Geo Week News

July 15, 2019

nFrames SURE combines photogrammetry and lidar for large-scale data production

nframes fig1

Header image: 3D Mesh extracted fully automatically from the Leica CityMapper sensor using the nFrames SURE Software. Mesh geometry from Lidar only (left) and from both imagery and Lidar (right). Data courtesy of Geomni/Verisk (George Halley) 

Now that the mapping industry produces and consumes data than ever before, new and automated photogrammetric workflows help to raise data production rates. A good example of this is the latest version of nFrames’ SURE software, that now offers the extraction of true orphophotos and supports lidar data, which enables the creation of 3D meshes.

nFrames is a provider of photogrammetric software that uses intelligent algorithms that can automate the processing of constantly increasing data volumes at an ever-growing acquisition frequency. The software is used worldwide by private and public mapping organizations for the automated production of point clouds, digital surface models, true orthophotos and meshes from imagery for cities and entire countries by extracting  precise 3D information for every pixel.

While the automatically produced true orthophotos are driving automation for applications in the 2D GIS domain, meshes are now driving GIS applications in the 3D beyond city modelling, visualization and web-streaming as they elegantly combine geometry as a closed surface in combination with high-resolution imagery. Particularly interesting applications are direct feature extraction for cadasters, BIM or simulations on the mesh surface, for example for 5G wave propagation or occlusion analysis. The fast-evolving 3D GIS software solutions offering these tools are now using also streaming-ready formats, such as Cesium tiles, the Esri Scene Layer Package or Open Scene Graph, which can be directly produced by the nFrames SURE Software.

Lidar support

Besides image-based surface information retrieved from photogrammetry, lidar point clouds can now optionally be imported in the latest SURE release. More and more lidar systems include high-resolution medium format cameras, while on the other hand oblique camera systems are now made available with additional lidar sensors that enable fast hybrid data capture. This way, the complementary behavior of both sensor systems can be exploited with SURE, enabling more completeness and robustness of the extracted surface, while being able to recover more detail from the images.

Whereas photogrammetry is strong for high spatial resolution, edges and radiometric information, lidar is strong as an active system, enabling additional measurements in poorly textured image areas like shadows. Furthermore, only one laser pulse is required to derive a measurement, which is beneficial for urban environments with narrow street canyons or for vegetation penetration. In the SURE workflow, lidar point clouds can now be directly imported after the imagery import using the industry standard formats .las and .laz enabling a seamless interface for producing a hybrid surface representation.

Automated extraction of true orthophotos

Besides 3D meshes, true orthophotos play an important role for automation, particularly in the large scale traditional country- or citywide mapping process. Within the traditional orthophoto workflow, the DTM break-line and seamline editing led to delivery delays of months, resulting in additional complexity and higher costs. With the automated extraction of true orthophotos with SURE software, this additional effort becomes obsolete.

Traditional Orthophoto with building lean and occlusions (left). True Orthophoto fully automatically generated using nFrames SURE Software without needing manual editing like DTM breakline or seamline editing (right). Data courtesy of Aerowest.

Based on the precisely reconstructed 3D surface data, the perspective image distortions like the building lean can be removed automatically, enabling a truly orthographic representation. This leads to consistency with mapping data such as building footprints, or the ability to extract spatially correctly located data directly from the multispectral true orthophoto. At the same time, the DSM or a point cloud representation are consistently available. This consistency is also beneficial for automatic analysis like with machine learning applications, particularly when multiple time epochs are available.

Scaling data processing workflow automation

The workflow automation and the data capturing performance of the sensors enable small organizations to capture and process high data volumes in a short time. For instance, the acquisition of an entire city of more than 100 square kilometers coverage at a 5cm ground resolution within a few hours with nowadays oblique camera systems are becoming the norm. This means that even small organizations are reaching yearly input data volumes in the petabyte domain easily.

In order to meet the processing efficiency, the computation within SURE has been highly optimized through multi-threading and optional graphics card use. Furthermore, distributed processing capabilities enable processing on local clusters and cloud environments. This way, the benefits of automation through intelligent algorithms can be scaled flexibly to high throughput, empowering both small and large organizations to enter the next level large-scale data production.

Want more stories like this? Subscribe today!

Read Next

Related Articles


Join the Discussion