimate, which is extremely important in this case.” This is not some scientific bullshit, but it protects you that you did not choose a particular method without any particular reason, but that it has a research history.
Sidenote: If possible, don’t use Wikipedia references, because it might look discredited! I admit that there is a lot of good and useful content on Wikipedia, but since it is not always verified, it cannot be considered a serious starting point.
I first wrote the title of this chapter as “Data display”, then I realized that this would not be good, since not only the raw files must be provided with properly tasteful style, but also the shape of a report and the intertextual figures must be taken care of.
Prepare the data file in such a way that even a person without GIS background can interpret it. You are the specialist, the customer is interested in the result, which must be well designed and easy to interpret. This duality is a particular problem when handing over data files, because the customer often asks for maps, point clouds or 3D models, which in many cases he does not know what to do with. For this very reason, the raw data file must always be cleaned and freed from unnecessarily surveyed areas.
If you have to submit a report, make sure that the chapters and paragraphs are structured properly, and that you use easy-to-read fonts and font size. Furthermore, you should also pay attention to the proportionality of the chapters. If a special analysis is expected from you, then the presentation of the contractual background should not account for 90% of the report. Try to avoid unnecessary repetition of words, as it can be very confusing. Try to be clear.
In the reports, it is essential to depict the result maps and models generated during the processing. Here, it is extremely important that these figures stand on their own. This means that if someone does not read the text, but only examines an image, they can still get the most important information. In order for this to happen, it is necessary that the font, font size and color of the writings on the diagram can be read. In addition to all this, the colors and symbols used should have an explanation. If you use color scales – for example when displaying a DEM – you can choose both continuous and categorized versions. In the case of continuous coloring, the color explanation must be clearly visible, which shade means which values. If you choose a categorized display, the colors of each class must be clearly distinguishable from each other. It can also be confusing to understand the files if you deviate from the traditional coloring, for example, if the red shade represents the lowest values and the blue represents the highest. When it comes to map or 3D coloring, it must also be mentioned that even in the case of a more complex file – where vector and raster models are mixed – it is worth using strongly contrasting colors in order to separate the individual layers. You should also pay attention to the layer order, the general structure of which is as follows: rasters and polygons at the bottom, then lines and points at the top.
If you work with 3D files, then when creating an explanatory ilustration, it may happen that the map presentation will still be the winner. Several times I ran into the fact that screenshots of beautiful 3D models had zero information content as still images. At the same time, they became immediately interpretable in a map projection with appropriate accompanying text. Mainly because it is easier to put the maps into context, since so many basic maps (satellite or derived) can help us to display the data together with its wider environment. Don’t get me wrong, I’m not map-obsessed (if such a term exists), I was just trying to point out that 3D can easily lose its charm if we have to present it as a still image.
In the previous chapter, I already covered what display aspects you should consider. At the same time, you must be able to share the displayed data with the customer. To do this, you first need to make sure what exactly the customer needs. If they just want to see a report, no problem, it can be sent via email. There are also many cases where these raw data are requested even without professional knowledge. I am aware that it is difficult to give a large map file, point cloud or 3D model to someone who can open it and take measurements without special software or expertise. Having experienced this problem, we started the development of SurveyTransfer, which we hope will help as many professionals as possible. 🙂
If the customer has the software and expertise to open these files, several data sharing solutions can be considered: external hard drive, e-mail sender for large files, cloud-based data sharing, FTP server. If you are interested in this topic, do not miss our article. Of course, even in this case, the request may arise that the customer wants to view the shared files immediately on the web without any download time, and SurveyTransfer is a good solution for this as well. Such a collaboration interface provides an opportunity for quick information transfer, thus shortening the closing of a project even by days.
If you really liked what you read, you can share it with your friends. 🙂
Did you like what you read? Do you want to read similar ones?