UAV PHOTOGRAMMETRY FOR MAPPING AND 3D MODELING – CURRENT STATUS AND FUTURE PERSPECTIVES


UAVs are currently an important data source for inspection, surveillance, mapping, and 3D modelling. New applications in short- and close-range are presented, with UAVs replacing human aerial photogrammetry. Rotary or fixed wing UAVs may fly in manual, semi-automated, or autonomous modes,

.

A UAV, according to the UVS International definition, is a generic aircraft designed to fly without the presence of a human pilot. In the geomatics community, the word UAV is often used, although other terminology such as Remotely Piloted Vehicle (RPV), Remotely Operated Aircraft (ROA), Remote Controlled (RC) Helicopter, Unmanned Vehicle Systems (UVS), and Model Helicopter are also commonly used. UVS International divides UAVs into three categories based on size, weight, endurance, range, and flying altitude: tactical UAVs, which include micro, mini, close-, short-, medium-, medium-range, medium-range endurance, low altitude deep penetration, low altitude long endurance, and medium altitude long endurance systems. The mass spans from a few kilogrammes to 1,000 kilogrammes, the range from a few kilometres to 500 kilometres, the flying altitude from a few hundred metres to five kilometres, and the endurance from a few minutes to two or three days. In the past, military objectives and uses drove the development of UAV systems and platforms.
The main military objectives were unmanned inspection, surveillance, reconnaissance, and mapping of hostile regions. Przybilla and Wester-Ebbinghaus were the first to experiment with geomatics applications (1979). In recent years, more and more UAV applications in the geomatics sector have become commonplace. Indeed, UAV photogrammetry offers up a slew of new applications in the close-range aerial realm, as well as low-cost alternatives to traditional human aerial photogrammetry (Colomina et al., 2008; Eisenbeiss, 2009). The expansion of low-cost platforms paired with amateur or SRL digital cameras and GNSS/INS systems, which are required to guide the UAV with high accuracy to predetermined acquisition locations, explains this trend. The limited size and payload of certain UAV platforms make it difficult to carry high-quality IMU devices like those used in conjunction with airborne cameras or mapping LiDAR sensors. Because GNSS is primarily utilised for code-based location, it is insufficient for precise direct sensor orientation. RTK methods would enhance positioning quality to the decimeter level, but the system would become too complicated, costly, and heavy to employ.3D Modeling Services used for  furniture and Augmented reality have emerged as the most recent trends in home decorating and design. It enables users to take on the role of interior designers by creating their own environments and customising them with their own finishes, décor elements, and furnishings. Furthermore, using Augmented Reality, users may put up the virtual lens and experience the locations as if they were really there!
The most affordable systems are simple and hand-launched UAVs that fly autonomously utilising a GPS-driven autopilot and, in most cases, an IMU sensor, however platform stability in windy places may be an issue. More stable systems, usually with a gasoline engine, allow for more professional cameras aboard or even surveying with them, due to a bigger payload. The article provided an overview of current UAV systems, with a focus on platforms that are utilised for geomatics applications. The ability of real UAV systems to transmit high temporal and spatial resolution picture information fast and enable a speedy reaction in a variety of critical scenarios when instant access to 3D geoinformation is needed is a significant benefit. Indeed, UAVs have real-time data collecting, transmission, and perhaps processing capabilities.
Because rotary wing UAV platforms can take off and land vertically, there is no need for a runway. UAVs may be used as a supplement or substitute for terrestrial collection in certain cases (images or range data). The resulting high-resolution pictures may be utilised for texture mapping on existing 3D data, as well as mosaic, map, and drawing production, in addition to geometric modelling. When compared to typical aerial platforms, they have lower operating expenses and a lower danger of access in hostile settings while maintaining excellent accuracy. However, in order to attain the same picture coverage at a similar quality, compact or medium format cameras, which are often used, especially on low-cost and small payload systems, demand the capture of a larger number of photos. Although camera and platform stabilisers may lessen weather reliance, the stability and durability of low-cost and light platforms remains an outstanding topic, particularly in windy places. High altitude surveys may cause problems for gasoline and turbine engines, while payload constraints compel the usage of light-weight IMUs, preventing direct geo-referencing. A disadvantage might be the need of at least two people for system movements and transportation. The capture of picture blocks with acceptable geometry for the photogrammetric process remains a key issue, particularly for large-scale projects and nonflat objects (e.g. buildings, towers, rock faces, etc.). While flight planning is straightforward when employing nadiral views, it becomes considerably more difficult when dealing with 3D objects that need convergent photos and, maybe, vertical strips. Future work will focus on developing tools to make this effort easier.
Despite the fact that automated image processing is currently achievable with fairly dependable and exact results, there are still advancements that may be made in the near future. Advanced DSM generation algorithms may give surface models in a quick period owing to GPU programming, while high-end navigation sensors like DGPS and costly INS would enable immediate georeferencing of the captured photos on location. Advanced SLAM methods could be used to achieve real-time image orientation in low-end navigation systems (Nuechter et al., 2007; Konolige and Agrawal, 2008; Strasdat et al., 2010), though working incrementally on a frame by frame basis can lead to error accumulation and drift errors (Steffen and Foerstner, 2008). For (real-time) picture analysis and object space feature extraction, sequential estimation techniques based on Givens transformations (Gruen and Kersten, 1992) might be used. In any event, for applications needing great precision, lab post-processing and some user intervention will almost certainly be required.
UAV rules are now being developed in a number of nations throughout the globe in order to establish technical criteria and locations where these devices may be deployed, such as over metropolitan areas, therefore expanding their range of uses.

Read more