Bezpilotní prostředky pro detekci a monitoring invaznních rostlinných druhů

Unmanned Aerial Vehicles for Alient Plant Species Detection and Monitoring

článek v časopise ve Scopus, Jsc

en

Invasive species spread rapidly and their eradication is difficult. New methods enabling fast and efficient monitoring are urgently needed for their successful control. Remote sensing can improve early detection of invading plants and make their management more efficient and less expensive. In an ongoing project in the Czech Republic, we aim at developing innovative methods of mapping invasive plant species (semi-automatic detection algorithms) by using purposely designed unmanned aircraft (UAV). We examine possibilities for detection of two tree and two herb invasive species. Our aim is to establish fast, repeatable and efficient computer-assisted method of timely monitoring, reducing the costs of extensive field campaigns. For finding the best detection algorithm we test various classification approaches (object-, pixel-based and hybrid). Thanks to its flexibility and low cost, UAV enables assessing the effect of phenological stage and spatial resolution, and is most suitable for monitoring the efficiency of eradication efforts. However, several challenges exist in UAV application, such as geometrical and radiometric distortions, high amount of data to be processed and legal constrains for the UAV flight missions over urban areas (often highly invaded). The newly proposed UAV approach shall serve invasive species researchers, management practitioners and policy makers.

Příspěvek popisuje aplikaci bezpilotních leteckých prostředků pro detekci a monitoring invazních rostlinných druhů jako velmi účinnou metodu.

Invasive species spread rapidly and their eradication is difficult. New methods enabling fast and efficient monitoring are urgently needed for their successful control. Remote sensing can improve early detection of invading plants and make their management more efficient and less expensive. In an ongoing project in the Czech Republic, we aim at developing innovative methods of mapping invasive plant species (semi-automatic detection algorithms) by using purposely designed unmanned aircraft (UAV). We examine possibilities for detection of two tree and two herb invasive species. Our aim is to establish fast, repeatable and efficient computer-assisted method of timely monitoring, reducing the costs of extensive field campaigns. For finding the best detection algorithm we test various classification approaches (object-, pixel-based and hybrid). Thanks to its flexibility and low cost, UAV enables assessing the effect of phenological stage and spatial resolution, and is most suitable for monitoring the efficiency of eradication efforts. However, several challenges exist in UAV application, such as geometrical and radiometric distortions, high amount of data to be processed and legal constrains for the UAV flight missions over urban areas (often highly invaded). The newly proposed UAV approach shall serve invasive species researchers, management practitioners and policy makers.

Automatické zpracování; APM Plane; letoun; Geometrické a radiometrické korekce; invazní druhy; Mapování; Objektove a pixelově orientovaná klasifikace

Automated processing; APM Plane; Fixed-wing platform; Geometric and radiometric correction; Invasive species; Mapping; Object- and pixel-based classification

2015

02.09.2015

International Society of Photogrammetry and Remote Sensing (ISPRS), Copernicus

Toronto, Canada

2194-9034

XL-1/W4

ICWG I/Vb

83–90

8