What's New in Photogrammetry?
The imagery that photogrammetry captures through planes and unmanned drones enables surveyors and professionals in walks of life like construction, mining, crime forensics and government to obtain detailed imagery of physical areas that can facilitate decision making and business operations.
In recent years, there have been firms that have dumped photogrammetry in favor of satellite-recorded imagery that is available over free Internet mapping services such as Google Maps, but these commercial applications have their limits. They can’t always guarantee high-precision accuracy that can solve the questions that users ask.
Last year, Darryl Murdock, vice president of professional development for the U.S. Geospatial Intelligence Foundation, described how an engineer used Internet-based mapping to map a street and the size of a crack that a lawsuit plaintiff tripped over in a slip and fall lawsuit. But the information about the grade in the sidewalk, the date that the photo was taken and the orientation of the camera were missing. The defendant’s lawyer chose instead to use photogrammetry to investigate the scene of the accident. Photogrammetry’s superior precision ultimately won the case for the defendant.
The Photogrammetry Edge
“The accuracy of photogrammetry, which is down to the thousandth of an inch, enables us to pinpoint issues such as surface level differences that can cause falls,” says Lee DeChant, president of DeChant Consulting, which, among other projects, provides expert witness testimony in federal court for tort slip and fall cases. “To compare, if you use an Internet mapping application for accident reconstruction, you might not be able to get accuracy that is better than within one half meter of precision.”
Lisa Chen, technical communication manager at Swiss-based Pix4D, which provides photogrammetry software, also acknowledges the broad adoption of photogrammetry in law enforcement forensics.
“By using photogrammetry, law enforcement and forensics experts can avoid human errors for on-site documentation and measurement,” Chen says. “This speeds up the process and obtains better results at a lower cost. “In such cases, time is the key. Within 10 or 20 minutes, image/video acquisition is done and everything can be finished up afterwards in the office. This means less risk of evidence being washed out due to critical weather conditions, or of the need to shut a highway down.”
DeChant agrees with Chen, and adds that the addition of unmanned aerial vehicles (UAVs) like drones makes the job even easier.
“I do a lot of work in accident reconstruction and law enforcement,” DeChant says. “There’s a big advantage in doing mapping with UAVs because they can get in and out of an area quickly. On a busy interstate highway, if we use patrol cars to investigate and document an accident it could take three or four hours, and that would impact traffic. The drone can do this in 10 minutes. It records the scene, we load this information onto the laptop, we push a button and we generate a point cloud. It’s a huge benefit.”
Photogrammetry is also a technology of choice in other industries, such as tower inspection for telecommunications and electric utilities.
“By using a rotary drone with photogrammetry software, a tower inspection can be done in a much more efficient way without putting workers in danger,” Chen says. “Traditionally, workers need to climb up the tower under the threat of height and electromagnetic radiations, taking close-up pictures and noting down the issues and their location they observe. Now they can fly with a drone, taking overlapped close-up images, which they can reconstruct in 3D with software. With a cloud-based interface, they can even correspond any point of interest to the related high-res images.”
Photogrammetry or LiDAR?
When it comes to UAV field work, there are best practice recommendations, such as using a technology like LiDAR to map beneath obstructions like a tree canopy, or to model narrow objects like transmission lines, pipes or roof edges. Conversely, photogrammetry works best in bare earth imaging of mine sites, earth works projects and other areas that are not occluded by trees, buildings or equipment; or in small mapping projects.
The takeaway is that both technologies can be worked together for best results.
“LiDAR is a totally different technique compared to passive remote sensing techniques like photogrammetry,” Chen says. “It is capable of capturing more valid data for photogrammetry limitations, such as more valid points on reflective surfaces (glasses, metals), uniform texture surfaces (snow, ice, white walls), and penetrating more in dense forests.”
However, Chen points out that LiDAR is more expensive and requires a steeper learning curve.
“It is less adaptable to the new platforms compared to photogrammetry, which only involves cameras,” she says. “With drone photogrammetry, geospatial outputs can be generated much faster and with higher accuracy and much lower cost. Satellite imagery can be used as an economic way for large-scale mapping with lower resolution, but with that, the cloud coverage is a problem, especially for certain landscapes.”
Chen says that over the next three years, companies can expect photogrammetry to break new ground due to the advancement of image processing power and machine learning technologies. “At the same time, these advances will be complemented with a fuller integration of photogrammetry with different technologies such as LiDAR, virtual reality (VR) and augmented reality (AR).”
Equally important are more usable photogrammetry tools that are now available and that require less training than the tools of the past.
Analogue photogrammetry especially required a great deal of training, Chen says. Now, digital photogrammetry, sensors, platforms and computer vision technology have reduced the need to possess profound photogrammetry knowledge to complete a mapping project. While she says it does still help for solving problems and improving productions, the key is choosing the right tools.
DeChant sees drones as a major driver of photogrammetry. Over the past few years he has watched photogrammetry become more actively used in UAVs.
“The optics are wonderful. Sensors on board drones and other assets now have 20-megapixel resolutions. With a Wi-Fi direct connect, you can get an image onto a laptop immediately and produce photos,” he says. “The biggest drawback is licensing, which is limited for commercial applications in the U.S.”
Photogrammetry Best Practices
For companies that want to use photogrammetry, Chen lists four best practices.
- Make sure that the right tools are chosen, including platforms, sensors, and processing hardware and software.
- Understand the advantages and the limitations of photogrammetry, and create the most optimal plans for mapping target regions.
- Pay attention to regulations like those on UAV flights and to output requirements like accuracy of the geospatial results.
- Always keep an eye on the new technologies that are coming out that can generate common-format outputs, which can easily be integrated with other technologies.