研究目的
To address the bottleneck in UAS-based field observations and high-throughput phenotyping by developing a flight planning tool that incorporates photographic concepts, which are often ignored in existing tools, to improve image quality and reduce uncertainty in remote sensing applications.
研究成果
The PhenoFly Planning Tool successfully integrates photographic concepts into UAS flight planning, addressing gaps in existing tools. It reduces uncertainty in remote sensing by optimizing parameters like motion blur and GSD, and supports tasks from equipment selection to flight preparation. Field trials confirmed its effectiveness in improving image quality and mission success.
研究不足
The tool is an offline tool without autopilot functionality, requiring integration with third-party software for flight implementation. It assumes nadir view and planar terrain, which may not hold in all real-world conditions. The evaluation of other tools was limited to those with affordable fees or public documentation, potentially missing some commercial solutions.
1:Experimental Design and Method Selection:
The study involved designing and developing the PhenoFly Planning Tool software in R using R Shiny, based on theoretical models of photography and mapping with UASs, including concepts like sensor characteristics, lens properties, view, sharpness, exposure, mapping areas, GCPs, viewing geometry, and way-point flights.
2:Sample Selection and Data Sources:
Field trials were conducted using a Matrice 600 Pro UAS and a Sony α9 camera with a Sonnar T* FE 55 mm F
3:8 ZA lens, imaging UV-coated GCP prints, a DIN A4 IT7 color checker panel, and experimental wheat plots. Data from experiments were used to validate the tool. List of Experimental Equipment and Materials:
UAS: Matrice 600 Pro (DJI), Camera: Sony α9 (ILCE-9) with Sony Sonnar T* FE 55 mm F
4:8 ZA lens, Gimbal:
Ronin-MX (DJI), GCPs: UV-coated prints (
5:2x2m), Color checker:
IT
6:7 panel (Targets.coloraid.de), Software:
R, R Shiny, and various R packages (ggplot2, gridExtra, NMOF, RJSONIO, rlist, rgdal, readr, zoo, data.table, raster).
7:Experimental Procedures and Operational Workflow:
Experiments included varying flight heights and speeds to assess GSD and motion blur effects, and a full mapping flight with GCP placement. Photos were processed using Agisoft PhotoScan Professional for SfM and image projection to determine exposure stations and viewing geometries.
8:Data Analysis Methods:
Data were analyzed using the developed software to calculate parameters like GSD, motion blur, and viewing geometry frequencies. Statistical analysis involved comparing predicted and realized flight parameters.
独家科研数据包,助您复现前沿成果,加速创新突破
获取完整内容