A Backpack-Mounted Omnidirectional Camera with Off-the-Shelf Navigation Sensors for Mobile Terrestrial Mapping: Development and Forest Application

• DOI: 10.3390/s18030827
• Publication Date: 2018-03-09T17:17:41Z
• Authors: Mariana Campos; Antonio Tommaselli; Eija Honkavaara; Fabricio Prol; Harri Kaartinen; Aimad El Issaoui; Teemu Hakala
• Keywords: PMTS data quality; Personal mobile terrestrial system; Chemical technology; ta1171; personal mobile terrestrial system; omnidirectional cameras; low-cost sensors; forest mapping; PMTS data quality; personal mobile terrestrial system; TP1-1185; 01 natural sciences; Article; 620; omnidirectional cameras; low-cost sensors; Forest mapping; Omnidirectional cameras; Low-cost sensors; forest mapping; 0105 earth and related environmental sciences

The use of Personal Mobile Terrestrial System (PMTS) has increased considerably for mobile mapping applications because these systems offer dynamic data acquisition with ground perspective in places where the use of wheeled platforms is unfeasible, such as forests and indoor buildings. PMTS has become more popular with emerging technologies, such as miniaturized navigation sensors and off-the-shelf omnidirectional cameras, which enable low-cost mobile mapping approaches. However, most of these sensors have not been developed for high-accuracy metric purposes and therefore require rigorous methods of data acquisition and data processing to obtain satisfactory results for some mapping applications. To contribute to the development of light, low-cost PMTS and potential applications of these off-the-shelf sensors for forest mapping, this paper presents a low-cost PMTS approach comprising an omnidirectional camera with off-the-shelf navigation systems and its evaluation in a forest environment. Experimental assessments showed that the integrated sensor orientation approach using navigation data as the initial information can increase the trajectory accuracy, especially in covered areas. The point cloud generated with the PMTS data had accuracy consistent with the Ground Sample Distance (GSD) range of omnidirectional images (3.5–7 cm). These results are consistent with those obtained for other PMTS approaches.