Kinect v2 sensor-based mobile terrestrial laser scanner for agricultural outdoor applications

Rosell Polo, Joan Ramon; Gregorio López, Eduard; Gené Mola, Jordi; Llorens Calveras, Jordi; Torrent Martí, Xavier; Arnó Satorra, Jaume; Escolà i Agustí, Alexandre

doi:https://doi.org/10.1109/TMECH.2017.2663436

View/Open

Postprint (805.1Kb)

Issue date

2017-02-02

Author

Rosell Polo, Joan Ramon

Gregorio López, Eduard

Gené Mola, Jordi

Llorens Calveras, Jordi

Torrent Martí, Xavier

Arnó Satorra, Jaume

Escolà i Agustí, Alexandre

Suggested citation

Rosell Polo, Joan Ramon; Gregorio López, Eduard; Gené Mola, Jordi; Llorens Calveras, Jordi; Torrent Martí, Xavier; Arnó Satorra, Jaume; Escolà i Agustí, Alexandre; . (2017) . Kinect v2 sensor-based mobile terrestrial laser scanner for agricultural outdoor applications. IEEE-Asme Transactions on Mechatronics, 2017, vol. 22, núm. 6, p. 2420 - 2427. https://doi.org/10.1109/TMECH.2017.2663436.

Abstract

Mobile terrestrial laser scanners (MTLS), based on light detection and ranging (LiDAR) sensors, are used worldwide in agricultural applications. MTLS are applied to characterize the geometry and the structure of plants and crops for technical and scientific purposes. Although MTLS exhibit outstanding

performance, their high cost is still a drawback for most agricultural applications. This paper presents a low-cost alternative to MTLS based on the combination of a Kinect v2 depth sensor and a Real Time Kinematic (RTK) global navigation satellite system (GNSS) with extended color information capability. The theoretical foundations of this system are exposed along with some experimental results illustrating their performance and limitations. This work is focused on open-field agricultural applications, although most conclusions can also be extrapolated to similar outdoor uses. The developed Kinect-based MTLS (K2-MTLS) system allows to select different acquisition frequencies and fields of view (FOV), from one to 512 vertical slices. The authors conclude that the better performance is obtained when a FOV of a single slice is used, but at the price of a very low measuring speed. With that particular configuration, plants, crops, and objects are reproduced accurately. Future efforts will be directed to increase the scanning efficiency by improving both the hardware and software components and to make it feasible using both partial and full FOV.

URI

http://hdl.handle.net/10459.1/59504

DOI

https://doi.org/10.1109/TMECH.2017.2663436

Is part of

IEEE-Asme Transactions on Mechatronics, 2017, vol. 22, núm. 6, p. 2420 - 2427