Detection of invasive plants using remote sensing : a case study of ragweed in the Rhone-Alps region, France

• Published in: International journal of remote sensing Vol. 29; no. 3-4; pp. 1109 - 1124
• Main Authors: AUDA, Y, DECHAMP, C, DEDIEU, G, BLASCO, F, DUISIT, D, PONTIER, J.-L
• Format: Journal Article
• Language: English
• Published: Abingdon Taylor and Francis 2008; Taylor & Francis
• Subjects: Animal, plant and microbial ecology; Applied geophysics; Biological and medical sciences; Continental interfaces, environment; Earth sciences; Earth, ocean, space; Environmental Sciences; Exact sciences and technology; Fundamental and applied biological sciences. Psychology; General aspects. Techniques; Global Changes; Internal geophysics; Sciences of the Universe; Teledetection and vegetation maps; Isere France; Rhône Alpes; Drome France; France; Europe; Western Europe; Allergy; maps; pollen; multispectral remote sensing; surveys; geographic information systems; vegetation; Annual plant; impacts; Survey; Invasive species; Health status; Infestation; methodology; Public health; communities; Environment quality; COVER; "KNAPWEED CENTAUREA-MACULOSA; KNAPWEED CENTAUREA-MACULOSA; HYPERSPECTRAL IMAGERY; AGREEMENT; INFESTATIONS
• ISSN: 0143-1161; 1366-5901
• DOI: 10.1080/01431160701355231

Ragweed, Ambrosia artemisiifolia, is an annual plant, the pollen of which is responsible for respiratory allergies. In 2005, these allergies affected up to 20% of the population in parts of the department of Rhne. This is a major health problem for six of the eight departments of the Rhne-Alpes region (Ain, Ardche, Drme, Isre, Loire and Rhne). Our objective was to validate a method to map ragweed infestation in a village and then to extend this experiment to a department. Using methods developed by Auda et al. (2002a), we undertook a survey of ragweed infestation at the end of July 2005 in the village of Estrablin (Isre). Plots of land were registered in a Geographical Information System, with indications of crop type and ragweed infestation included in each plot. The sample area covered 3650km2 and constituted 30% of the total area of the village. The data were used to validate a SPOT 5 multispectral satellite image captured on 16 August 2005. Analysis of the ground data confirmed the extent of the ragweed infestation. More than 90% of the sample area was found to be infested. The image processing was based on crossing the crop type with the degree of infestation, using maximum likelihood classification. The accuracy of the method of detection used was demonstrated by the percentage of correctly classified pixels (45%), the reasons for confusion and the clarity of the visual representations. The success of our approach is highly influenced by the availability of high-quality images. Its feasibility, which is linked therefore to technical constraints, is examined in relation to space agency programmes.