Dálkový průzkum v ekologii invazí? Detekce, monitoring a kontrola invazních rostlin

Remote sensing in invasive ecology? Detection, monitoring and control of alien plant species

conference paper

en

Plant invasions represent a serious threat to changing landscapes. Once fully established, the invaders spread rapidly, outcompete native species and their removal and elimination is difficult. Early detection enabled by the remote sensing (RS) means could make the management more efficient and less expensive. In an ongoing project, we aim at developing innovative methods of mapping invasive plant species (semi-automatic detection algorithms) by using purposely designed unmanned aircraft (UAV) and combination of aerial and satellite optical RS data. We examine the possibilities for detection of several invasive herbs, such as Heracleum mantegazzianum, Robinia pseudoaccacia, Ailanthus altissima, and Fallopia s.l. (F. japonica, F. sachalinensis and F. × bohemica). Our aim is to establish fast, repeatable and efficient computer-assisted method of timely monitoring of invasions applicable over large areas, reducing costs of extensive field campaigns. We evaluate the imagery of different origin (satellite, aerial and UAV), spectral (panchromatic, multispectral and color), spatial (very high to medium) and temporal resolution, and various technical approaches (object-based – OBIA, pixel-based and hybrid approaches). The rule-based OBIA hierarchical classification was successfully applied for detection of H. mantegazzianum even on low spectral resolution historical panchromatic aerial photographs and the results let us to study the process and progressing of invasion in detail. Thanks to its flexibility and low cost, UAV enables to assess the effect of phenological stage and spatial resolution on the recognition of the species, and is most suitable for monitoring the efficiency of eradication efforts. However, several challenges exist in UAV application, such as geometrical distortions, radiometric differences during the flight, high amount of data to be processed and legal constrains for the UAV flight missions over urban areas (often highly invaded). In our study, we address tradeoffs between spectral, spatial and temporal resolutions required for balance between the precision of detection and economic feasibility. The resulting data enable assessment of the invasibility of different types of habitat, modelling the potential species distribution and identifying the drivers of spread; the derived knowledge can serve as a valuable input for prediction, monitoring and prioritization purposes in other areas.

Rostlinné invaze představují vážnou hrozbu pro moderní krajinu... ... V předkládaném textu jsou evaluovány možnosti metod dálkového průzkumu Země pro detekci a monitorování rostlinných invazí.

Plant invasions represent a serious threat to changing landscapes. Once fully established, the invaders spread rapidly, outcompete native species and their removal and elimination is difficult. Early detection enabled by the remote sensing (RS) means could make the management more efficient and less expensive. In an ongoing project, we aim at developing innovative methods of mapping invasive plant species (semi-automatic detection algorithms) by using purposely designed unmanned aircraft (UAV) and combination of aerial and satellite optical RS data. We examine the possibilities for detection of several invasive herbs, such as Heracleum mantegazzianum, Robinia pseudoaccacia, Ailanthus altissima, and Fallopia s.l. (F. japonica, F. sachalinensis and F. × bohemica). Our aim is to establish fast, repeatable and efficient computer-assisted method of timely monitoring of invasions applicable over large areas, reducing costs of extensive field campaigns. We evaluate the imagery of different origin (satellite, aerial and UAV), spectral (panchromatic, multispectral and color), spatial (very high to medium) and temporal resolution, and various technical approaches (object-based – OBIA, pixel-based and hybrid approaches). The rule-based OBIA hierarchical classification was successfully applied for detection of H. mantegazzianum even on low spectral resolution historical panchromatic aerial photographs and the results let us to study the process and progressing of invasion in detail. Thanks to its flexibility and low cost, UAV enables to assess the effect of phenological stage and spatial resolution on the recognition of the species, and is most suitable for monitoring the efficiency of eradication efforts. However, several challenges exist in UAV application, such as geometrical distortions, radiometric differences during the flight, high amount of data to be processed and legal constrains for the UAV flight missions over urban areas (often highly invaded). In our study, we address tradeoffs between spectral, spatial and temporal resolutions required for balance between the precision of detection and economic feasibility. The resulting data enable assessment of the invasibility of different types of habitat, modelling the potential species distribution and identifying the drivers of spread; the derived knowledge can serve as a valuable input for prediction, monitoring and prioritization purposes in other areas.

invazní druhy; dálkový průzkum, rozlišení, klasifikace

invasive species; remote sensing, resolution, clasiffication

19.07.2015

Masaryk University

Brno

978-80-210-7860-4

58th Annual Symposium of the International Association for Vegetation Science

1

268–268

1