Oblique imagery got thrust into the spotlight with Microsoft’s publication of the Bird’s Eye imagery to the public. Geo-referenced oblique images are useful for highly automatic creation of 3D city models on their own or in combination with airborne Lidar and planar topographic maps containing building outlines. With the advent of multi-ray matching algorithms using rotation invariant features such as SIFT and SURF, cameras capable of simultaneously capturing multiple view geometries have become highly desirable, making Lidar systems an added complication instead of a complimentary tool. The utility of oblique imagery is now even higher allowing the creation of fully textured 3D models of entire cities given and efficient capture system.

In a typical configuration in addition to vertical, oblique images are also taken, enabled by a sensor system configured with five cameras, one directed nadir, the others viewing forward, backward, left and right. The shape of the ground coverage captured simultaneously by the five cameras looks like a Maltese Cross. Depending on the along-track and across-track overlap, each point on the ground may be visible in multiple of images, provided there is no occlusion. Often the five cameras are mounted rigidly together and their geometric configuration is calibrated to enable accurate measurements in both the vertical and oblique images. In other cases the rigidity of the system is countered by flexibility allowing the cameras to move and achieve greater footprint and oblique angle diversity, this is illustrated by the patented step-stare configuration created by Google or the wide-angle sweep approach adopted by Visionmap A3 Edge. The main advantage of oblique images over vertical is better and more intuitive interpretation, making photogrammetry accessible to non-professionals. Oblique images show parts of buildings and other structures invisible in vertical images, making them useful for 3D cadastral applications, determining property taxes and providing building permits. Outside the GIS box the generated 3D models are used by architects, city planners, building developers and special effects companies alike to bring their vision to life. Looked at the other way around, they can also act as a spatial access means to databases containing building information.

Here is a round-up of some of the Oblique Camera systems currently in operation. This list is by no means exhaustive, but it gives a fair sample of variety in the oblique camera market and points towards products we will see in the near future.

Ultracam Osprey – More than a standard camera, the UltraCam Osprey houses two cameras in one photogrammetric grade housing, using cutting edge technology to collect photogrammetry-grade nadir images (PAN, RGB, NIR) and oblique images (RGB) simultaneously, serving applications such as cadastre, infrastructure planning, DTMOrtho or DSMOrtho generation. Like all UltraCam systems, the UltraCam Osprey delivers subpixel accuracy, high dynamic range, and integrates all system components in the sensor head including the optional UltraNav direct georeferencing and flight management subsystem, as well as full UltraMap software support.

osprey

Microsoft Ultracam Osprey used in Bing Maps oblique imagery capture instead of Pictometry

A3 Edge  – A3 Edge was initially designed as a wide-swath highly efficient orthophotography capture system. However with a wide field-of-view (up to 106 degrees), A3 also collects multi-directional images of each point in both vertical and oblique angles. To achieve an overall Maltese cross footprint the camera needs to be flown in a crisscross pattern.

Midas 400 (Chimera) – This system uses modified Phase One cameras known as TCam. The main advantage of the TCam lies in its replaceable shutter. This allows the user to change a broken shutter in mid-air, within minutes, and continue flying without the need for re-calibration or re-boresighting. The 400 mp Chimera is a very large camera system with a view angle of more than 120° which has completely revolutionized the way an aerial camera is installed in an aircraft. A new stabilized mount system has been specially developed to allow the Chimera to be installed as low as possible in a standard 19in (48cm) camera hole. This system, Puppetair, gives the Chimera a full stabilization range of +/- 10 degrees in pitch and roll as well as +/- 30 degrees of drift correction. In the patented Puppetair principle, the stabilized mount mechanism is actually above the Chimera and the camera moves in the camera hole.

Leica RCD30 Oblique – The new Leica RCD30 Oblique camera system is specifically designed for high accuracy 3D urban mapping and 3D corridor mapping applications. Based on the Leica RCD30, the world’s first 60MP multispectral medium format camera. The cameras are laid out in the classic cross pattern offering high photogrammetric accuracy and flexibility.

RCD30_oblique_footprint

Leica RCD30 Oblique showing Maltese Cross footprint

Pictometry – Pictometry is one of the earliest movers in the Aerial oblique camera arena, supplying the bird’s Eye Imagery to Microsoft prior to the development of Microsoft’s own Ultracam Osprey. The pictometry camera system uses the tried and tested 5 camera cross formation. In addition Pictometry allows gereferencing of every pixel in the oblique imagery allowing direct digitisation from the datasets. The cameras used in this system are specialised IMPERX ones and do not have the variety of lens options COTS cameras have.

step_stare

Google oblique camera system with step-stare mode

Google Custom Oblique – Not to be left behind Google has patented its own oblique aerial imagery capture system combining both the multiple camera systems in established configuration with the flexibility of a moving camera system such as the A3 Edge. This creates a huge data volume which Google is uniquely placed to process and deliver. The variety of imaging angles will create greater photogrammetric rigidity, lead to fewer occlusions and allow faster captures due to the increased overall swath.

hyperpod_1_shot

NearMap Hyperpod with 14 cameras

NearMap Hyperpod – What Google does using moving cameras can be done using lots of cameras in parallel. This system only has vertical, rear and right looking cameras. To keep the overall camera count down the NearMap system sacrifices flying efficiency. The cross pattern is formed by flying the same area in a forward and reverse direction. There are also overview frames at very low resolution to fill up any gaps in the very small footprint high resolution cameras. The downside of such a large camera count system is that all cameras need to be individually calibrated.

With so many oblique capture options there are more vendors coming to the market offering oblique mosaics and 3D models. There is still a lot of room for innovation in terms of camera bodies, lenses, electromechanics, capture platforms and mounting options.