Building on our past blog, 'How to Produce an Accurate Aerial Survey', we wanted to offer some tips for flight parameters in order to produce usable and accurate orthomosaic (.tiff) and point cloud (.las) files. These rules of thumb are based on our experience collecting data and should be adjusted depending project specifics and the desired outcome.
Based on our experience in data collection the key factors that directly affect ground sampling distance (“GSD”),point cloud density are image overlap, flight patterns, altitude, camera angle/orientation and camera megapixels. At the same time, a critical element you should also be aware of is creating too much data. If the GSD requirements are specified too high, more flight lines--and therefore more images--will be required and the amount of data per flight will increase. All this leads to difficulty in file handling and increased processing times. Being clear on your accuracy needs before data collection will help to minimize data overload. We hope these tips are useful.
There are many UAV flight management software applications on the market and they all have similar features. For a firm just getting started, we recommend finding one that has a user-friendly interface and is easy to use. We have used Maps Made Easy, which now supports Airmap and Low Altitude Authorization and Notification Capability system(LAANC). Generally speaking, most UAV flight software has two primary flight modes, manual and autonomous. This setting is found in commercially available flight planning and management software. For consistency of data quality, you should always use the autonomous flight mode setting.
To achieve the best results with our proprietary software mentioned earlier in this article we recommend a minimum ground sampling distance (GSD) of 3cm(1.18 in) for accurate object detection.Depending on the intended use of the output, the GSD could be higher or lower for your specific project needs. Keep in mind as the pixel GSD increases, the spatial resolution of the image will decrease and objects will become less visible, making it harder, or impossible to trace them. This is why good flight parameters are critical in order to create an accurate output.
Once you have decided on your GSD the next step will be to setup your flight pattern for data collection. After working with many iterations of flight patterns, altitudes and camera angles, we recommend flying a double grid pattern at 150’ with the camera set at oblique/30 degrees in combination with a separate single grid pattern with camera set at nadir/90 degrees. In areas with vegetation or homogeneous terrain, you should increase the altitude of the single grid/90 degree by at least 50’. Flying only nadir images does not successfully capture details of facades, oblique images are very important in the UAV photogrammetric process for 3D modeling applications.
Keeping in the spirit of managing both data quality as well assize we have also learned through experience that image forward and side overlap should be set at 80/85. It might be tempting to fly with a greater overlap but the downside for the user is a huge amount of data, which leads to longer processing time and higher required hardware resources.
The quality of your camera sensor to produce higher resolution photos with a greater pixel density is also a key factor. Simply put, an image with more pixels contains more information, which produces a denser point cloud. A20-megapixel camera or higher will deliver a GSD of less than 3cm at 150’ in altitude. Often professional UAV’s are equipped with higher quality camera of 40 megapixels. In our experience these cameras don’t always lead to better model resolution, as often these large images have to be down-sized (broken into smaller pieces) before processing depending on the size of the area you’re flying. Be sure to determine your GSD based on your particular equipment.
If you are currently flying UAVs in your current organization, these settings will help you achieve the best quality outputs. If you are looking to outsource your UAV data collection, there are a number of piloting platforms available that give you the ability to have pilots bid on the data collection scope of work. One of the resources we have used is droners.io which is one of the largest networks of commercial UAV pilots available for hire.In one of our next articles we’ll describe in more detail how to hire and onboard a UAV pilot, so that you can achieve the data collection quality desired for your project.