Detection of Tree Disease by Large Scale Aerial Photography. An Experience in Trentino - North-Italy

1989

S. 1-70

8 Abb., 7 Tab., 37 Lit. Ang.

Unselbständiges Werk

200033976

Annali dell'Istituto Sperimentale per l'Assestamento Forestale e per l'Alpicoltura 1985-87 ; 10, S. 1-70

The aim of the research was to verify the degree of agreement between individual evaluations of the state of health of a sample of trees, carried out both from the real evidence on the ground and photographic evidence, (taken at low altitude from a helicopter), identifying which type of film used, (normal colour or infra-red false colour), would prove capable of producing the best results in relation to the stated aim. From amongst several forested areas in the province of Trento ten sample sites were chosen, with a vegetation of predominantly silver-fir and spruce, which were affected with acute tree disease. On these sites photographs were taken from a helicopter, at low altitude, using a pair of 35 mmm camercas mounted together and loaded respectively with normal colour film and infrared false colour film. At the same times as this remote sensing, surveys were carried out on the ground by two surveyors, who normally undertake similar work on behalf of the provincial forestry service. The two surveyors, independently, carried out the classification of the state f health of the sample subjects. Two photo-interpreters, suitably trained, then independently undertook the interpretation of the photographic evidence in terms of the recognition of the species and the state of health (defoliation and de-pigmentation) of each of the sample trees. Finally, the following analyses of the agreement of the results were carried out. Between the surveyors' classifications of the state of health of the trees carried out on the ground. Between the photo-interpreters (for each type of film) in their classification of the species and the state of health of the trees, based on the photographi evidence. Between the averages of the classifications relating to the species and to the state of health of the trees, carried out by the two photointerpreters respectively on the normal film photographs, the false colour photographs, and on both types of photography analysed together. Between the averages of the classifications of the species and the state of health of the trees carried out on the various types of photographic evidence, in comparison with similar classificatins carried out on the ground. The degree of agreement between the classifications of the state of health of the trees, on the photographs and on the ground, varies, broadly speaking, between 65% and 95% and is therefore shown to be not always satisfactory. Whilst, on the other had, the results relting to the identification of species, on both types of film, seem satisfactory. Given the variability in the evaluations of the photointerpreters, and the differing tendencies of the classifications of the state of tree disease, (varying according to the species, the symptoms, and the type of photographs considered) it is not possible to regard as particularly convincing the differences in agreement .....